Merge pull request #27362 from open-webui/dev

0.11.0

.env.example

@@ -13,6 +13,12 @@ OPENAI_API_KEY=''
 # CORS_ALLOW_ORIGIN='http://localhost:5173;http://localhost:8080'
 CORS_ALLOW_ORIGIN='*'
 
+# Set to false to keep memory tools enabled without adding memory context to the system context.
+ENABLE_MEMORY_SYSTEM_CONTEXT=true
+
+# Set to false to disable workspace Tools and Functions.
+ENABLE_PLUGINS=true
+
 # For production you should set this to match the proxy configuration (127.0.0.1)
 FORWARDED_ALLOW_IPS='*'
 

.github/FUNDING.yml

@@ -1 +1 @@
-github: tjbck
+github: open-webui

.github/ISSUE_TEMPLATE/config.yml

@@ -1 +1,5 @@
 blank_issues_enabled: false
+contact_links:
+  - name: 🔒 Report a Security Vulnerability
+    url: https://github.com/open-webui/open-webui/security
+    about: Do NOT open a public issue for security vulnerabilities, suspected vulnerabilities, or any security-related concern. Please review our Security Policy and report privately via the "Report a vulnerability" button so it can be handled as a private advisory.

.github/workflows/backend.yaml

@@ -7,10 +7,10 @@ name: Python CI
 on:
   push:
     branches: [main, dev]
-    paths: ['backend/**', 'pyproject.toml', 'uv.lock']
+    paths: ['backend/**', 'pyproject.toml', 'uv.lock', '.github/workflows/backend.yaml']
   pull_request:
     branches: [main, dev]
-    paths: ['backend/**', 'pyproject.toml', 'uv.lock']
+    paths: ['backend/**', 'pyproject.toml', 'uv.lock', '.github/workflows/backend.yaml']
 
 concurrency:
   group: backend-${{ github.ref }}
@@ -38,3 +38,6 @@ jobs:
 
       - name: Verify formatting
         run: ruff format --check . --exclude .venv --exclude venv
+
+      - name: Detect logic errors
+        run: ruff check --select=F --ignore=F401,F403,F405,F541,F811,F841 --output-format=github .

.github/workflows/docker.yaml

@@ -75,6 +75,17 @@ jobs:
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v3
 
+      - name: Prepare CI Dockerfile
+        run: |
+          awk '
+            /^FROM --platform=\$BUILDPLATFORM node:/ {
+              print
+              print "ENV NODE_OPTIONS=\"--max-old-space-size=12288\""
+              next
+            }
+            { print }
+          ' Dockerfile > "${RUNNER_TEMP}/Dockerfile"
+
       - name: Log in to the Container registry
         uses: docker/login-action@v3
         with:
@@ -115,6 +126,7 @@ jobs:
         id: build
         with:
           context: .
+          file: ${{ runner.temp }}/Dockerfile
           push: true
           platforms: ${{ matrix.platform.arch }}
           labels: ${{ steps.meta.outputs.labels }}
@@ -231,6 +243,70 @@ jobs:
         run: |
           docker buildx imagetools inspect ${{ env.FULL_IMAGE_NAME }}:${{ steps.meta.outputs.version }}
 
+  notify-helm-charts:
+    runs-on: ubuntu-latest
+    needs: [merge]
+    if: ${{ !cancelled() && needs.merge.result == 'success' && (github.ref == 'refs/heads/dev' || startsWith(github.ref, 'refs/tags/v')) }}
+    steps:
+      - name: Create Helm charts app token
+        id: helm-app-token
+        uses: actions/create-github-app-token@v2
+        with:
+          app-id: ${{ secrets.HELM_CHARTS_APP_ID }}
+          private-key: ${{ secrets.HELM_CHARTS_APP_PRIVATE_KEY }}
+          owner: ${{ github.repository_owner }}
+          repositories: helm-charts
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Verify published Open WebUI image
+        id: image
+        run: |
+          set -euo pipefail
+
+          image_name="ghcr.io/${GITHUB_REPOSITORY,,}"
+          ref_name="${GITHUB_REF_NAME}"
+
+          if [ "${GITHUB_REF}" = "refs/heads/dev" ]; then
+            image_tag="dev"
+          else
+            image_tag="${ref_name#v}"
+          fi
+
+          docker buildx imagetools inspect "${image_name}:${image_tag}"
+          echo "tag=${image_tag}" >> "${GITHUB_OUTPUT}"
+
+      - name: Dispatch Helm chart automation
+        uses: actions/github-script@v8
+        with:
+          github-token: ${{ steps.helm-app-token.outputs.token }}
+          script: |
+            const isDev = context.ref === 'refs/heads/dev';
+            const eventType = isDev
+              ? 'open-webui-dev-image-published'
+              : 'open-webui-release-published';
+            const refName = context.ref.replace('refs/heads/', '').replace('refs/tags/', '');
+            const appVersion = refName.startsWith('v') ? refName.slice(1) : refName;
+            const payload = {
+              image_tag: isDev ? 'dev' : appVersion,
+              source_ref: context.ref,
+              source_sha: context.sha,
+              source_run_id: String(context.runId),
+              source_repository: context.repo.repo,
+            };
+
+            if (!isDev) {
+              payload.app_version = appVersion;
+            }
+
+            await github.rest.repos.createDispatchEvent({
+              owner: context.repo.owner,
+              repo: 'helm-charts',
+              event_type: eventType,
+              client_payload: payload,
+            });
+
   copy-to-dockerhub:
     runs-on: ubuntu-latest
     if: ${{ !cancelled() && (github.ref == 'refs/heads/main' || startsWith(github.ref, 'refs/tags/v')) }}

.github/workflows/frontend.yaml

@@ -43,6 +43,8 @@ jobs:
 
       - name: Production build
         run: npm run build
+        env:
+          NODE_OPTIONS: --max-old-space-size=8192
 
   # ── Vitest unit tests ────────────────────────────────────────────────────
   unit-tests:

.github/workflows/issue-label.yaml

@@ -0,0 +1,43 @@
+name: Issue Labeler
+
+on:
+  issues:
+    types: [opened]
+
+permissions:
+  issues: write
+
+jobs:
+  label-bug-reports:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Add "bug" label to unlabeled bug reports
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const issue = context.payload.issue;
+
+            // Web-form submissions already carry the label from the issue template
+            if (issue.labels.some((label) => label.name === 'bug')) {
+              return;
+            }
+
+            const title = issue.title ?? '';
+            const body = issue.body ?? '';
+
+            // Freeform bug reports: "issue: ...", "bug: ...", "fix: ...", "[Bug] ...", "issue/UX: ..."
+            const bugLikeTitle = /^\s*\[?(bug|issue|fix)\]?\s*[:/\-]/i.test(title);
+
+            // API/CLI-created issues that reproduce the bug report form structure.
+            // Only headings distinctive to the bug form (both are required fields there) —
+            // generic headings like "Expected Behavior" also appear in freeform feature requests.
+            const bugFormBody = /###\s*(Installation Method|Open WebUI Version)/i.test(body);
+
+            if (bugLikeTitle || bugFormBody) {
+              await github.rest.issues.addLabels({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                issue_number: issue.number,
+                labels: ['bug']
+              });
+            }

.gitignore

@@ -310,3 +310,4 @@ dist
 cypress/videos
 cypress/screenshots
 .vscode/settings.json
+.cptr

CHANGELOG.md

@@ -5,6 +5,557 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.11.0] - 2026-07-27
+
+### Added
+
+- 🎨 **Redesigned interface.** Open WebUI has been visually rebuilt from the ground up. All aspects of the User Interface, from the chat view to the admin panel. Now with a narrower conversation column, lighter typography, tidier spacing, consistent menus and dropdowns, clearly outlined text boxes, and settings rearranged. [Commit](https://github.com/open-webui/open-webui/commit/aedb6bef4e2eb12234c02085a545ff395d96db18), [Commit](https://github.com/open-webui/open-webui/commit/b3255a36569f295766271b8a2b0bd969b4083b9f), [Commit](https://github.com/open-webui/open-webui/commit/ba067258dea2229a9956077b3b0d7b1c68b56f66), [Commit](https://github.com/open-webui/open-webui/commit/8dd862d3383978f21111e63fb2d6029711abed9a), [Commit](https://github.com/open-webui/open-webui/commit/263bbc77d803e83b9af4b04c0cae29705af5f072), [Commit](https://github.com/open-webui/open-webui/commit/f8ea15b84a274712dca33daa970f63ed7368043e), [Commit](https://github.com/open-webui/open-webui/commit/9f17c5960a0e47a09773da4bba12997a31222fc8), [Commit](https://github.com/open-webui/open-webui/commit/6772b1cb4f4e0d3dc166956014e6e7b9bddc721a), [Commit](https://github.com/open-webui/open-webui/commit/d3fd860c131846a9458888f9c256a9a29f3767f2), [Commit](https://github.com/open-webui/open-webui/commit/f1584b5a3764f72de2de6caad507e7c39ad19c23), [Commit](https://github.com/open-webui/open-webui/commit/2e8d92c7b1a9bb8d35f4a27ba3c73368d735c480), [Commit](https://github.com/open-webui/open-webui/commit/e58a4633b15ae53d33fc3b46cb97c76d86be325f), [Commit](https://github.com/open-webui/open-webui/commit/04b146f2cec7e6a01e9d3590eb83655c128fa3c7), [Commit](https://github.com/open-webui/open-webui/commit/e5e2cd78769639b2df83776f1b991966f922f8b4), [Commit](https://github.com/open-webui/open-webui/commit/3316ba76aabe5429596ffd130fd36be4d5c3aa6c), [Commit](https://github.com/open-webui/open-webui/commit/6fcb38fe2e0aded9b85f655cd8f3279e9f4e765e), [Commit](https://github.com/open-webui/open-webui/commit/421da674468f638f72cc5266c5a3874aa3bca3b7), [Commit](https://github.com/open-webui/open-webui/commit/d0bea60581eaa07d41f92ad8f86007e83247e061), [Commit](https://github.com/open-webui/open-webui/commit/21e180182a5096481d4cbb1a8f94212c0a515a40), [Commit](https://github.com/open-webui/open-webui/commit/d027a32ed134ae104f2f142ba45ff38e56215c5f), [Commit](https://github.com/open-webui/open-webui/commit/437c06c4795a72700295d7690d5fd65d1153372c), [Commit](https://github.com/open-webui/open-webui/commit/1bf05ebc7d135d74969438824778c9f73243ba8d), [Commit](https://github.com/open-webui/open-webui/commit/fd07e3a8e3e619f3712067765b416f0925fa80d3), [Commit](https://github.com/open-webui/open-webui/commit/d3ea51fd466a8741afc4dfd4f0d0f2f77fb6467f), [Commit](https://github.com/open-webui/open-webui/commit/4da2ff2655d9abb851805da127cf60b4d9ad1aa7), [Commit](https://github.com/open-webui/open-webui/commit/2fcb36267f034f2b83f936bfacedc20b680a2710), [Commit](https://github.com/open-webui/open-webui/commit/1428a4ddce4998cb3664a5ce37e176442dd426fa), [Commit](https://github.com/open-webui/open-webui/commit/bc8d24c951e9a2c973fc2dd1f2832a2b0855bc0e), [Commit](https://github.com/open-webui/open-webui/commit/704d07e9a20a830aad7bfc5131b0d92621cf0691), [Commit](https://github.com/open-webui/open-webui/commit/e88d2e053c2a63cce3823c9b6006f4a184c4fef2), [Commit](https://github.com/open-webui/open-webui/commit/6940297486d4a5de127efd1a5148b0adcebe87e3), [Commit](https://github.com/open-webui/open-webui/commit/9ca8cf528af1c49da3f0a2bc3c6ca95c1dedbcf5), [Commit](https://github.com/open-webui/open-webui/commit/c4efa81d08c425678c810c51b4d62716e1e57117), [Commit](https://github.com/open-webui/open-webui/commit/bb12b1a18b77d80829cedb2d5bf965808222415b), [Commit](https://github.com/open-webui/open-webui/commit/49abfbdd155dc22882fdcb09989e4f4964db16ee), [#27178](https://github.com/open-webui/open-webui/pull/27178), [Commit](https://github.com/open-webui/open-webui/commit/dcc7fb1e8ef144205531829f8a56e52171c4d63d), [Commit](https://github.com/open-webui/open-webui/commit/5c505c1119fec6170c1bc092ed162f86262a887e)
+- 🤖 **Sub-agents.** Administrators can now enable sub-agents, which let a model hand parts of a task to background helper agents that run their own tool-driven conversations and report results back into the chat, tuned through new "ENABLE_SUBAGENTS", concurrency, iteration, and system-prompt settings. [Commit](https://github.com/open-webui/open-webui/commit/7088d245bb45fc69c0b22748563b9f3c6f0daa73), [Commit](https://github.com/open-webui/open-webui/commit/2f37e853d1259a901f736a823bad29dcc2c3b130), [Commit](https://github.com/open-webui/open-webui/commit/959558fd82eb2a3c980231acd500b73ba4b698b3), [Commit](https://github.com/open-webui/open-webui/commit/3005b7bc71fcbd5abc6e73c3e4caa4ea781cdb76)
+- 📂 **Folder pages.** Opening a folder now takes you to its own page, where its chats load a page at a time, can be sorted by title or last updated, and you can start a new chat straight from the folder. [Commit](https://github.com/open-webui/open-webui/commit/409fb39717be9ab7becd9e8c01801a08c5bae318)
+- ⏲️ **Chat timers.** The assistant can now set a timer that brings a prompt back into the conversation later, after a delay or at a set time, and can drop it automatically if you read the chat or reply before it fires. [Commit](https://github.com/open-webui/open-webui/commit/b23ddeb2800098c6352203ec8fbe9fca40ba415c)
+- 🔔 **Notification targets.** Notifications now have their own settings tab where you can send them to several webhook destinations, each picking which events it wants, from chats finishing or failing to channel messages and calendar alerts, with a test button and a choice between always notifying or only when you are away, and any webhook you already had is carried over for you. [Commit](https://github.com/open-webui/open-webui/commit/c55e373b994d3a14c99a97f44261422012f63266), [Commit](https://github.com/open-webui/open-webui/commit/cf235738f5a44db415012b3b0ebc1f6e752f5439), [Commit](https://github.com/open-webui/open-webui/commit/200d447f6289faca42f2a666bbabae2c7f3ebadf), [#24750](https://github.com/open-webui/open-webui/issues/24750)
+- 🗯️ **Full replies in channels.** A reply from the assistant in a channel is now saved and shown in full, with its reasoning, tool calls and other structured parts, where it previously came through blank. [Commit](https://github.com/open-webui/open-webui/commit/498cdab9a548d7d2fd19c389204ee26236fc7efe), [#26720](https://github.com/open-webui/open-webui/pull/26720), [#27409](https://github.com/open-webui/open-webui/pull/27409), [#26707](https://github.com/open-webui/open-webui/issues/26707), [#26656](https://github.com/open-webui/open-webui/issues/26656)
+- 📣 **Notifications from the assistant.** The assistant can now send you a notification itself when something is worth your attention, so a long task can reach you after you have moved on to something else. [Commit](https://github.com/open-webui/open-webui/commit/c55e373b994d3a14c99a97f44261422012f63266), [Commit](https://github.com/open-webui/open-webui/commit/200d447f6289faca42f2a666bbabae2c7f3ebadf)
+- 🌎 **Share a chat with anyone holding the link.** A shared chat can now be set to Open so it opens without signing in, with visitors no longer bounced to the sign-in page on their way to it, which administrators must first allow through a new "Chats Open Sharing" permission that stays off by default, and such pages ask search engines not to index them. [Commit](https://github.com/open-webui/open-webui/commit/1f0dc90abe879a55f654f2333e29fb0f630831c7), [Commit](https://github.com/open-webui/open-webui/commit/0e0d08382ac0d05b1ad98c47e8a2e37df2a185bb)
+- 🔖 **Chat variables.** A model's system prompt can now declare fields such as text boxes and dropdown lists that you fill in for a conversation, with the values saved alongside the chat and carried over when it is forked or cloned. [Commit](https://github.com/open-webui/open-webui/commit/bef8ae4b2f05ca49ed88a02ab7a3cdc11b62c4f1), [Commit](https://github.com/open-webui/open-webui/commit/4e869011cd5040b5d6a197fc83d5f50d2425dbc2), [Commit](https://github.com/open-webui/open-webui/commit/1e88367cc837b39c0e9958fefbe053803336dce2), [Commit](https://github.com/open-webui/open-webui/commit/8cbb7f765cfc9c9b3237a6c5593cd93f849033f0), [Commit](https://github.com/open-webui/open-webui/commit/b35e2d265a4e4a2f2a075b31917d48be1dd9ef19), [Commit](https://github.com/open-webui/open-webui/commit/239cb740077a14e452ad002a1e671a09ab558e40), [#26915](https://github.com/open-webui/open-webui/discussions/26915)
+- 🗄️ **LDAP group synchronization.** Administrators can now map LDAP groups to Open WebUI groups from the authentication settings, with optional automatic creation of missing groups, so a user's group memberships are kept in step with the directory each time they sign in. [#27263](https://github.com/open-webui/open-webui/pull/27263), [#18015](https://github.com/open-webui/open-webui/issues/18015)
+- 👥 **Restrict sharing with groups.** Admins can now stop resources from being shared with entire groups through a new "USER_PERMISSIONS_ACCESS_GRANTS_ALLOW_GROUPS" permission, which stays enabled by default so existing group sharing keeps working untouched. [Commit](https://github.com/open-webui/open-webui/commit/4ed19d504bd30c0fc801e9228d9816669ec1c09c), [Commit](https://github.com/open-webui/open-webui/commit/f84dabe3d97ff701c097055023f28f3f2f7ebd07), [Commit](https://github.com/open-webui/open-webui/commit/77da3d8c81b9a6fda4354619de94f5d433328d8e), [Commit](https://github.com/open-webui/open-webui/commit/84e4d6ef8277f4b4f3ac4d355b3219e9b5a37268), [#27124](https://github.com/open-webui/open-webui/pull/27124)
+- 🤝 **Shared folder collaboration.** People with access to a shared folder can now use its files and system prompt as knowledge in chat and, with write access, rename and manage the folder, all according to their read or write permission. [Commit](https://github.com/open-webui/open-webui/commit/797293c74957bd79e42262d1dc0fd637a45d0357), [Commit](https://github.com/open-webui/open-webui/commit/caa2457c17e592587b804f21054cc000944af75c), [Commit](https://github.com/open-webui/open-webui/commit/009715cd63d1c8b5320aba68e9a70afcde519016), [Commit](https://github.com/open-webui/open-webui/commit/53ccd718a53de25bb6d61476a6617bfb3f130a44)
+- 👁️ **Chat previews in the sidebar.** Hovering a chat in the sidebar now shows a compact preview of its recent messages, so you can find the conversation you want without opening it. [Commit](https://github.com/open-webui/open-webui/commit/d0f7da4f45b8831b09b2ab3ec8f91aa354d90ba3), [Commit](https://github.com/open-webui/open-webui/commit/aaf2834db758bfec69408ab4cabcf324965c221c), [Commit](https://github.com/open-webui/open-webui/commit/1513ddaf58fe18029086880461d7cad0649a699c), [Commit](https://github.com/open-webui/open-webui/commit/93bd05271c07c249978d69abf3297fd2841900f9)
+- 🕗 **Local message timestamps.** Message timestamps now appear on hover in your device's local date and time format, with the full weekday and date shown in a tooltip. [Commit](https://github.com/open-webui/open-webui/commit/797293c74957bd79e42262d1dc0fd637a45d0357), [Commit](https://github.com/open-webui/open-webui/commit/f84dabe3d97ff701c097055023f28f3f2f7ebd07)
+- 📇 **User variables.** You can now store your own values in account settings, such as your role or how you like answers written, and a model's system prompt can insert them wherever they are needed. [Commit](https://github.com/open-webui/open-webui/commit/bd5d7b2e879511882429075d804d9222956f4a1a), [Commit](https://github.com/open-webui/open-webui/commit/212eec408ca320edfa2271e604d10e14b6a9bc1a), [Commit](https://github.com/open-webui/open-webui/commit/793a43d9c48225925929eb312fe2b70d5914d1da)
+- 🧺 **Automations that file their chats away.** An automation can now be pointed at one of your folders, from the dialog, the editor or by asking the assistant, so each run lands there instead of loose in your chat list, and the folder is cleared automatically if it is later deleted. [Commit](https://github.com/open-webui/open-webui/commit/f798d05586a140f1a6b51f1e51b2b2a63d079d45), [Commit](https://github.com/open-webui/open-webui/commit/bab71ed08b5af6f4a8ff2daa02792baae9edab03), [Commit](https://github.com/open-webui/open-webui/commit/db5c092299471444c356216d5ef39b382ba1aa1e)
+- 🔵 **See what you have not read yet.** Folders in the sidebar now carry a count of chats with something new in them, a folder's own page marks unread chats with a dot, shows a spinner on any still generating, clears the dot as you open one, and keeps itself up to date as replies finish elsewhere, unread chats sort to the top of a folder, and you can mark a single chat unread again mark everything in a folder read, or mark every chat read at once from the sidebar. [Commit](https://github.com/open-webui/open-webui/commit/f798d05586a140f1a6b51f1e51b2b2a63d079d45), [Commit](https://github.com/open-webui/open-webui/commit/f867825bf3b7699bc2bd967ef46b2bb63e48b098), [Commit](https://github.com/open-webui/open-webui/commit/1de36d600f7191c28a98bf4b347b44cf8f1bef43), [Commit](https://github.com/open-webui/open-webui/commit/85c47fb467177ed811ba77dfe62461dfbe8e2548), [Commit](https://github.com/open-webui/open-webui/commit/b7489bbc6c4e376c017edffd8da3c2eb4e6c1c8e), [Commit](https://github.com/open-webui/open-webui/commit/3cd72ee6a8e93dc39a4d4c173117056e25326c8a), [Commit](https://github.com/open-webui/open-webui/commit/6f93ecd4fd77b0d51a5fbbd2fc3fd6d151036a55), [Commit](https://github.com/open-webui/open-webui/commit/e5a08d52208e8b1ed07ff94906e27d174146b1ca), [Commit](https://github.com/open-webui/open-webui/commit/8ddf119570b3c0b04b673d41cb2363370c23939b)
+- 🗜️ **Compact a chat on demand.** Typing a compact command in a long conversation now summarizes the earlier turns straight away, instead of waiting for it to happen automatically once the conversation grows past the threshold. [Commit](https://github.com/open-webui/open-webui/commit/7a9928ef172b7c280c377c86cb52957e39340158), [Commit](https://github.com/open-webui/open-webui/commit/75894161e46aafe30a7db4ecc946af60f04495e4)
+- 🌿 **Fork a chat.** Every response now has a fork button that copies the conversation up to that point into a new chat which remembers where it branched, so you can carry on down a different path without touching the original. [Commit](https://github.com/open-webui/open-webui/commit/63ada247066dfc51e0e9559366f0cfd9a98db40b), [Commit](https://github.com/open-webui/open-webui/commit/cf887b68ea58bcd1d8b035842c4ea35a5113ed8b), [Commit](https://github.com/open-webui/open-webui/commit/73421c5b42ac5c2ebc5faa520b7ed3fa0e39f10d), [Commit](https://github.com/open-webui/open-webui/commit/e769f9ff4f9fa7b0faebdd4f34ff98fe0dcc300d)
+- 📌 **Pin the conversation map.** The chat overview now has a pin control that stops it recentring on the newest message, so you can keep looking at the branch you were reading while a reply comes in. [#25736](https://github.com/open-webui/open-webui/pull/25736)
+- 📊 **Chat status at a glance.** The slash menu now shows how full the context window is, and a new status command opens a panel with context usage, queued messages, running tasks, and the chat ID. [Commit](https://github.com/open-webui/open-webui/commit/7a9928ef172b7c280c377c86cb52957e39340158), [Commit](https://github.com/open-webui/open-webui/commit/263bbc77d803e83b9af4b04c0cae29705af5f072)
+- 🎹 **Customizable keyboard shortcuts.** Most keyboard shortcuts can now be rebound to key combinations of your choosing in settings, which saves them to your account, warns you when two actions share a combination, and offers a reset to the defaults, with moving to the previous or next chat and opening the controls panel available to bind as well. [Commit](https://github.com/open-webui/open-webui/commit/343eb1d659262cc9d762a2ce49bef25434a03bfa), [Commit](https://github.com/open-webui/open-webui/commit/de681aa543b356d456b83c13ad88c7f9941319e7), [#26624](https://github.com/open-webui/open-webui/pull/26624)
+- ⌨️ **Turn keyboard shortcuts off.** A new switch in the keyboard settings disables every configurable shortcut and hides its hint, so combinations that clash with your browser or operating system pass straight through. [#27300](https://github.com/open-webui/open-webui/pull/27300), [#1008](https://github.com/open-webui/open-webui/issues/1008)
+- ⌨️ **Skills in slash commands.** Typing a slash in the message input now lists your skills alongside your prompts, grouped under headings and with descriptions on hover, so you can attach a skill without leaving the keyboard. [Commit](https://github.com/open-webui/open-webui/commit/9588c97e64e10d161a9a0ab1ab9ba3fee6cbb94d)
+- 📎 **Attach anything with the at menu.** Typing an at sign in the message input now searches your folders, knowledge collections, and individual files as well as your models, and pasting a link offers it as a web page or YouTube attachment. [Commit](https://github.com/open-webui/open-webui/commit/e8b4c7f9e212253267b8c79dc5360894f0d91bec)
+- 📝 **Chat with a note.** Chatting with a note now gives you the full chat experience, including model choice, tools and file attachments, alongside suggested prompts, a button to insert a response straight into the note, edits that appear in the note as the assistant makes them, and as many separate conversations per note as you want to keep. [Commit](https://github.com/open-webui/open-webui/commit/423cafd4e75e34b487f3b5d10ec1c506f073b3da), [Commit](https://github.com/open-webui/open-webui/commit/185bca8552ee3f87ea95fdcad32a433924881b9a)
+- ↕️ **Sort your lists.** The notes, prompts, models, knowledge, skills, tools and functions lists can now be sorted by title or by when they were last updated, in either direction, by clicking the column headings. [Commit](https://github.com/open-webui/open-webui/commit/30c91e46e5d237bee3c8805bd54749408cc2727a), [Commit](https://github.com/open-webui/open-webui/commit/56f2cb530259df5393ea1ae844bad5cc6b2c810c), [#27457](https://github.com/open-webui/open-webui/pull/27457), [#27456](https://github.com/open-webui/open-webui/discussions/27456)
+- 🗒️ **Notes without stored contents.** A note whose contents were never filled in now opens and saves normally instead of failing. [Commit](https://github.com/open-webui/open-webui/commit/6c59ef313fdaac7fa8e9089be758c4651f8b9412)
+- 📄 **Note attachments.** Notes now have an upload option in their menu and show attached files above the note itself, where you can open or remove them, instead of only accepting files dropped onto the page. [Commit](https://github.com/open-webui/open-webui/commit/c4c4ab57e33bb4359b51a4c255f4569fa38a5058), [Commit](https://github.com/open-webui/open-webui/commit/0dc93b8ae798ad834e7f3327571452bfdf4c218f)
+- 🗂️ **The assistant can search your attachments.** A new Files capability lets the model list the files attached to the chat and search them by meaning or by exact text, and read the parts it needs, rather than having their whole contents pushed into the conversation up front, and knowledge collections or notes attached to a chat are now announced to the model so it can query those the same way. [Commit](https://github.com/open-webui/open-webui/commit/57e60423b9963c4a69fdfda7ae5799efc5583010), [Commit](https://github.com/open-webui/open-webui/commit/55e0801dab8fe5f8bceff7d7c49772676724b8b1), [#26711](https://github.com/open-webui/open-webui/pull/26711), [#27232](https://github.com/open-webui/open-webui/issues/27232), [#26708](https://github.com/open-webui/open-webui/issues/26708)
+- 🔎 **Search in the attachment menu.** The attachment menu now lets you search your knowledge bases, notes, files, and chats instead of scrolling to find them, with matching text shown for chats. [Commit](https://github.com/open-webui/open-webui/commit/668f9fe3905fea5fdfccb2b9b308c8f4d7d2f061)
+- ⚗️ **Default file upload mode.** You can now choose in settings how attached files are handled by default, rather than picking that on each upload. [#20900](https://github.com/open-webui/open-webui/pull/20900), [#18431](https://github.com/open-webui/open-webui/issues/18431)
+- ⬇️ **Response auto-scroll toggle.** A new interface setting lets you stop the view following a reply as it is written, so you can read earlier text while generation continues. [Commit](https://github.com/open-webui/open-webui/commit/cea991260f279f004489dab5d930bc25b0abf612), [#26826](https://github.com/open-webui/open-webui/pull/26826)
+- 📜 **Client certificates for SearXNG.** Web search can now present a client certificate to a SearXNG instance that requires one, through new "SEARXNG_CLIENT_CERT_FILE" and "SEARXNG_CLIENT_KEY_FILE" settings. [Commit](https://github.com/open-webui/open-webui/commit/def26ce266c2e1d80e31b3ad12099db61a674832), [#26992](https://github.com/open-webui/open-webui/issues/26992)
+- 🔭 **OpenSERP web search.** Web search can now run against a self-hosted OpenSERP instance, which returns results from several major search engines without any API key, configured through a new "OPENSERP_BASE_URL" setting. [#27437](https://github.com/open-webui/open-webui/pull/27437), [#27438](https://github.com/open-webui/open-webui/issues/27438)
+- 🥇 **Model order as a setting.** Administrators can now set the order models appear in through a new "MODEL_ORDER_LIST" variable, so the arrangement survives a restart on instances that do not persist configuration. [#27420](https://github.com/open-webui/open-webui/pull/27420), [#27206](https://github.com/open-webui/open-webui/issues/27206)
+- ⏱️ **Idle cap for streamed replies.** Administrators can now set an "AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT" that ends a streamed reply when the provider stops sending anything for that long, instead of holding the connection open until the overall timeout expires. [Commit](https://github.com/open-webui/open-webui/commit/4a7d4ebadac27d652ec200fa3939f10e9a5c17ed), [Commit](https://github.com/open-webui/open-webui/commit/c727643e05f3597395ee1a60b17117d04f693a18)
+- 🖼️ **Media types an extraction engine may handle.** Administrators can now list which image and video types the configured content extraction engine is allowed to process, instead of media being passed to it only when the engine is the external one, so an engine with its own text recognition can take images. [Commit](https://github.com/open-webui/open-webui/commit/db2d24896b0682191a54f41c6b9f0b9d2971637f), [#26940](https://github.com/open-webui/open-webui/pull/26940), [#14768](https://github.com/open-webui/open-webui/issues/14768)
+- 🧵 **Where a channel reply lands.** Administrators can now choose whether a reply to a mention posts in a thread under that message or straight into the channel. [Commit](https://github.com/open-webui/open-webui/commit/db2d24896b0682191a54f41c6b9f0b9d2971637f), [#27410](https://github.com/open-webui/open-webui/pull/27410)
+- 📚 **Limits for knowledge tools.** Administrators can now set how much a knowledge search or file view may return, how many files one search may scan, and how many matches are reported, and a knowledge command's whole output is now capped so a single call cannot flood the conversation. [Commit](https://github.com/open-webui/open-webui/commit/11e61b69ebd922602edc37ded7b42fdd43bb8456), [#27524](https://github.com/open-webui/open-webui/pull/27524), [#27327](https://github.com/open-webui/open-webui/issues/27327), [#26139](https://github.com/open-webui/open-webui/issues/26139)
+- 🎛️ **File streaming chunk size.** Administrators can now tune how large each chunk of a streamed file transfer is through a new "AIOHTTP_FILE_STREAM_CHUNK_SIZE" setting. [Commit](https://github.com/open-webui/open-webui/commit/429f2df50cd2f0ec8d0a1bb4136a46e2f94a4bf5)
+- 🪛 **Model for summarizing long chats.** Administrators can now pick a dedicated model to write context compaction summaries, separate from the task model, with the conversation's own model used when none is chosen. [#26806](https://github.com/open-webui/open-webui/pull/26806), [#27051](https://github.com/open-webui/open-webui/issues/27051)
+- 📏 **Context compaction token cap.** Administrators can now set a "Token Cap" that limits how high per-model context compaction thresholds are allowed to reach, giving finer control over long-conversation summarization. [Commit](https://github.com/open-webui/open-webui/commit/5c389ad93f0668d4bab717d14bd189b679338ef2), [Commit](https://github.com/open-webui/open-webui/commit/31996a5acfe1458720fa19f1b9fb4da95749b5e6), [Commit](https://github.com/open-webui/open-webui/commit/44c2a27ce0695d8e9c7e72f9a84dc325cb15096a)
+- ⚖️ **Retained messages after compaction.** Administrators can now set what share of recent messages survives when a long conversation is summarized, between a tenth and half of it. [Commit](https://github.com/open-webui/open-webui/commit/33cf3fbb7f017ab1b79dce5c5ca4d4e1c3092844), [#27050](https://github.com/open-webui/open-webui/issues/27050)
+- 🧠 **Memory as a per-model capability.** Whether a model receives your stored memories is now a switch on the model itself, so it can be left on for everyday assistants and off for ones that should start from nothing. [Commit](https://github.com/open-webui/open-webui/commit/6732852ce6c1a2a445bc001b122c60cf12e0278b), [#26861](https://github.com/open-webui/open-webui/pull/26861), [#18610](https://github.com/open-webui/open-webui/discussions/18610)
+- ☑️ **Searchable model pickers.** When editing a model, the Tools, Skills, Knowledge, Voice, Filters and Actions pickers now let you search and toggle items in place, select or clear them all at once, and see what is active at a glance. [Commit](https://github.com/open-webui/open-webui/commit/e355959e9156fd61a1105953d731d007fbb4bae3), [Commit](https://github.com/open-webui/open-webui/commit/e1f96aa20ef80c8b01e3a81039d2cca3001d0ef5), [Commit](https://github.com/open-webui/open-webui/commit/fa889837e9e90c7284def56fac2daed20a3ce699), [Commit](https://github.com/open-webui/open-webui/commit/5424ac58917d2a21f21256b62d1c42f1a8c51c51), [Commit](https://github.com/open-webui/open-webui/commit/ea2e3d0afc76fa99f2af665fd50425dd28d000d5), [Commit](https://github.com/open-webui/open-webui/commit/cda5bdb9d42886dfe74d05907179e3f96097030c), [#26758](https://github.com/open-webui/open-webui/issues/26758)
+- 🎚️ **Switch for single sign-on.** OAuth and OIDC now have their own on and off switch in the authentication settings, matching the LDAP one above it, so sign-in through a provider can be turned off without clearing the configuration. [#26988](https://github.com/open-webui/open-webui/pull/26988)
+- 🖲️ **One sign-in attempt at a time.** The sign-in, sign-up and LDAP form now disables its buttons while a request is in flight, so a slow response no longer turns repeated clicks or Enter presses into several concurrent attempts. [#27416](https://github.com/open-webui/open-webui/pull/27416), [#27264](https://github.com/open-webui/open-webui/issues/27264)
+- 🛂 **Trusted clients for token exchange.** Administrators can now list which OAuth clients may have their tokens exchanged for a session, through a new "OAUTH_TOKEN_EXCHANGE_TRUSTED_CLIENT_IDS" setting, so a token a person obtained by signing in to an unrelated application of the same provider can no longer be turned into a session as that person. [#27546](https://github.com/open-webui/open-webui/pull/27546), [Commit](https://github.com/open-webui/open-webui/commit/b190dcf3caa00dc8b7b9c7312828298d9143f60d), [Commit](https://github.com/open-webui/open-webui/commit/c4332be71e6e9c314e8a13b9d2819a6932561630)
+- 🚪 **Throttle for token exchange.** Administrators can now cap how often the OAuth token exchange endpoint may be called from one address through new "OAUTH_TOKEN_EXCHANGE_RATE_LIMIT" and "OAUTH_TOKEN_EXCHANGE_RATE_LIMIT_WINDOW" settings, which bound automated attempts with leaked or guessed tokens and stay off until set. [Commit](https://github.com/open-webui/open-webui/commit/453b9fb0291c0de8957a2713988c7c53dbcc5465)
+- 🔏 **PKCE for every sign-in provider.** The code challenge setting now applies to Google, Microsoft and GitHub sign-in as well as OpenID Connect, so the same protection covers every provider. [Commit](https://github.com/open-webui/open-webui/commit/40320c113637f80e0466e30cae63ab9ac1ba596e), [#27302](https://github.com/open-webui/open-webui/pull/27302)
+- 🔤 **Embeddings through the OpenAI-compatible API.** Integrations built on OpenAI client libraries can now create embeddings through the Ollama proxy, so embedding requests go through the same sign-in and model access rules as chat instead of needing direct access to Ollama. [#27332](https://github.com/open-webui/open-webui/pull/27332), [Commit](https://github.com/open-webui/open-webui/commit/9f00b62b3a005b030ffaf638fd1da4c27e3c0586), [#27328](https://github.com/open-webui/open-webui/discussions/27328), [Docs:#1331](https://github.com/open-webui/docs/pull/1331)
+- 🎚️ **Passthrough parameters per connection.** Administrators can now list request parameters that a connection should receive untranslated, under a new Advanced section in connection settings, so provider-specific options reach the upstream API intact. [Commit](https://github.com/open-webui/open-webui/commit/bb12b1a18b77d80829cedb2d5bf965808222415b)
+- 🅰️ **Anthropic requests passed straight through.** Requests to the Anthropic-compatible API aimed at an Anthropic or LiteLLM connection now reach the provider untouched rather than being translated on the way, and LiteLLM is selectable as a connection type. [Commit](https://github.com/open-webui/open-webui/commit/b81627b2c95aad184a6abf59145ca3b18a32bb2c)
+- 💭 **Reasoning in Anthropic responses.** Responses from the Anthropic-compatible API now carry the model's reasoning as thinking blocks, in both streamed and complete responses. [Commit](https://github.com/open-webui/open-webui/commit/bb12b1a18b77d80829cedb2d5bf965808222415b)
+- 🧩 **Structured output through the Anthropic-compatible API.** Requests can now ask for a JSON schema or JSON object response and set a reasoning effort, which are carried through to the upstream model. [Commit](https://github.com/open-webui/open-webui/commit/bb12b1a18b77d80829cedb2d5bf965808222415b)
+- 🪧 **Group names in forwarded headers.** Custom headers on a connection can now carry the groups a person belongs to, by name or by id, so an upstream service or gateway can apply its own rules per group. [#27236](https://github.com/open-webui/open-webui/pull/27236), [#26834](https://github.com/open-webui/open-webui/issues/26834)
+- 🪪 **User identity forwarded to Mistral OCR.** Document extraction through Mistral OCR now carries the requesting user's identity when user info forwarding is enabled, so a gateway in front of it can attribute requests per user like other outbound integrations already do. [#27253](https://github.com/open-webui/open-webui/pull/27253), [#27250](https://github.com/open-webui/open-webui/issues/27250)
+- 🔢 **Anthropic token-counting endpoint.** The Anthropic-compatible API now offers a token-counting endpoint, so integrations can check how many input tokens a request will use before sending it. [Commit](https://github.com/open-webui/open-webui/commit/08dacd19da1b0eefd9d274d24ed59ec1e5d5a2de), [Commit](https://github.com/open-webui/open-webui/commit/23062e9fcaace42cf06db33f9533127bbbcd33d9)
+- 🖲️ **Terminal instructions read fresh.** The instructions a terminal server provides are now fetched for each request, so changing them on the server takes effect immediately instead of after re-saving the connection or restarting. [#27242](https://github.com/open-webui/open-webui/pull/27242)
+- 🖥️ **Live terminal server policies.** Administrators can now read an orchestrator terminal server's current policy and lifecycle settings directly in connection settings rather than relying on a locally cached copy. [Commit](https://github.com/open-webui/open-webui/commit/2f37e853d1259a901f736a823bad29dcc2c3b130), [Commit](https://github.com/open-webui/open-webui/commit/3005b7bc71fcbd5abc6e73c3e4caa4ea781cdb76)
+- 🌍 **Model privacy at a glance.** Admins can now make a model public or private straight from its menu in the model list, where each model is marked as public, shared, or private. [Commit](https://github.com/open-webui/open-webui/commit/fb2ea272952ed96b0db5ac59a861637204100cb6)
+- 📈 **Personal usage dashboard.** A new Usage tab in settings shows your own activity over time, including a token-activity heatmap, current and longest streaks, lifetime and peak token counts, your longest active chat, and your most used models and tools. [Commit](https://github.com/open-webui/open-webui/commit/af9a315ac30b83241f3df5556d7a2abfcd5d25b0)
+- 🧠 **Memories in settings.** Your memories are now listed directly in personalization settings where you can search, add, edit, and remove them, instead of being tucked behind a separate manage dialog. [Commit](https://github.com/open-webui/open-webui/commit/db934a3b4ff16532b48670d7cc75e7048a6db993)
+- 💾 **Import notes and automations.** Notes can now be brought in from text and markdown files, and automations can be exported and imported as files, so you can move them between instances. [Commit](https://github.com/open-webui/open-webui/commit/f8350360dfd60ff890b73fe2f39aaf20a52ad28b), [Commit](https://github.com/open-webui/open-webui/commit/2018546a7baeb853bb4d98e2fb2a092ef20aa431)
+- 🧮 **Counts in the tabs.** The workspace tabs now show how many models, knowledge bases, prompts, skills, and tools you have, and the admin tabs do the same for users, groups, leaderboard entries, and feedback, so you can see the size of each section without opening it. [Commit](https://github.com/open-webui/open-webui/commit/05e3f713175c1eea43a99f29521eb01700c21d3c), [Commit](https://github.com/open-webui/open-webui/commit/727041da78bcfcb44c0e4c83c4e6a641b7061b90), [Commit](https://github.com/open-webui/open-webui/commit/f8ea15b84a274712dca33daa970f63ed7368043e)
+- 🧾 **Group permissions at a glance.** The groups list now shows whether each group uses custom or default permissions, without opening it. [Commit](https://github.com/open-webui/open-webui/commit/ccb1ab7739fbeb77c810036dcac240570034a56a), [Commit](https://github.com/open-webui/open-webui/commit/1d1f60ab440b167b9c6ab8f4011b884caa99f27d)
+- 📤 **Streamed file transfers.** Uploading a model, pipeline or audio file now sends it in chunks instead of holding the whole thing in memory, and reading and writing files no longer blocks other requests, so large transfers no longer spike memory or stall the server. [Commit](https://github.com/open-webui/open-webui/commit/429f2df50cd2f0ec8d0a1bb4136a46e2f94a4bf5), [#27351](https://github.com/open-webui/open-webui/pull/27351), [#27349](https://github.com/open-webui/open-webui/issues/27349)
+- 🧰 **Built-in tool descriptions built once.** The descriptions handed to the model for the built-in tools are now worked out once at startup rather than rebuilt on every message. [Commit](https://github.com/open-webui/open-webui/commit/d727ee4d1febb2b72d5f6c26668c562eadca54f1), [Commit](https://github.com/open-webui/open-webui/commit/12974c9e4ed97b2d68c4ad129c057ff0e774254e), [#27374](https://github.com/open-webui/open-webui/pull/27374), [#27396](https://github.com/open-webui/open-webui/pull/27396)
+- 🪺 **Records read without a double pass.** Loading a model, tool, prompt, skill, note, knowledge base, channel or calendar no longer converts the record twice on the way out. [Commit](https://github.com/open-webui/open-webui/commit/f1409266feb224e74fe023d7459f1e2b5aad0b29), [#27377](https://github.com/open-webui/open-webui/pull/27377)
+- 🔧 **Faster tool and knowledge base listings.** Listing tools no longer loads each one's full source, and working out which tools and knowledge bases you can see takes a single check rather than one per item. [#27387](https://github.com/open-webui/open-webui/pull/27387)
+- 🧊 **Quicker collection checks on Chroma.** Checking whether a collection exists now asks for that one collection instead of listing them all, which grew slower with every knowledge base and file. [Commit](https://github.com/open-webui/open-webui/commit/48ee357156fd7f567ea13eb5ddba0a25701c0351), [#27394](https://github.com/open-webui/open-webui/pull/27394)
+- 🔠 **Tokenizer loaded once.** The tokenizer used to split documents is now kept after first use rather than being loaded again for every file. [Commit](https://github.com/open-webui/open-webui/commit/7e31f64bc81b2264136a85efc7bb86c00f346784), [#27394](https://github.com/open-webui/open-webui/pull/27394)
+- 📗 **Faster knowledge base file lists.** Opening a knowledge base now loads just the file names and details instead of the entire extracted text of every document, so large collections appear almost instantly. [#27386](https://github.com/open-webui/open-webui/pull/27386), [#26144](https://github.com/open-webui/open-webui/issues/26144)
+- 🗝️ **Faster file access checks.** Working out whether you can open a file no longer scales with how many workspace models and knowledge bases exist, so opening files and listing folder contents stays quick on large instances. [#27383](https://github.com/open-webui/open-webui/pull/27383)
+- 🕰️ **Faster automation scheduling.** Working out when an automation that repeats every few minutes or hours runs next is now near instant, instead of taking twenty seconds or more and slowing further each year. [Commit](https://github.com/open-webui/open-webui/commit/b3aead23da6cf8ebeedbd9fa3b97c7ac1a3f54ec), [#26954](https://github.com/open-webui/open-webui/issues/26954)
+- 📁 **Faster folder loading.** Your folder list no longer runs a separate lookup for every folder to check where it sits, so it loads in a single pass. [Commit](https://github.com/open-webui/open-webui/commit/9a49b271aaf5d6eeaec24ab974be38a8c68ddd76)
+- 🎯 **One round of requests per folder click.** Selecting a folder in the sidebar now fetches the folder, the folder tree, and each expanded folder's chats once instead of two to four times. [#27540](https://github.com/open-webui/open-webui/pull/27540), [#27539](https://github.com/open-webui/open-webui/issues/27539)
+- 🎧 **No wasted work when nobody is listening.** Updates for a chat whose tab has been closed, or for requests made through the API, are no longer packaged up only to be discarded, which matters most on long streamed replies. [#27366](https://github.com/open-webui/open-webui/pull/27366), [Commit](https://github.com/open-webui/open-webui/commit/858e9236df1c3d84782e373c22b56cfc312b6db8)
+- 📑 **Cheaper audit logging.** With audit logging on, each request is no longer authenticated a second time just to record the log entry, so audited instances carry noticeably less overhead. [#27373](https://github.com/open-webui/open-webui/pull/27373)
+- 🪧 **Cheaper tagging after each reply.** Saving the tags generated for a conversation now updates just that field instead of loading, rewriting and re-reading the whole conversation, which cost more the longer the chat. [#27382](https://github.com/open-webui/open-webui/pull/27382)
+- ✍️ **Faster saves across the app.** Saving a chat, note, prompt, tool or user setting no longer re-reads the record it just wrote, so writes finish sooner, most noticeably on long conversations. [#27381](https://github.com/open-webui/open-webui/pull/27381), [#27379](https://github.com/open-webui/open-webui/pull/27379), [Commit](https://github.com/open-webui/open-webui/commit/c182a95ffdb87bae3d47d93387ec1a26c97740a2), [Commit](https://github.com/open-webui/open-webui/commit/977c7930623b860949410a73922579daf81705a7)
+- 🛢️ **Less database overhead per request.** SQLite installations no longer run a connection check before every database call, and requests that never touch the database skip the bookkeeping that used to run regardless. [#27385](https://github.com/open-webui/open-webui/pull/27385)
+- ⚡ **Faster memory lookups.** Stored memories are now indexed so retrieving them stays quick as the number you have grows. [Commit](https://github.com/open-webui/open-webui/commit/28bdcb063b8d5d6a0b10943b1b2f87b16ff63621), [#26957](https://github.com/open-webui/open-webui/pull/26957)
+- 🪪 **Fewer checks before a reply starts.** Working out whether you may use a model now looks up the model and your group memberships once instead of repeating both, including for every model a workspace model is built on. [#27378](https://github.com/open-webui/open-webui/pull/27378)
+- 👤 **Lighter user activity checks.** Checking whether someone is currently active now reads only that timestamp rather than their whole profile, including their profile image. [Commit](https://github.com/open-webui/open-webui/commit/c8f2e09fdcafc800c1e3af6da1cd6f2581cd9191), [#27224](https://github.com/open-webui/open-webui/pull/27224)
+- 📨 **Fewer settings lookups when sending a message.** Sending a chat message now reads the settings behind tools, file retrieval, voice, skills and the code interpreter in fewer trips to the database, so replies start sooner. [#27223](https://github.com/open-webui/open-webui/pull/27223)
+- 🪄 **Lighter conversion for Ollama requests.** Preparing a request for an Ollama model no longer copies the entire conversation before sending it, which cost more with every message and repeated on each tool-call round. [#27371](https://github.com/open-webui/open-webui/pull/27371)
+- 🦙 **Fewer settings lookups on Ollama requests.** Ollama chat, generation and embedding requests now read their connection settings once instead of up to four times, so each request reaches the server sooner. [#27226](https://github.com/open-webui/open-webui/pull/27226)
+- 🧹 **Less repeated work on every response.** Security headers are now worked out once at startup rather than rebuilt for each response, and ordinary page requests skip the redirect handling they never needed, so responses carry less overhead. [#27229](https://github.com/open-webui/open-webui/pull/27229)
+- 🚀 **Lower per-request overhead.** Requests no longer each perform a settings lookup before they are handled, trimming a little latency from everything the app does. [Commit](https://github.com/open-webui/open-webui/commit/4493b56e424db29fa9e72310b1ca6b025c3e5f8b), [#27395](https://github.com/open-webui/open-webui/pull/27395), [Commit](https://github.com/open-webui/open-webui/commit/6ff1df326c76824f0706671b0974df4035cb453f), [Commit](https://github.com/open-webui/open-webui/commit/85664f650cc111a6b170b97ed0c391d962717bec), [#27227](https://github.com/open-webui/open-webui/pull/27227)
+- 💨 **Leaner filter handling while streaming.** Filters applied to a streaming reply no longer re-read their settings and each plugin's full source from the database for every chunk, so responses with filters enabled cost the server far less work. [#27228](https://github.com/open-webui/open-webui/pull/27228), [#27372](https://github.com/open-webui/open-webui/pull/27372), [Commit](https://github.com/open-webui/open-webui/commit/f9107edeebc7ee545d7e3c1f1b7d449c123ab398), [Commit](https://github.com/open-webui/open-webui/commit/f578d8d67ec2c109b0d8c38d90eaeb4448f83610), [Commit](https://github.com/open-webui/open-webui/commit/9acbe3aa0f258a3bda593bb98ec81c6acc458d20), [#27392](https://github.com/open-webui/open-webui/pull/27392)
+- 🚦 **No filter bookkeeping without filters.** Streamed API responses only build up the full reply for outlet filters when the model actually has one configured, instead of doing it for every request. [Commit](https://github.com/open-webui/open-webui/commit/315a6b5995663eabe1c96776d66b6593860d47d6), [#27391](https://github.com/open-webui/open-webui/pull/27391)
+- ✂️ **Cheaper tag detection while streaming.** Watching a reply for reasoning and code blocks now examines only the newly arrived text rather than rescanning the whole answer on every chunk, so a long answer no longer costs progressively more as it grows. [#27360](https://github.com/open-webui/open-webui/pull/27360)
+- 🌊 **Steadier long responses.** Building up a streamed reply no longer costs more work as it grows, so long answers keep pace instead of slowing down toward the end. [#27231](https://github.com/open-webui/open-webui/pull/27231), [#27359](https://github.com/open-webui/open-webui/pull/27359), [Commit](https://github.com/open-webui/open-webui/commit/ba556bd8f0517881cb250bb512631ddf8a0c82c3), [#27390](https://github.com/open-webui/open-webui/pull/27390)
+- 📦 **Faster JSON handling as an option.** Administrators can now switch the whole application to a faster encoder through a new "ENABLE_ORJSON" setting, covering request bodies, responses, upstream provider payloads and live updates, where the encoding of live updates was the largest single cost on the workers handling them in clustered deployments; it stays off by default because the faster encoder is stricter about what it accepts. [#27583](https://github.com/open-webui/open-webui/pull/27583)
+- ⚙️ **Faster Redis handling.** The compiled "hiredis" parser now ships as a dependency and is used automatically, so deployments backed by Redis spend noticeably less processor time reading responses. [#27282](https://github.com/open-webui/open-webui/pull/27282)
+- 🔗 **Fewer Redis round trips per chat.** Deployments backed by Redis now look up the model and connected sessions once per request instead of twice, and fetch the model list in a single call. [#27225](https://github.com/open-webui/open-webui/pull/27225)
+- 🛰️ **Fewer Sentinel lookups.** Redis Sentinel deployments no longer ask which server is the primary and open a fresh connection before every single command, which had caused heavy connection churn and stalls under load. [Commit](https://github.com/open-webui/open-webui/commit/75a8a0046b5b2ebd9942b25035b346aa953f81cc), [#27213](https://github.com/open-webui/open-webui/pull/27213), [#27210](https://github.com/open-webui/open-webui/issues/27210)
+- 📡 **Lighter live connection handling.** Typing indicators, shared document edits and reconnections no longer re-read your account or copy the full participant list each time, and idle sessions are no longer rewritten every few seconds. [Commit](https://github.com/open-webui/open-webui/commit/021c4c7a2e8b5b213f49800ecd331b1b18c2ef99), [#27393](https://github.com/open-webui/open-webui/pull/27393)
+- 🏎️ **Faster chat search on PostgreSQL.** Searching chats on PostgreSQL now reads from the message table instead of unpacking each conversation's stored data row by row, so results stay quick as your history grows. [Commit](https://github.com/open-webui/open-webui/commit/cc9a44569ef08b64ff44d15607c43966f362ce75), [#27221](https://github.com/open-webui/open-webui/issues/27221)
+- ⚡ **Lighter model lists.** Model lists no longer carry embedded profile images in their data, so they load faster. [Commit](https://github.com/open-webui/open-webui/commit/9281adc5647b7046e3ddcc53ac4b84be7f650221), [Commit](https://github.com/open-webui/open-webui/commit/f3a35507845e4a911c3d278d680a9989bb8d99ad)
+- 🏗️ **Fewer queries when building the model list.** Assembling the model list now makes fewer database round trips and no longer fetches every plugin's source code along the way, so it comes together faster. [Commit](https://github.com/open-webui/open-webui/commit/6b655689ccbdf2111620d005a8e6dedb8fb673f8), [#27389](https://github.com/open-webui/open-webui/pull/27389)
+- 🪶 **Model lists without knowledge text.** Model lists no longer include the extracted text of files attached to a model as knowledge, so they stay small regardless of how large those knowledge bases are. [Commit](https://github.com/open-webui/open-webui/commit/48625e657ff11161c3588af2747598f102c1a4d1), [#27287](https://github.com/open-webui/open-webui/issues/27287)
+- 🔛 **Functions can react to being switched on or off.** Two new events fire just before a function is enabled or disabled, and the function being enabled receives its own event even though it is not active yet, so it can run whatever setup or teardown it needs. [Commit](https://github.com/open-webui/open-webui/commit/94a60b04573acf6423e9c0519997b779f82e0560), [#26754](https://github.com/open-webui/open-webui/pull/26754), [#26748](https://github.com/open-webui/open-webui/discussions/26748)
+- 🔛 **Multiple choice settings in plugins.** A tool or function can now offer a setting where you tick several options from a list, fixed or worked out at the time it is shown, instead of asking you to type a comma-separated list of allowed values. [#26884](https://github.com/open-webui/open-webui/pull/26884), [#26848](https://github.com/open-webui/open-webui/issues/26848)
+- 🔌 **Disable plugins entirely.** Administrators can now completely turn off the built-in Tools and Functions plugin surfaces through a new "ENABLE_PLUGINS" setting, which hides them across the workspace and admin areas and removes their execution paths. [Commit](https://github.com/open-webui/open-webui/commit/bd6e0b61c2ae073aba9556ae46c345f4749acb84), [Commit](https://github.com/open-webui/open-webui/commit/8e46450acd7ae11a4dee166d19a7c9833d991e79), [Commit](https://github.com/open-webui/open-webui/commit/951f96021a970fbd4837a4ee441565c0cf3d2824), [Commit](https://github.com/open-webui/open-webui/commit/252e6fd855099e1c880f4def18aa09aedbe1733a)
+- 🧵 **Lighter chat listings and search.** Building a page of chat search results or a folder listing no longer copies each full conversation to read its title and dates, so those pages come together faster and use far less memory while they are built. [#27388](https://github.com/open-webui/open-webui/pull/27388)
+- 📮 **Name lookups off the thread pool.** Looking up a hostname no longer occupies one of the limited threads shared by every other piece of blocking work, so model calls, searches, page fetches and tool calls stop queueing behind each other once a few lookups are slow. [#27440](https://github.com/open-webui/open-webui/pull/27440)
+- 🥬 **Faster web page parsing.** Pages pulled in by web search and web retrieval are now read with a faster parser, cutting roughly a tenth off the time spent on a ten result search. [#27439](https://github.com/open-webui/open-webui/pull/27439)
+- 🧭 **No pointless lookups when filtering search results.** Filtering web search results against a domain list no longer resolves every result to an address first, which had turned a three second search into half a minute wherever the resolver was slow or a name did not resolve. [Commit](https://github.com/open-webui/open-webui/commit/42ea8a5a2f04b6a57ccb47a61611a62479b60b78), [#26920](https://github.com/open-webui/open-webui/issues/26920)
+- 🚄 **Leaner passthrough streaming.** Responses the server only relays now go straight through in whole network reads instead of being split line by line, roughly halving the work spent shuttling a streamed reply on those routes. [#27384](https://github.com/open-webui/open-webui/pull/27384)
+- 🧶 **Web page parsing off the critical path.** Reading those pages no longer holds up everything else on the server, so other people's replies, live updates and health checks keep flowing during a search instead of stalling for a second or more. [#27446](https://github.com/open-webui/open-webui/pull/27446)
+- 🈶 **Faster uploads of non-English text files.** Working out the encoding of an uploaded text file now samples the part that needs it rather than scanning the whole file, taking a four megabyte Japanese or Chinese document from several seconds down to well under one. [#27445](https://github.com/open-webui/open-webui/pull/27445)
+- ♿ **Improved UI accessibility.** Keyboard and screen reader users can now tell which chat in the sidebar is the one being viewed, open reasoning and detail blocks in a response, expand sidebar sections and open a folder without a mouse, sort the admin user list from the keyboard and hear which column it is sorted by, open a dropdown and its submenus with the keyboard, close them again with Escape and land back where they started, hear which value a dropdown is set to rather than only its label, hear what each admin settings switch, group permission toggle, checkbox, API key field and advanced model parameter slider controls, have the message box announced by its placeholder instead of as an unnamed field, press Enter on Cancel in a confirmation dialog without triggering the delete, reach the regenerate control, jump straight past the sidebar to the conversation with a skip link, hear what an icon-only button does across chat, calls, file previews, modals and the admin pages rather than an unlabelled button, placeholder text, section headings, field descriptions, inactive tab labels, timestamps, counters and icons are now readable against their background when High Contrast Mode is on, and sidebar buttons across notes, automations, the playground, and admin pages announce whether they open or close the sidebar. [#27510](https://github.com/open-webui/open-webui/pull/27510), [#27513](https://github.com/open-webui/open-webui/pull/27513), [#27503](https://github.com/open-webui/open-webui/pull/27503), [#27494](https://github.com/open-webui/open-webui/pull/27494), [#27491](https://github.com/open-webui/open-webui/pull/27491), [#27490](https://github.com/open-webui/open-webui/pull/27490), [#27489](https://github.com/open-webui/open-webui/pull/27489), [#27488](https://github.com/open-webui/open-webui/pull/27488), [#27555](https://github.com/open-webui/open-webui/pull/27555), [#27556](https://github.com/open-webui/open-webui/pull/27556), [#27554](https://github.com/open-webui/open-webui/pull/27554), [#27558](https://github.com/open-webui/open-webui/pull/27558), [#27501](https://github.com/open-webui/open-webui/pull/27501), [#27492](https://github.com/open-webui/open-webui/pull/27492), [#27509](https://github.com/open-webui/open-webui/pull/27509), [#27502](https://github.com/open-webui/open-webui/pull/27502), [#26769](https://github.com/open-webui/open-webui/pull/26769), [Commit](https://github.com/open-webui/open-webui/commit/89caa7c849c471561dfd76140d0c3c3ce7a68df8), [Commit](https://github.com/open-webui/open-webui/commit/7801909d27b18331a9a2bc399e1d618ab99ba5bf), [#26768](https://github.com/open-webui/open-webui/pull/26768), [#26770](https://github.com/open-webui/open-webui/pull/26770), [Commit](https://github.com/open-webui/open-webui/commit/421834b2de287b8d5291d4b695ba6ed4528a5e7f), [Commit](https://github.com/open-webui/open-webui/commit/7bfc4bb2c25249d3922acd54d5bc516db52c8519), [Commit](https://github.com/open-webui/open-webui/commit/e8fda1c7a07d1a0f91201ff977b77fd5a43e014a), [#27508](https://github.com/open-webui/open-webui/pull/27508)
+- 🔄 **General improvements.** Various improvements were implemented across the application to enhance performance, stability, and security.
+- 🌐 **Translation updates.** Slovenian is now available, and translations for English (UK), Finnish, German, Japanese, Portuguese (Brazil) and Portuguese (Portugal) were enhanced and expanded.
+
+### Fixed
+
+- 🛡️ **Security Advisory**: This release includes security and access-control fixes. We recommend updating production deployments at your earliest convenience. Not all security fixes in this version may be enumerated in the fixed section. Some may be withheld for a short time to give administrators time to upgrade. [Advisories](https://github.com/open-webui/open-webui/security)
+- 🔒 **Terminal file preview isolation.** Previewing an HTML file in the system terminal now runs it in an isolated context by default, closing a cross-site scripting hole that could expose your login session or, for privileged accounts, run code on the server. [#26907](https://github.com/open-webui/open-webui/pull/26907)
+- ➗ **Malformed maths in a message.** Maths that fails to render is now shown as plain text rather than being placed into the page as markup, closing a way for a crafted formula in a chat, channel or shared conversation to run code in the browser of anyone reading it. [#26718](https://github.com/open-webui/open-webui/pull/26718)
+- 🔩 **Updated file upload parsing library.** The library that parses file uploads and form submissions has been updated to a release that addresses a security advisory affecting that parsing path. [#26991](https://github.com/open-webui/open-webui/pull/26991)
+- 🛑 **Deactivated accounts lose live access.** Real-time connections now apply the same role check as the rest of the application, so an account moved out of the user or admin role can no longer keep its channels and shared notes open on an existing token. [#27537](https://github.com/open-webui/open-webui/pull/27537)
+- 🛅 **Writing into someone else's chat.** Completion and action requests now confirm you own the chat they name before anything is written to it, so a filter or action can no longer be pointed at another person's conversation. [#27486](https://github.com/open-webui/open-webui/pull/27486)
+- 🎟️ **Ollama version no longer readable anonymously.** Reading the configured Ollama backend's version now requires signing in, closing a route that let anyone learn the version in use and count how many backends are configured. [#27199](https://github.com/open-webui/open-webui/pull/27199)
+- 🔐 **Folder sharing permission.** The folder sharing setting in default and group permissions now saves instead of being silently discarded, so allowing or restricting folder sharing actually takes effect. [#27296](https://github.com/open-webui/open-webui/pull/27296), [#27120](https://github.com/open-webui/open-webui/issues/27120)
+- 🔕 **Webhook permission enforcement.** People without permission to use webhooks can no longer save webhook notification destinations to their settings, so the permission is enforced when settings are saved rather than only reflected in the interface. [#27297](https://github.com/open-webui/open-webui/pull/27297), [Commit](https://github.com/open-webui/open-webui/commit/af629177f46fa4595175f914c971c47701d1a676)
+- 🛎️ **Stopping someone else's generation.** Deleting a chat now checks who you are before anything is cancelled, so knowing another person's chat id no longer lets you cut off their reply or title generation on a request that is refused anyway. [#27006](https://github.com/open-webui/open-webui/pull/27006)
+- 🚥 **Automation limits in chat.** Automations that the assistant creates or reschedules on your behalf now respect the same maximum count and minimum interval as the ones you set up yourself, instead of being able to exceed both. [#27523](https://github.com/open-webui/open-webui/pull/27523), [#27121](https://github.com/open-webui/open-webui/issues/27121)
+- ⏲️ **Cancelling someone else's timers.** Marking a chat as read now only clears your own pending timers on it, instead of clearing everyone's, which had let another person's scheduled prompt be silently cancelled without them being told. [#27472](https://github.com/open-webui/open-webui/pull/27472)
+- 🗑️ **Deleting a shared folder's subfolders.** Deleting a folder is now limited to its owner or an administrator at every level, so someone with write access to a shared folder can no longer delete a subfolder and take the owner's chats with it. [#27003](https://github.com/open-webui/open-webui/pull/27003)
+- 📕 **Tool source shown to people who can only use it.** Opening a tool you were given read access to no longer returns its source code, which read access was never meant to include. [#27005](https://github.com/open-webui/open-webui/pull/27005)
+- 🎯 **Model settings in the list endpoint.** Listing models no longer includes each one's parameters and system prompt for people with read access only, matching what opening a single model already returned. [#27004](https://github.com/open-webui/open-webui/pull/27004)
+- 🖌️ **Image generation and web search without permission.** Turning on image generation or web search through the older request format now checks your permission first, so someone denied those features can no longer trigger them, and the billing that comes with them, by asking for that format. [#26703](https://github.com/open-webui/open-webui/pull/26703)
+- 🎗️ **Terminal single sign-on tokens.** The token forwarded to a terminal server for single sign-on is now taken from your own session on the server rather than from a header the browser supplied, so a caller can no longer send someone else's token in its place. [#26719](https://github.com/open-webui/open-webui/pull/26719)
+- 🫗 **Web search results scoped to you.** The temporary collections holding a web search's pages are now tied to the person who ran the search, closing the one place where that scoping was not applied. [#26706](https://github.com/open-webui/open-webui/pull/26706)
+- 🧺 **Knowledge base cleanup reaching other collections.** Tidying up a knowledge base now acts only on files and folders that belong to it, so someone with write access to one knowledge base can no longer delete folders or search data belonging to another. [#26722](https://github.com/open-webui/open-webui/pull/26722)
+- ⌛ **Searches that could stall the server.** A search pattern inside knowledge base commands now runs under a time budget, so a pattern that would take minutes to evaluate can no longer hold up everyone else on the instance. [#27471](https://github.com/open-webui/open-webui/pull/27471)
+- 🚫 **Disabled terminal servers are refused.** A terminal connection an administrator has turned off can no longer be reached by browsing its files, opening a session, or calling its tools, rather than only disappearing from the interface. [Commit](https://github.com/open-webui/open-webui/commit/7537989235675ac84a40bf70c91ec3e16fc0d8db)
+- 🧫 **Files attached to a shared folder.** Adding files to a folder is now refused unless the folder's owner can read them, and a folder's files are checked against what its owner can still read before they are used as knowledge in chat, so a collaborator can no longer place files into someone else's folder or keep serving files the owner has since lost access to. [#27464](https://github.com/open-webui/open-webui/pull/27464), [Commit](https://github.com/open-webui/open-webui/commit/56183fcb17142088e2a34d1e35228f013749030c)
+- 🧷 **Knowledge claimed by a direct connection.** Files listed as knowledge on a model supplied by the browser for a direct connection are now filtered against your own access before anything is retrieved, so a crafted request can no longer pull in documents you cannot otherwise open. [Commit](https://github.com/open-webui/open-webui/commit/305880f2e2aeb2dda2f4b2a18a20bdcd558f7134), [#26723](https://github.com/open-webui/open-webui/pull/26723)
+- 🪜 **Reaching a restricted model through a shared one.** A shared workspace model can no longer be used to reach an underlying model the person could not otherwise use, which previously slipped through when that model had no entry of its own. [#26905](https://github.com/open-webui/open-webui/pull/26905), [#26900](https://github.com/open-webui/open-webui/issues/26900)
+- 🖌️ **Shared image checkpoint changes.** Only administrators can now change the instance-wide Automatic1111 checkpoint, so an ordinary image generation request no longer switches the image model for everyone. [#27244](https://github.com/open-webui/open-webui/pull/27244)
+- 💬 **Channel message ownership.** Only the author of a channel message, or an administrator, can now edit or delete it, instead of anyone able to post in that channel. [#27197](https://github.com/open-webui/open-webui/pull/27197)
+- 🗄️ **Chats shared with an administrator.** An administrator can now open a chat that was deliberately shared with them even when broad admin access to other people's chats is turned off, instead of being refused a chat any other recipient could read. [#27127](https://github.com/open-webui/open-webui/pull/27127)
+- 📓 **Notes in folder knowledge are access checked.** Notes attached to a folder are now filtered against your own access before the list reaches the assistant, rather than relying on later checks further along. [#26739](https://github.com/open-webui/open-webui/pull/26739)
+- 🧱 **Code interpreter module blocking.** Modules an administrator has blocked for the code interpreter are now actually blocked, and other imports inside interpreter code work again. [#27245](https://github.com/open-webui/open-webui/pull/27245)
+- 📉 **Charts in the code interpreter.** Code that draws a chart now runs in the default code interpreter setup, instead of failing with a syntax error unless file persistence was turned on. [#26800](https://github.com/open-webui/open-webui/pull/26800), [#26660](https://github.com/open-webui/open-webui/issues/26660)
+- 🎬 **Chat action availability.** Chat actions can no longer be triggered when they are disabled, not assigned to the model in use, or on a model the caller cannot access, matching the actions the interface actually offers. [#27243](https://github.com/open-webui/open-webui/pull/27243)
+- 🗨️ **Response text where it was missing.** Assistant replies are no longer stored without their text, so copying, exporting, searching and reusing a conversation return the reply instead of nothing. [Commit](https://github.com/open-webui/open-webui/commit/33cf3fbb7f017ab1b79dce5c5ca4d4e1c3092844), [#26799](https://github.com/open-webui/open-webui/pull/26799), [#26436](https://github.com/open-webui/open-webui/issues/26436)
+- 🧪 **Filter edits that survive a reload.** A change a filter makes to a finished response is now saved with the conversation, instead of showing on screen and reverting the next time the chat is opened. [#27414](https://github.com/open-webui/open-webui/pull/27414), [#27017](https://github.com/open-webui/open-webui/issues/27017)
+- 📃 **Action functions receive the response text.** Running an action on a response now passes the assistant's text to the function, instead of handing it an empty message. [#26798](https://github.com/open-webui/open-webui/pull/26798), [#26672](https://github.com/open-webui/open-webui/issues/26672)
+- 🍎 **Blank messages on Safari.** Assistant responses no longer render as empty in Safari and on iPhone and iPad, where a browser painting bug left on-screen messages unpainted. [#26805](https://github.com/open-webui/open-webui/pull/26805), [#26712](https://github.com/open-webui/open-webui/issues/26712), [#26844](https://github.com/open-webui/open-webui/issues/26844)
+- ➡️ **Prompts opened from a link.** A prompt passed in through a link that sends automatically now waits for tool servers to finish loading, so external tools are available on that first message instead of the model reporting it has none. [Commit](https://github.com/open-webui/open-webui/commit/d7513e4ce81ada1936c0c34115f947e115d6f2cf), [#24176](https://github.com/open-webui/open-webui/issues/24176)
+- 🪟 **Tool result prompt submission.** Interactive tool result embeds that send a prompt back to the chat work again, showing the confirmation dialog before submitting instead of silently doing nothing. [#26914](https://github.com/open-webui/open-webui/pull/26914), [#26912](https://github.com/open-webui/open-webui/issues/26912)
+- 📻 **Live updates in a second tab.** Opening Open WebUI again while already connected now joins the new tab to your event stream, so notifications and chat updates reach every open tab instead of only the first one. [Commit](https://github.com/open-webui/open-webui/commit/d14fddf25405cd58184fdef3d2af012503e4edd8)
+- 🔁 **Connection recovery on new chats.** Chats started from the home page now recover automatically after a dropped connection, such as from mobile backgrounding, a VPN or IP change, or waking from sleep, instead of getting stuck loading until a manual refresh. [#26913](https://github.com/open-webui/open-webui/pull/26913), [#26844](https://github.com/open-webui/open-webui/issues/26844)
+- 🪫 **Terminal choice cleared on load.** Your selected terminal is no longer dropped while the list of terminals is still loading, so it survives a page refresh. [Commit](https://github.com/open-webui/open-webui/commit/9707d3a5c21d2fedada602f2cfd8a104cdc5e5d1), [Commit](https://github.com/open-webui/open-webui/commit/f59d86a10cb31e48d2dae0033081a09dbbaafc8c), [Commit](https://github.com/open-webui/open-webui/commit/2f2bf38e3481077597b5ad60f57685813e8d71b8), [#26775](https://github.com/open-webui/open-webui/pull/26775), [#26677](https://github.com/open-webui/open-webui/issues/26677)
+- 🔌 **Dropped sessions during keepalive.** Live connections no longer break on a routine keepalive check, which had cut the session so that anything the server needed to run in your browser failed afterwards, most visibly the code execution tool reporting the client as disconnected on every run. [#27553](https://github.com/open-webui/open-webui/pull/27553), [#27550](https://github.com/open-webui/open-webui/issues/27550)
+- ✂️ **Context compaction turn boundaries.** Long-conversation compaction now summarizes only completed earlier turns instead of sometimes cutting through the middle of a single turn, keeping the current turn's tool calls and results intact. [#27035](https://github.com/open-webui/open-webui/issues/27035), [Commit](https://github.com/open-webui/open-webui/commit/959558fd82eb2a3c980231acd500b73ba4b698b3), [Commit](https://github.com/open-webui/open-webui/commit/17e6496538e5f3147203b7520012a985cab044b7)
+- 🪆 **Summaries on a direct connection.** Summarizing a long conversation on a direct connection can now use the configured summary model rather than being limited to the connection's own model. [#26806](https://github.com/open-webui/open-webui/pull/26806)
+- 🪟 **System prompt through compaction.** The system message now stays at the front of the conversation when a long chat is summarized, instead of being folded into the summary and lost from that point on. [Commit](https://github.com/open-webui/open-webui/commit/70549c5c8a50315aa3bf909ebedea8cc3e602077), [Commit](https://github.com/open-webui/open-webui/commit/15688686af9dd73ec974e35f96a7ea24294dbe4f), [Commit](https://github.com/open-webui/open-webui/commit/44f4f9dce48f1ad2af0c5f3210fa5aa701bad624), [#26713](https://github.com/open-webui/open-webui/pull/26713), [#26710](https://github.com/open-webui/open-webui/issues/26710)
+- 🧷 **Context compaction continuity.** After a compaction, the retained recent messages now stay in the prompt on every following turn instead of disappearing after the first, preserving conversational continuity and prompt caching. [#27037](https://github.com/open-webui/open-webui/issues/27037), [Commit](https://github.com/open-webui/open-webui/commit/0c23466a3e9a1fb7d32875a0614f1ca8e583bc73), [Commit](https://github.com/open-webui/open-webui/commit/f730733bc44eff5812eff0e51ebca0bbcfa1bc6e)
+- 🔟 **Context size after tool calls.** The context meter and long-conversation compaction now read the size of the latest request rather than adding up every call in a tool loop, and understand the counts reported by Ollama and llama.cpp as well as the OpenAI-style ones, so compaction no longer fires far below its threshold, or never at all, and the usage shown is no longer inflated. [Commit](https://github.com/open-webui/open-webui/commit/df94268e892cbb66675170a6c78846aef23f6e89), [Commit](https://github.com/open-webui/open-webui/commit/e8f2c123e63c9073c9ae4ee00573144ce6d4b2e9), [#27031](https://github.com/open-webui/open-webui/issues/27031), [#26752](https://github.com/open-webui/open-webui/pull/26752), [#24410](https://github.com/open-webui/open-webui/discussions/24410)
+- 💭 **Reasoning that arrives late or empty.** Reasoning sent by a provider after the answer has started is now shown in its proper place above the answer rather than appended after it, and reasoning notes carrying nothing no longer open an empty thinking block. [Commit](https://github.com/open-webui/open-webui/commit/051a1f6c41d1d37e5a12ce2068fc6f8488940591), [#26687](https://github.com/open-webui/open-webui/pull/26687), [#26645](https://github.com/open-webui/open-webui/issues/26645)
+- 📐 **System prompt lost during tool calls.** A model's system prompt now stays in place through every round of tool calls, instead of being dropped after the first one and, with memories enabled, replaced by the memory block alone. [#26857](https://github.com/open-webui/open-webui/pull/26857), [#26836](https://github.com/open-webui/open-webui/issues/26836)
+- 🪶 **Memories from structured replies.** A reply delivered as structured output is now read when memories are reviewed after a turn, so nothing worth remembering is skipped just because of how the answer arrived. [Commit](https://github.com/open-webui/open-webui/commit/3fe03583a3b240da3cc42364088f22bc3a059487), [#26705](https://github.com/open-webui/open-webui/pull/26705), [#26651](https://github.com/open-webui/open-webui/issues/26651)
+- 🎲 **Stable skill ordering.** Skills available to a model are now listed in the same order on every request, instead of shuffling between requests and quietly defeating prompt caching. [Commit](https://github.com/open-webui/open-webui/commit/b9d72741bb2f649cb67942ce7c635ea173f32b70), [#26986](https://github.com/open-webui/open-webui/issues/26986)
+- 🛑 **Stopping an answer the moment it starts.** Each answer in a chat now carries its own task identifier from the first event onward, so stopping one immediately after sending no longer misses. [Commit](https://github.com/open-webui/open-webui/commit/aadab2f480a8c17a9265a244d262947da23ddc79)
+- ⏸️ **Deleting while a reply is being written.** The delete control is now hidden on messages while a response is generating or a task is running, so a conversation can no longer be left with the finished reply detached from the messages before it. [Commit](https://github.com/open-webui/open-webui/commit/b4d13793a3af2d75aaf33fe9793cbbf278958940), [#26668](https://github.com/open-webui/open-webui/issues/26668)
+- 🎁 **Feedback while a download is prepared.** Downloading a file or folder from the terminal now tells you it is being prepared, will not start the same archive twice if you click again, and reports a failure instead of quietly giving up or leaving a preview spinning. [#27421](https://github.com/open-webui/open-webui/pull/27421), [#27055](https://github.com/open-webui/open-webui/issues/27055)
+- 📥 **Moving an archived chat into a folder.** Moving an archived chat into a folder now takes it out of the archive so it appears there, and the folder's contents refresh straight away after a move from the menu. [#27485](https://github.com/open-webui/open-webui/pull/27485), [#27484](https://github.com/open-webui/open-webui/issues/27484)
+- 📜 **Chats past the first sixty in a folder.** Folder listings now page through every chat instead of stopping at a fixed limit, so older chats no longer appear to vanish from a folder once it grows past sixty. [#26786](https://github.com/open-webui/open-webui/issues/26786), [Commit](https://github.com/open-webui/open-webui/commit/409fb39717be9ab7becd9e8c01801a08c5bae318)
+- 📌 **Sidebar highlight follows the open chat.** The sidebar no longer keeps a chat highlighted after you move to another page, so deleting or archiving it there no longer throws you back to a new chat, and cloning no longer leaves two chats looking selected. [#26977](https://github.com/open-webui/open-webui/pull/26977)
+- 🔀 **Sidebar ordering during replies.** Background updates such as follow-up suggestions, sources, and status no longer bump a chat to the top of the sidebar or change its last-updated time, and neither does saving a chat's variables or settings, nor the automatic title generation on a new chat. [Commit](https://github.com/open-webui/open-webui/commit/f1ded9409a5523ec27d99635d8b7e7e1a297a4eb), [Commit](https://github.com/open-webui/open-webui/commit/a9617ca2187920734e2be5f90c5f119c09850ff5)
+- 🖱️ **One hover preview at a time.** Moving between chats in the sidebar, or between avatars in the admin user list, channel messages and member lists, no longer leaves an earlier preview open behind the new one. [#27549](https://github.com/open-webui/open-webui/pull/27549), [#27548](https://github.com/open-webui/open-webui/issues/27548), [#27578](https://github.com/open-webui/open-webui/pull/27578), [#27577](https://github.com/open-webui/open-webui/issues/27577)
+- ✨ **Folder lists no longer flash.** Clicking a folder title in the sidebar no longer empties the chat lists of your expanded folders for a moment before they reappear. [#27535](https://github.com/open-webui/open-webui/pull/27535), [#27533](https://github.com/open-webui/open-webui/issues/27533)
+- 🫧 **Flickering sidebar rows.** Moving the pointer across a chat in the sidebar no longer makes its title and timestamp flicker in and out, or draw the timestamp underneath the action buttons. [#27474](https://github.com/open-webui/open-webui/pull/27474), [#27473](https://github.com/open-webui/open-webui/issues/27473)
+- ⭐ **Rating scale in multi-model replies.** The rating scale in the feedback panel is no longer cut off when several models answer side by side, so every score can be picked. [#26846](https://github.com/open-webui/open-webui/issues/26846)
+- 🧑‍🤝‍🧑 **Duplicate models side by side.** Adding the same model twice in a side-by-side chat now keeps each column's own answer after a reload, instead of every column collapsing onto the first one. [#26980](https://github.com/open-webui/open-webui/pull/26980)
+- ⬅️ **Back button after opening admin or workspace.** Going back in the browser now returns you to the page you came from, instead of being pushed forward again to where you just were. [#27478](https://github.com/open-webui/open-webui/pull/27478), [#27477](https://github.com/open-webui/open-webui/issues/27477)
+- 🎛️ **Typing a top_k value.** The top_k box in advanced parameters now accepts whole numbers up to its limit and rejects anything else, instead of letting the slider and the box disagree over what is allowed. [Commit](https://github.com/open-webui/open-webui/commit/34920213619eb66467470105cbe3a275c812ccbc), [#26669](https://github.com/open-webui/open-webui/issues/26669)
+- 🌙 **Date pickers in dark mode.** The calendar and clock icons on date and time fields are now visible in dark mode, across the calendar, automation schedules, account settings and analytics. [#27275](https://github.com/open-webui/open-webui/pull/27275), [#27274](https://github.com/open-webui/open-webui/issues/27274)
+- 🪞 **Settings content stays inside the window.** Long chat titles in Archived Chats now shorten with the full title on hover, and the admin analytics tables and chart no longer stretch past the edge of the settings window. [#27306](https://github.com/open-webui/open-webui/pull/27306), [#27305](https://github.com/open-webui/open-webui/issues/27305), [#27329](https://github.com/open-webui/open-webui/issues/27329)
+- 🔗 **Settings links that open in place.** A link to a settings tab now opens it without a page refresh, and the Add Terminal button in the terminal menu goes straight to the Integrations tab instead of flashing the admin panel and doing nothing. [#27552](https://github.com/open-webui/open-webui/pull/27552), [#27551](https://github.com/open-webui/open-webui/issues/27551)
+- 🎰 **Model choice on a fresh chat.** Starting a new chat now falls back to your default model when the previous selection is no longer available, instead of leaving the picker empty, while a model named in the link still wins. [Commit](https://github.com/open-webui/open-webui/commit/f91ac068d09eed381e14d35472d80ae4670fe52b), [#26697](https://github.com/open-webui/open-webui/pull/26697)
+- 📱 **Model selector on small screens.** The model list now stays fully on screen and sizes itself to the space available, instead of running past the edge or hiding behind the on-screen keyboard on phones. [Commit](https://github.com/open-webui/open-webui/commit/79d3e34eea6dc2828d1945cc2b9fca5d662d825b), [Commit](https://github.com/open-webui/open-webui/commit/e39ff71532651438c32b2aa7ffe2b6068c94e6b2), [Commit](https://github.com/open-webui/open-webui/commit/ea31a3bd61fdf1a72206f9ed5f7252486d554c9b)
+- 📲 **Sidebar stays open over the calendar.** Opening the calendar from the account menu on a phone now closes the sidebar, as every other entry in that menu already did. [#26979](https://github.com/open-webui/open-webui/pull/26979)
+- 🗓️ **Automation dialog on narrow screens.** The buttons along the bottom of the automation dialog now sit on their own row on a phone, instead of the schedule and model pickers wrapping and pushing Cancel into the middle. [#27027](https://github.com/open-webui/open-webui/pull/27027)
+- 📐 **Input menu with keyboard open.** The message input's attachment menu now stays on screen and resizes to fit when the on-screen keyboard is open on mobile, instead of running off the edge. [Commit](https://github.com/open-webui/open-webui/commit/6e5efc1f757c614814ba88bbae6caa3aaddda528)
+- 🎈 **Dropdowns that follow their content.** A menu now stays in place as its contents grow or shrink, instead of running past the edge of the screen when a submenu swaps in taller content, and no longer bounces as it opens. [#27460](https://github.com/open-webui/open-webui/pull/27460), [#27458](https://github.com/open-webui/open-webui/issues/27458)
+- 🧾 **Attachment menus load once.** Opening a submenu of the attachment menu now requests its list a single time instead of twice. [#27461](https://github.com/open-webui/open-webui/pull/27461), [#27459](https://github.com/open-webui/open-webui/issues/27459)
+- 🔦 **Chat search on PostgreSQL.** Searching your chats now finds matches in current conversations on PostgreSQL setups, instead of only matching chats still stored in the older format. [Commit](https://github.com/open-webui/open-webui/commit/cc9a44569ef08b64ff44d15607c43966f362ce75)
+- 🧲 **Search quality with prefix-based embedding models.** Memories, knowledge base descriptions and searches against an external vector database now carry the query and content markers your embedding model expects, so results are no longer quietly worse than they should be on models that rely on them. [Commit](https://github.com/open-webui/open-webui/commit/c4f5ac65ee3cd20dd1d507eda04fca21a866910a), [#26958](https://github.com/open-webui/open-webui/pull/26958), [#26353](https://github.com/open-webui/open-webui/issues/26353)
+- 🥄 **Counting matches in knowledge base commands.** Piping text into a search inside knowledge base commands now honours the count and filenames-only flags, instead of returning the matching lines regardless. [#26721](https://github.com/open-webui/open-webui/pull/26721), [#26715](https://github.com/open-webui/open-webui/issues/26715)
+- 🔍 **Knowledge base file search.** Searching inside knowledge base files now returns matching lines with correct line numbers, and patterns that list alternatives separated by a pipe find matches instead of silently returning none. [#27249](https://github.com/open-webui/open-webui/pull/27249), [Commit](https://github.com/open-webui/open-webui/commit/e18e249d5da3d8fe701a885edc64341cc5dbf813), [Commit](https://github.com/open-webui/open-webui/commit/8d2fee5d4559d030b53575d377a0013e2c67b9fe), [#26795](https://github.com/open-webui/open-webui/pull/26795), [#26781](https://github.com/open-webui/open-webui/issues/26781), [#26744](https://github.com/open-webui/open-webui/issues/26744)
+- 🖨️ **PDF text recognition.** The text recognition package is now included again, so the application starts and PDFs with image text extraction enabled upload correctly instead of failing. [#26851](https://github.com/open-webui/open-webui/pull/26851), [#26646](https://github.com/open-webui/open-webui/issues/26646), [#26994](https://github.com/open-webui/open-webui/issues/26994)
+- 🧿 **Mistral OCR on a stock install.** Extracting documents with Mistral OCR now works out of the box, instead of failing on a missing name resolution library that the code assumed was present. [#27440](https://github.com/open-webui/open-webui/pull/27440)
+- 📧 **Outlook message uploads.** Uploading a .msg email now works, where it previously failed because the package it relied on could not be installed alongside the rest of the application at all. [#26704](https://github.com/open-webui/open-webui/pull/26704), [#26690](https://github.com/open-webui/open-webui/issues/26690)
+- 🖇️ **Uploads with PaddleOCR-VL selected.** With PaddleOCR-VL chosen as the document loader, only PDFs and images now go to it and everything else falls back to the usual handling, so text, markdown, spreadsheet and Word files index instead of being rejected. [#27529](https://github.com/open-webui/open-webui/pull/27529), [#24988](https://github.com/open-webui/open-webui/issues/24988), [#26759](https://github.com/open-webui/open-webui/issues/26759)
+- 🪙 **Documents containing special tokens.** Splitting text by tokens no longer fails when the content contains reserved marker sequences, so those pages and files can be fetched and added to a knowledge base. [Commit](https://github.com/open-webui/open-webui/commit/33cf3fbb7f017ab1b79dce5c5ca4d4e1c3092844), [#27094](https://github.com/open-webui/open-webui/issues/27094)
+- 📚 **Knowledge base upload reliability.** Adding a file directly to a knowledge base now finishes processing and linking the file before reporting success, so uploaded files are reliably searchable. [Commit](https://github.com/open-webui/open-webui/commit/f5b196c060805fd22e1aa1c9f738b60221ef0fd8)
+- 🛠️ **Web loader settings from the admin panel.** The web loader picked in admin settings is now actually used, along with its certificate checking, request pacing and proxy settings, so instances that fetch pages through an external loader work again instead of trying to reach the internet directly with whatever was configured at startup. [#26749](https://github.com/open-webui/open-webui/pull/26749), [#26747](https://github.com/open-webui/open-webui/issues/26747), [Commit](https://github.com/open-webui/open-webui/commit/304cbe4569cddbc9e4641186e51bc9e8d5154533), [#27083](https://github.com/open-webui/open-webui/pull/27083), [#27025](https://github.com/open-webui/open-webui/pull/27025), [#27061](https://github.com/open-webui/open-webui/issues/27061)
+- 🚧 **Quoted entries in the web fetch filter list.** Stray quote marks around a filter entry, which Docker Compose passes through literally, no longer turn the list into one that blocks every web address. [#26910](https://github.com/open-webui/open-webui/pull/26910), [#26908](https://github.com/open-webui/open-webui/issues/26908)
+- 🌐 **Web fetching with certain plugins installed.** Fetching a web page and loading web search results work again on instances where a tool or function pulls in a replacement networking library, which previously made every fetch fail and return nothing. [#26796](https://github.com/open-webui/open-webui/pull/26796), [#26791](https://github.com/open-webui/open-webui/issues/26791)
+- 📢 **Web search failures explained.** When a search finds pages but cannot store them, the chat now says what went wrong and points at the document settings, instead of reporting sites searched and then no sources found. [#26883](https://github.com/open-webui/open-webui/pull/26883)
+- 🕸️ **Mixed web page extraction.** Fetching several web pages at once now reads each one according to its own format, instead of applying the first page's format to the whole batch and garbling the rest. [#27367](https://github.com/open-webui/open-webui/pull/27367)
+- 🧯 **Leftover browser sessions on web fetches.** Fetching pages through a remote Playwright server now closes each page and the browser even when a page times out or the search is abandoned partway, instead of leaving sessions open and slowing every later search until that server was restarted. [#27526](https://github.com/open-webui/open-webui/pull/27526), [#25880](https://github.com/open-webui/open-webui/issues/25880)
+- 🖇️ **Sign-in profile pictures fetched safely.** The profile picture pulled in when someone signs in through a provider is now fetched through the same protected path as other outbound requests, so a host that changes its address between the check and the fetch can no longer point it at an internal service, taking the forwarded sign-in token with it. [#26699](https://github.com/open-webui/open-webui/pull/26699)
+- 🪃 **Backslashes in terminal proxy paths.** A request to the terminal proxy containing a backslash is now refused, closing a way to smuggle directory traversal past the path check to an upstream that treats it as a separator. [#27198](https://github.com/open-webui/open-webui/pull/27198)
+- 🧱 **Internal addresses disguised as public ones.** A web address that hides an internal target inside an IPv6 address, through the mapped, 6to4, Teredo or NAT64 forms, is now recognised and refused like any other internal address. [Commit](https://github.com/open-webui/open-webui/commit/1717b493d83c86afa82aa8bc50139250852dd2f3)
+- 🪤 **Tighter checks when a page is fetched.** Every request a fetched page makes is now checked against the address rules rather than only the page itself, each hop of a redirect is checked in turn, and background workers and socket connections the page tries to open are refused. [Commit](https://github.com/open-webui/open-webui/commit/bef63a2ae915571d50d2722a635e8bfa753d7877), [#27042](https://github.com/open-webui/open-webui/pull/27042), [#27008](https://github.com/open-webui/open-webui/pull/27008)
+- 🐢 **Dropped pages when fetches are paced.** Pages fetched through Firecrawl, Tavily, Microsoft Web IQ or Playwright are no longer discarded whenever the loader has to pause between requests, which quietly lost any page following close behind another and sometimes blamed it on a failed security check. [#27528](https://github.com/open-webui/open-webui/pull/27528), [#26079](https://github.com/open-webui/open-webui/issues/26079)
+- 🎙️ **Dictation repeating earlier speech.** Dictating into the message box no longer re-inserts everything you said in previous recordings, and cancelling a recording no longer inserts the text anyway. [#26793](https://github.com/open-webui/open-webui/pull/26793), [#26784](https://github.com/open-webui/open-webui/issues/26784)
+- 🧩 **Order of long transcriptions.** A long recording split into pieces for transcription is now reassembled in the order it was spoken, instead of sections sometimes appearing out of sequence in the transcript and everything read from it. [#27417](https://github.com/open-webui/open-webui/pull/27417), [#27143](https://github.com/open-webui/open-webui/issues/27143)
+- 🔊 **Text-to-speech reliability.** Text-to-speech playback and other streamed responses no longer intermittently cut out partway through when several requests run at once. [#26924](https://github.com/open-webui/open-webui/pull/26924), [#26922](https://github.com/open-webui/open-webui/issues/26922)
+- 🧮 **Anthropic usage reporting.** Responses from the Anthropic-compatible API now report accurate input and output token counts, pass through cache and server tool figures where the provider gives them, and leave the input count out entirely rather than reporting zero when it is unknown. [Commit](https://github.com/open-webui/open-webui/commit/e8b59b2ef35ecb727fa760cd565d6da20c9e7e79), [Commit](https://github.com/open-webui/open-webui/commit/51ff386fd6461c07225d10a0de530019eebdd157), [Commit](https://github.com/open-webui/open-webui/commit/0576e8eeb5797a36b43eba5790a4e2a5dd8e5a4d), [Commit](https://github.com/open-webui/open-webui/commit/8e74cac8decc0a54137d7214e70c33b0dd52a99a), [Commit](https://github.com/open-webui/open-webui/commit/93a34bb25b32ae0b3a876fd3161c7778227b76bf), [Commit](https://github.com/open-webui/open-webui/commit/4c2d864b3f4c1ae6c4c9bd93aea078d4bf520463), [#26790](https://github.com/open-webui/open-webui/pull/26790), [#27293](https://github.com/open-webui/open-webui/pull/27293), [Docs:#1328](https://github.com/open-webui/docs/issues/1328)
+- 📨 **Non-streaming requests to strict providers.** A request that is not streaming no longer carries the streaming-only usage option, which some providers reject outright.
+- 🪝 **Tool calls with structured arguments.** A provider that sends a tool call's arguments as an object, or as nothing at all, no longer breaks the reply partway through. [#27195](https://github.com/open-webui/open-webui/issues/27195)
+- 🧬 **Shared pipe model tool calls.** Non-admin users of a shared model built on a pipe or manifold model no longer see the response silently stop right after a tool call. [#26906](https://github.com/open-webui/open-webui/pull/26906), [#26900](https://github.com/open-webui/open-webui/issues/26900)
+- 🧑‍🔧 **Startup as an arbitrary user.** Running the image as a non-root account, as OpenShift and similar setups do, no longer fills the boot log with permission errors while it writes its own icons and manifest. [#26664](https://github.com/open-webui/open-webui/pull/26664), [#26662](https://github.com/open-webui/open-webui/issues/26662)
+- 🩹 **Startup with an ownerless tool or function.** A tool or function left without an owner no longer prevents the application from starting, which had blocked all chat responses until it was removed. [#26850](https://github.com/open-webui/open-webui/pull/26850), [#26843](https://github.com/open-webui/open-webui/issues/26843)
+- 🏷️ **Model names containing a connection prefix.** A prefix set on a connection is now removed only from the front of the model name, so a model whose own name contains that text is no longer mangled before the request is sent. [Commit](https://github.com/open-webui/open-webui/commit/ed663f16ecaaf99de194922d1634ecc5d906c703)
+- 🦙 **Newly pulled Ollama models.** Sending a message to a model that was pulled after the list was last built now refreshes the list and proceeds, instead of reporting the model as not found. [Commit](https://github.com/open-webui/open-webui/commit/ed663f16ecaaf99de194922d1634ecc5d906c703), [#27353](https://github.com/open-webui/open-webui/pull/27353)
+- 🗑️ **Deleting a model from the selector.** Removing a workspace model from the model selector menu now deletes just that model and leaves the underlying one in place, instead of failing with a not found error. [#26819](https://github.com/open-webui/open-webui/pull/26819)
+- 🔑 **Connecting a remote MCP server over OAuth.** Setting up a remote MCP server now reports plainly when its sign-in details cannot be discovered, rather than saving an unusable connection that failed with a server error the moment you tried to authorise it. [#26654](https://github.com/open-webui/open-webui/pull/26654), [#26647](https://github.com/open-webui/open-webui/issues/26647)
+- 🪢 **Tool servers with cross-referencing types.** A tool server whose description defines types that refer to each other now loads its tools instead of failing outright, so the integration appears in model and tool selection again. [#27413](https://github.com/open-webui/open-webui/pull/27413), [#27239](https://github.com/open-webui/open-webui/issues/27239)
+- 👥 **Previewing what someone can use.** The preview of a person's access now includes the models, knowledge bases and tools they own, not just the ones shared with them. [Commit](https://github.com/open-webui/open-webui/commit/a9a3e5b95c8e641881fedc1ce7431eedab9a371b), [#27423](https://github.com/open-webui/open-webui/pull/27423), [#27407](https://github.com/open-webui/open-webui/discussions/27407)
+- 🧰 **Model editor loading.** The model editor no longer fails to open when its tool list can't be loaded, falling back gracefully instead. [Commit](https://github.com/open-webui/open-webui/commit/10724d057af13a826c52e92b1c01a031656768d5)
+- 🗃️ **Milvus Lite collection creation.** Setting up collections now succeeds on embedded Milvus Lite, which previously could fail while creating the resource index. [#26911](https://github.com/open-webui/open-webui/pull/26911)
+- 🧽 **Milvus log noise.** Instances backed by Milvus no longer fill their logs with deprecation warnings while indexing and retrieving, and keep working with future PyMilvus releases that drop the old interface entirely. [#27521](https://github.com/open-webui/open-webui/pull/27521), [#26978](https://github.com/open-webui/open-webui/issues/26978)
+- 🚏 **Stray terminal containers.** Terminal orchestrator connections that use a policy now send every request through that policy, so each person no longer ends up with a second unintended container alongside the intended one. [#26945](https://github.com/open-webui/open-webui/issues/26945), [Commit](https://github.com/open-webui/open-webui/commit/7088d245bb45fc69c0b22748563b9f3c6f0daa73)
+- 🔦 **Connections on hardened instances.** With the admin access bypass turned off, a connection that has no access grants yet is now reachable by administrators again, instead of being hidden from everyone including the admin who created it. [#27581](https://github.com/open-webui/open-webui/pull/27581), [#27580](https://github.com/open-webui/open-webui/issues/27580), [#27064](https://github.com/open-webui/open-webui/issues/27064)
+- ♻️ **Connection changes take effect immediately.** Saving connection settings now refreshes the model list straight away, instead of leaving the previous models in place until the server was restarted.
+- 🚫 **Disabled OpenAI connections are enforced.** Turning off the OpenAI API now blocks chat requests to it and clears its models, rather than only hiding it from the interface.
+- 🪛 **Deleting an Ollama connection.** Removing an Ollama connection now saves straight away, instead of reappearing until the Ollama API switch was toggled afterwards. [#27483](https://github.com/open-webui/open-webui/pull/27483), [#27482](https://github.com/open-webui/open-webui/issues/27482)
+- 🧹 **Orphaned sessions get cleaned up.** The instance that reaps sessions left behind by a crashed worker now keeps trying if another instance holds the job, rather than one instance giving up for good and leaving stale sessions to accumulate, and the lock it uses can no longer be released or renewed by an instance that does not hold it. [Commit](https://github.com/open-webui/open-webui/commit/bf35f64a7f14161933dfa608577a977d107b1569), [Commit](https://github.com/open-webui/open-webui/commit/846ba80a9d5e75837d3db37185e9a23b1e6bfe78)
+- 🧊 **Redis cluster connections.** A deployment using Redis in cluster mode is no longer handed a connection built for a single server, or the reverse, when both point at the same address. [Commit](https://github.com/open-webui/open-webui/commit/fc4906c9e9df3fa42bb9073ac197383347caa853)
+- 🚏 **Stopping a reply when Redis is configured.** The stop button now actually halts generation on Redis-backed deployments, where the listener that carries stop requests between instances quietly died after a few idle seconds and left tokens streaming on, and a new "REDIS_SOCKET_TIMEOUT" setting controls that timeout. [#27104](https://github.com/open-webui/open-webui/pull/27104), [#26779](https://github.com/open-webui/open-webui/issues/26779)
+- 🛟 **Redis failover on timeouts.** A Redis connection that times out now retries against a freshly resolved primary instead of failing, so Sentinel setups recover from a failover rather than erroring out. [Commit](https://github.com/open-webui/open-webui/commit/75a8a0046b5b2ebd9942b25035b346aa953f81cc), [#27210](https://github.com/open-webui/open-webui/issues/27210)
+- 👣 **First sign-in through a trusted header.** Two requests arriving together for someone signing in for the first time through a trusted header no longer create two accounts for the same person, and the database now refuses a second account for an address that already exists, whatever its capitalisation. [Commit](https://github.com/open-webui/open-webui/commit/b190dcf3caa00dc8b7b9c7312828298d9143f60d), [Commit](https://github.com/open-webui/open-webui/commit/50e050e1957de40caa9df479b4c0d9b814f1f623), [#27571](https://github.com/open-webui/open-webui/pull/27571), [#27117](https://github.com/open-webui/open-webui/issues/27117)
+- 🔧 **Sign-on settings from environment variables.** Single sign-on settings supplied through environment variables are no longer overridden by stale values saved at first startup, so changing them takes effect. [#26928](https://github.com/open-webui/open-webui/pull/26928), [#26917](https://github.com/open-webui/open-webui/issues/26917)
+- 🎫 **Expired identity tokens sent to tools.** A sign-in session is now refreshed before the earliest of its tokens expires, so tools and pipes that forward your identity no longer hand a downstream service a token it rejects. [#27520](https://github.com/open-webui/open-webui/pull/27520), [#27066](https://github.com/open-webui/open-webui/issues/27066)
+- 🎫 **Sign-in tokens that never expire.** A provider that returns no expiry and no way to refresh is now taken at its word, instead of being given an invented one-hour lifetime that left the session unusable afterwards. [Commit](https://github.com/open-webui/open-webui/commit/98656b7c5e29383b61d2113164466b8d4ab1d424), [#26802](https://github.com/open-webui/open-webui/pull/26802), [#26141](https://github.com/open-webui/open-webui/issues/26141)
+- 🔓 **Single sign-on after a key rotation.** Signing in with OIDC now recovers when the provider rotates its signing key, refreshing the cached keys and retrying instead of failing with an invalid credentials error. [#27310](https://github.com/open-webui/open-webui/pull/27310), [#26407](https://github.com/open-webui/open-webui/issues/26407)
+- 🔑 **Signing in after a session expires.** An expired session now cleanly returns you to the sign-in page and back to where you were afterwards, instead of bouncing you away from the sign-in page or leaving a stale session behind. [Commit](https://github.com/open-webui/open-webui/commit/609cc6ad9b597c6a3f4df6f9dba93d6ff6ec1f18), [Commit](https://github.com/open-webui/open-webui/commit/29782aba01b8f34625949170dc9ee9e1c5872893), [#26751](https://github.com/open-webui/open-webui/pull/26751), [#26731](https://github.com/open-webui/open-webui/issues/26731)
+- 🫥 **Temporary chats and channels write nothing.** Generating or editing an image and status updates in a temporary chat or a channel message no longer try to save themselves against a conversation that was never stored, and the task list tools are no longer offered there at all rather than being offered and then failing. [Commit](https://github.com/open-webui/open-webui/commit/d484a2a99e3a0c21fdcad007a50ebc412fffbb2e), [Commit](https://github.com/open-webui/open-webui/commit/d2936c880cfc8cb71bb5c235926048ae189bba25), [Commit](https://github.com/open-webui/open-webui/commit/b45c020f68a9499b66e598e854b17a3f232b6cf2), [Commit](https://github.com/open-webui/open-webui/commit/71c4da8c065491a96e41da3c9f0c663e5f759468), [#27432](https://github.com/open-webui/open-webui/issues/27432)
+- 🎞️ **Artifacts panel reopening itself.** The artifacts panel now opens once when a finished block is detected, so closing it partway through a reply no longer sees it forced back open on every word that follows. [Commit](https://github.com/open-webui/open-webui/commit/4856afcef8251969f751ade5760cefea9577c051), [#27399](https://github.com/open-webui/open-webui/issues/27399)
+- 🏞️ **Images returned by a tool.** Images a tool produces are now passed to the model in a form the OpenAI-compatible providers accept, so it can actually look at them instead of receiving a result it cannot read. [Commit](https://github.com/open-webui/open-webui/commit/dd86b984bd508cf2841f08dae90411dfb5fe407f)
+- 🖼️ **External message images.** Images hosted on other sites and referenced in a message now display inline instead of being replaced with a placeholder. [Commit](https://github.com/open-webui/open-webui/commit/890bfd0d9771d1919ce24f04e11b6c96589fca5b)
+- 🔣 **Names containing a vertical bar.** What you insert with the at sign or a slash is now recorded by the key you typed rather than guessed from its name, so a prompt or model whose name contains a vertical bar is no longer mistaken for a skill. [Commit](https://github.com/open-webui/open-webui/commit/e28b391e514384ec329ca871d02189aa81fb1a00)
+- 〰️ **Text above a collapsible block.** A line written directly above a collapsible section is no longer turned into a large heading, and the section itself still renders as a collapsible widget rather than leaking its markup. [Commit](https://github.com/open-webui/open-webui/commit/7d77efe0f1cfa4782893ffde78419a96a95240f4), [#27148](https://github.com/open-webui/open-webui/pull/27148), [#27001](https://github.com/open-webui/open-webui/issues/27001)
+- ✳️ **Asterisks in the message input.** Wrapping a word in asterisks no longer silently turns it italic and swallows the asterisks, so your prompt reaches the model exactly as you typed it. [Commit](https://github.com/open-webui/open-webui/commit/001775d8e868ce44f125e0d683dd50363f0e8318)
+- 📶 **Reconnect warnings on mobile.** Switching back to Open WebUI after using another app no longer flashes a connection lost warning while the tab wakes up and reconnects on its own. [Commit](https://github.com/open-webui/open-webui/commit/63ada247066dfc51e0e9559366f0cfd9a98db40b)
+- 🧭 **Sidebar access from the automation editor.** Opening an automation on a phone no longer hides the sidebar button, so you can move around without leaving the editor first.
+- 🔣 **Chats containing unusual characters.** Broken character sequences are now cleaned out of text before it is stored, so a conversation that picked one up still saves and still opens instead of failing to load. [Commit](https://github.com/open-webui/open-webui/commit/43e7eefa959918baf9fbf867a12b5d721fd782af), [#27201](https://github.com/open-webui/open-webui/pull/27201), [#27081](https://github.com/open-webui/open-webui/issues/27081)
+- 📛 **Failures after a tool call.** A reply that fails while continuing after a tool call or a code interpreter run now says so, instead of stopping mid-answer with nothing to explain why. [Commit](https://github.com/open-webui/open-webui/commit/8ab44ed3b153dd8d8d57a444c98d100f281f7f7e), [#27426](https://github.com/open-webui/open-webui/pull/27426), [#27411](https://github.com/open-webui/open-webui/issues/27411)
+- 💾 **Errors kept after reloading.** An error that ends a streamed reply is now saved to the conversation, so it is still there when you reload instead of disappearing. [#27365](https://github.com/open-webui/open-webui/pull/27365), [#27074](https://github.com/open-webui/open-webui/issues/27074)
+- 💬 **Readable error messages.** Errors in a conversation now always show readable text that wraps instead of running off the edge, including errors that arrive wrapped inside another error.
+- 🪝 **Blocked webhook targets look like failures.** A webhook pointing at an address that is not publicly reachable is now skipped with a short warning, instead of an error and a full traceback that read like the server crashing on startup. [Commit](https://github.com/open-webui/open-webui/commit/0671b7aa2b59f5c6235bfa85d6aba54e2ba91353), [#26975](https://github.com/open-webui/open-webui/issues/26975)
+- 🕵️ **Values printed in error logs.** A failure no longer prints the contents of nearby variables alongside its traceback, which could put keys and message content into the logs, and a new "LOGURU_DIAGNOSE" setting turns that detail back on for debugging. [Commit](https://github.com/open-webui/open-webui/commit/6aebfd88e938d1cd139068b5737a83a82ff393ed), [#26814](https://github.com/open-webui/open-webui/pull/26814)
+- 🪵 **Empty audit exclusion list.** Clearing the list of paths excluded from audit logging no longer switches off auditing altogether, so requests are recorded as intended. [#27370](https://github.com/open-webui/open-webui/pull/27370), [Commit](https://github.com/open-webui/open-webui/commit/2ef6c76f5126ccdef5d1c814004920941275f45d)
+- 🗒️ **Readable audit log bodies.** Audit logs that record response bodies now store them as readable text instead of compressed data, so entries are legible whenever a browser requested compression. [#27369](https://github.com/open-webui/open-webui/pull/27369)
+- 👍 **Rating in feedback events.** Events sent when someone rates a response now carry the rating that was given, instead of reporting it as empty. [Commit](https://github.com/open-webui/open-webui/commit/300302d43259e119cd88247b3f246bea82b4dd8e), [#26840](https://github.com/open-webui/open-webui/issues/26840)
+- ⏱️ **Accurate request timing header.** The processing time reported on each response now includes fractions of a second instead of rounding everything under a second down to zero. [#27368](https://github.com/open-webui/open-webui/pull/27368)
+- 📋 **Provider rejection logging.** When a model provider rejects a request, the reason it gave is now recorded in the server logs, so administrators can diagnose failures without querying the provider directly. [#27238](https://github.com/open-webui/open-webui/pull/27238), [#27237](https://github.com/open-webui/open-webui/issues/27237), [#26253](https://github.com/open-webui/open-webui/issues/26253)
+- ⏳ **Faster licensed startup.** Instances with a license key no longer wait on the license server during startup, so the app becomes ready to serve traffic without that delay. [Commit](https://github.com/open-webui/open-webui/commit/8f7753331752e72b17ef8f055318ab548a73f4b8), [Commit](https://github.com/open-webui/open-webui/commit/0c7ddbdb4f7dbd46f1dadc3242dbb13b81b47758)
+- 📅 **Calendar invitation responses.** Whether you have accepted an invitation is now decided by your own response rather than by whoever created the event, and invitations you decline disappear from your calendar. [#27007](https://github.com/open-webui/open-webui/pull/27007)
+- 🗓️ **Schedules written by hand.** A recurrence rule is now read the same way whether it is written in upper or lower case, a start date in the rule is respected, second-by-second rules are understood, and a rule that cannot be supported is refused with a clear message instead of behaving unpredictably. [Commit](https://github.com/open-webui/open-webui/commit/c4ae8c86786fed521960466f6d8eef8af22c2946), [Commit](https://github.com/open-webui/open-webui/commit/2d928df30443516a9a3d6b71b0f31426a2362499), [#27470](https://github.com/open-webui/open-webui/pull/27470)
+- 📅 **One calendar event stalling the server.** Working out when a repeating event happens next now walks its rule once rather than re-counting from the beginning for every occurrence, so an event repeating every minute from an old start date can no longer occupy the server for everyone. [#27468](https://github.com/open-webui/open-webui/pull/27468)
+- ⏰ **Recurring automation scheduling.** Automations that repeat every few minutes or hours now align to the clock and are no longer wrongly rejected as having no upcoming runs when the server clock is ahead of your timezone. [Commit](https://github.com/open-webui/open-webui/commit/b3aead23da6cf8ebeedbd9fa3b97c7ac1a3f54ec), [#26954](https://github.com/open-webui/open-webui/issues/26954)
+
+### Changed
+
+- ⚠️ **Database Migrations**: This release includes database schema changes; we strongly recommend backing up your database and all associated data before upgrading in production environments. If you are running a multi-worker, multi-server, or load-balanced deployment, all instances must be updated simultaneously, rolling updates are not supported and will cause application failures due to schema incompatibility.
+- 🛠️ **Admin settings moved into settings.** Admin settings and the analytics dashboard are no longer separate pages and now open alongside your personal settings in the settings window, under their own Admin section, with the old links redirecting there. [Commit](https://github.com/open-webui/open-webui/commit/c1460570b7e2897a3e648441a0206ad25e603b8c), [Commit](https://github.com/open-webui/open-webui/commit/3ce3c529365a3cd9e5631b14bfffead6baa2b1ed), [Commit](https://github.com/open-webui/open-webui/commit/667cba1a9561166941f59faf3bfa24038288b449)
+- 📁 **Workspace actions in one menu.** Creating, importing, and exporting workspace items no longer have their own buttons on each page and are now reached from a single Create menu in the workspace header, with creating a prompt or knowledge base opening a dialog rather than a separate page. [Commit](https://github.com/open-webui/open-webui/commit/05e3f713175c1eea43a99f29521eb01700c21d3c), [Commit](https://github.com/open-webui/open-webui/commit/f8350360dfd60ff890b73fe2f39aaf20a52ad28b), [Commit](https://github.com/open-webui/open-webui/commit/91277726cd666276ce4b41a722a142eb539089b8), [Commit](https://github.com/open-webui/open-webui/commit/1760b073c7595d4075a91b986520ff8eeeaebc35)
+- 🔐 **Administrators no longer reach other people's automations.** Viewing, editing, running and deleting an automation is now limited to the person who created it, so an administrator with a link to someone else's automation is refused rather than allowed through. [Commit](https://github.com/open-webui/open-webui/commit/f798d05586a140f1a6b51f1e51b2b2a63d079d45)
+- 🏷️ **Shorter titles without emojis.** Automatically generated titles for chats and notes are now two to four words and no longer include an emoji, and anyone who prefers the old style can restore it by editing the title generation prompt in admin settings. [Commit](https://github.com/open-webui/open-webui/commit/50d3c927bfed3b8dd94fd9f79bff258a84ecbd92), [Commit](https://github.com/open-webui/open-webui/commit/7a9928ef172b7c280c377c86cb52957e39340158)
+- 🗂️ **Archived chats moved to settings.** The Archived Chats shortcut is no longer in the user menu, and your archived conversations are now reached through Settings, where they can also be searched and sorted. [Commit](https://github.com/open-webui/open-webui/commit/9f17c5960a0e47a09773da4bba12997a31222fc8), [Commit](https://github.com/open-webui/open-webui/commit/8dd862d3383978f21111e63fb2d6029711abed9a)
+- 🔢 **Usage now reports the latest call separately.** In a response's usage block, "prompt_tokens" and "completion_tokens" now carry the counts from the most recent model call rather than the running total, while "input_tokens", "output_tokens" and "total_tokens" stay cumulative, so anything reading the first pair for billing should read the second set instead. [Commit](https://github.com/open-webui/open-webui/commit/df94268e892cbb66675170a6c78846aef23f6e89), [#27031](https://github.com/open-webui/open-webui/issues/27031)
+- 🧳 **The "python-jose" library is no longer installed.** Nothing in Open WebUI imports it anymore, so it and the two packages it pulled in have been dropped from the image, and any tool or function that imports it directly now needs to install it itself. [#27444](https://github.com/open-webui/open-webui/pull/27444)
+- 📦 **Storage emulator no longer bundled.** The optional Google Cloud Storage emulator is no longer installed as part of the full package, so anyone who relied on it for local storage testing now needs to install "gcp-storage-emulator" themselves. [Commit](https://github.com/open-webui/open-webui/commit/30415c925a18b1ea1c3f2739bd944dd939f020cf)
+
+## [0.10.2] - 2026-07-01
+
+### Added
+
+- 💭 **Streamed reasoning display.** Models that emit thinking or reasoning now show that content as it streams, and it renders correctly in the chat overview and in exported conversations. [Commit](https://github.com/open-webui/open-webui/commit/0b75445ff9a42e37640c034812d0de9b84039e60), [Commit](https://github.com/open-webui/open-webui/commit/af1c0eee89810fa4c36e3eb7e4eba6de685bd7ca), [Commit](https://github.com/open-webui/open-webui/commit/4b08d65597e5b634b7191b0bd6d28feeafcc2a48), [Commit](https://github.com/open-webui/open-webui/commit/fa2abe4cb6a085a4c9045bf8d8ff4b6beffdef6a)
+- 🗂️ **Folder uploads to knowledge bases.** Dragging a folder into a knowledge base, or syncing one, now recreates its subfolder structure instead of flattening everything into loose files. [#26130](https://github.com/open-webui/open-webui/issues/26130), [Commit](https://github.com/open-webui/open-webui/commit/2ed8934f5b2bc8a11c74ef2f34cdb62ef024809e)
+- 🧠 **Memory system context toggle.** Administrators can now keep memory tools available while choosing not to add stored memories to the system context, using the new 'Memory System Context' toggle in admin settings. [Commit](https://github.com/open-webui/open-webui/commit/4067e357b2ff9e2fb59866d24656e832908fb6fe)
+- 🧹 **Tidier automatic memories.** Automatically saved memories now focus on enduring details like preferences and goals and skip one-off things like meals, routine events, or passing mood unless you ask to remember them. [Commit](https://github.com/open-webui/open-webui/commit/80af65c24adac5140a39a2b5687a3b669b86719f)
+- 🎙️ **Speech-to-text request format.** OpenAI-compatible speech-to-text can now send audio as either a multipart upload or base64 JSON, selectable in admin audio settings. [Commit](https://github.com/open-webui/open-webui/commit/989c6c13f5d4c5cc255aec77aea14725104d9cb3)
+- 🧰 **API configs via environment.** Administrators can now set per-connection Ollama and OpenAI API configurations through the "OLLAMA_API_CONFIGS" and "OPENAI_API_CONFIGS" environment variables. [Commit](https://github.com/open-webui/open-webui/commit/19d8f03bd2c64013b510f2a0eeb3513d452814a3)
+- 📡 **Provider failure events.** Failed Ollama and OpenAI-compatible provider requests now emit a structured event describing the error type, provider, and status, giving administrators clearer visibility into upstream failures. [Commit](https://github.com/open-webui/open-webui/commit/4351c78b1e45bb0c5824f9d9eb911b395e343646)
+- 🏟️ **Arena models via environment.** Administrators can now define evaluation arena models through the "EVALUATION_ARENA_MODELS" environment variable. [#26174](https://github.com/open-webui/open-webui/issues/26174)
+- ♿ **Clearer high-contrast sidebar selection.** With high-contrast mode enabled, the currently selected chat in the sidebar now stands out with stronger colors, making it easier to tell which chat is active. [#26469](https://github.com/open-webui/open-webui/issues/26469), [Commit](https://github.com/open-webui/open-webui/commit/52ee5cb1b3aa8a8ad40b23d72d82db7c3522dbf8)
+- 🔄 **General improvements.** Various improvements were implemented across the application to enhance performance, stability, and security.
+- 🌐 **Translation updates.** Translations for Thai, Portuguese (Brazil), Catalan, and Spanish were enhanced and expanded.
+
+### Fixed
+
+- 🛡️ **Security Advisory**: This release includes security and access-control fixes. We recommend updating production deployments at your earliest convenience. Not all security fixes in this version may be enumerated in the fixed section — some may be withheld for a short time to give administrators time to upgrade. [Advisories](https://github.com/open-webui/open-webui/security)
+- 🔐 **Fewer unexpected logouts.** A single request returning an authorization error no longer signs you out while your session is still valid, since the app now confirms the session status before redirecting to login. [Commit](https://github.com/open-webui/open-webui/commit/56ee875e21cb3b137a5aa1eca6eb3c0731b14045)
+- 🔒 **Web search domain filtering.** Domain allow and block rules for web search results are now matched against the host, closing a gap where some URLs could slip past the filter. [Commit](https://github.com/open-webui/open-webui/commit/688bda09fbe26619613aa487beda8059a5fd55ef)
+- 🕵️ **Image prompt log privacy.** Image generation workflows are no longer written to server logs at the default log level, keeping user-authored prompt content out of operator-visible logs. [#26400](https://github.com/open-webui/open-webui/issues/26400), [Commit](https://github.com/open-webui/open-webui/commit/64b92ff08a09b1bf338d5d95b7bf6f9db0d7e857)
+- 🗄️ **Safer database upgrades.** Upgrading an existing SQLite database no longer crashes during the user-table migration or corrupts saved user settings, resolving failures that could block startup or break login after an upgrade. [#26403](https://github.com/open-webui/open-webui/issues/26403), [Commit](https://github.com/open-webui/open-webui/commit/c416c6cad69cc4ce44faacdad87cc65862625fb4)
+- ⚙️ **Saving settings as a non-admin.** Non-admin users can once again save their interface settings, such as the default model and theme, which previously failed with a server error while the interface incorrectly reported success. [#26627](https://github.com/open-webui/open-webui/issues/26627), [Commit](https://github.com/open-webui/open-webui/commit/9866a02863c12c466100aed832f0006225759493)
+- 🕒 **Sidebar chat timestamps.** Chats in the sidebar now show when they were last active instead of when they were created, so the time label matches their position in the list and refreshes after each new message. [#26454](https://github.com/open-webui/open-webui/pull/26454), [#26451](https://github.com/open-webui/open-webui/issues/26451)
+- 🎯 **Default model after refresh.** Your selected model is no longer cleared when you reload the page, as model selection now waits for the model list to finish loading. [Commit](https://github.com/open-webui/open-webui/commit/b6d4baeb7ea9bc83e5486ca98b3cae07d3869aa7)
+- ⏳ **Tool dialogs no longer hang.** Dismissing a tool or function input dialog by clicking outside it now cancels the pending request instead of leaving the chat spinning indefinitely. [#26417](https://github.com/open-webui/open-webui/issues/26417), [Commit](https://github.com/open-webui/open-webui/commit/0016266c0652757e32aa44f2e1fb7311ac08db51)
+- 🐍 **Reliable code execution loading.** Running Python code in chat now loads its runtime reliably, fixing sandbox startup failures that broke code execution and the Pyodide file viewer in recent releases. [#26625](https://github.com/open-webui/open-webui/pull/26625), [#26390](https://github.com/open-webui/open-webui/issues/26390)
+- 🤖 **Models with null capabilities.** Chatting with a model whose capabilities are unset no longer fails with an error when the memory feature or automations are involved. [#26412](https://github.com/open-webui/open-webui/issues/26412), [Commit](https://github.com/open-webui/open-webui/commit/650b81792582268345073ce3757ac14356189d75), [Commit](https://github.com/open-webui/open-webui/commit/0016266c0652757e32aa44f2e1fb7311ac08db51)
+- 🔎 **Searchable responses.** Chat search again finds assistant messages whose text is stored as structured output, which were previously skipped. [#26405](https://github.com/open-webui/open-webui/pull/26405)
+- 🔔 **Chat notification previews.** Background chat completion notifications and toasts now show a clean response preview instead of appearing blank for messages stored as structured output. [Commit](https://github.com/open-webui/open-webui/commit/c98d8ecaccc7284573b050e88b3857c4aeba3860)
+- 💾 **Banner and config startup.** Setting configuration such as "WEBUI_BANNERS" no longer causes a startup failure, since admin configuration values are now stored correctly regardless of their data type. [#26431](https://github.com/open-webui/open-webui/issues/26431), [Commit](https://github.com/open-webui/open-webui/commit/ab22fe64bdd10ba86845dcd46dfd7f619b22366f)
+- 📑 **RAG Template visibility.** The RAG Template editor now stays visible in admin document settings even when Bypass Embedding and Retrieval is enabled, since the template still applies to document content in that mode. [#26126](https://github.com/open-webui/open-webui/issues/26126), [Commit](https://github.com/open-webui/open-webui/commit/8fe480250f649824441a7b4d725a502683134713)
+- 🧬 **Editing derived models.** Editing a workspace model no longer clears its base model, including when that base is a preset or the model itself. [Commit](https://github.com/open-webui/open-webui/commit/092b5857bbda01be53619e26f42b1fa9b64a2b44), [Commit](https://github.com/open-webui/open-webui/commit/54f31c630afa75e9c81acaa048e1bd901254ba92)
+
+### Changed
+
+- ⚠️ **Database Migrations**: This release includes database schema changes; we strongly recommend backing up your database and all associated data before upgrading in production environments. If you are running a multi-worker, multi-server, or load-balanced deployment, all instances must be updated simultaneously, rolling updates are not supported and will cause application failures due to schema incompatibility.
+
+## [0.10.1] - 2026-06-29
+
+### Fixed
+
+- 🤝 **Shared folder read-only chats no longer sign users out.** Opening or reading chats from shared folders now keeps the current session active when a resource-level access error is returned, instead of incorrectly showing "Session expired. Please sign in again."
+
+## [0.10.0] - 2026-06-29
+
+### Added
+
+- 🤝 **Share folders with your team.** You can now share a folder and the chats inside it with specific users, groups, or everyone, with read or write access; people you share with see shared folders in their sidebar and open the chats in a read-only view when they are not the owner, and administrators control who is allowed to share folders with a new "Folders Sharing" permission that is off by default. [Commit](https://github.com/open-webui/open-webui/commit/5019af79a0c45743ede8c9ff37d68f768e7f6174), [Commit](https://github.com/open-webui/open-webui/commit/38920c0ed1f6ad5fe3bb9d12898fa968ead3634a), [Commit](https://github.com/open-webui/open-webui/commit/d65ac445a43348c5f0323d54c37397ae7f483cb8), [Commit](https://github.com/open-webui/open-webui/commit/c783fd30f20d6be5028cf337bc5e5c2f9afbd3f8), [Commit](https://github.com/open-webui/open-webui/commit/45fcf272ef51c84cb01c1454589da3f98e4adc2c), [Commit](https://github.com/open-webui/open-webui/commit/76854d14246660af8222a5302513020e1f36c4f3), [Commit](https://github.com/open-webui/open-webui/commit/084d040e220ee39f62757d928d839646e813fb25), [Commit](https://github.com/open-webui/open-webui/commit/10558173fb155c63403aa8d80f16f8a3ccfa72a6)
+- 🗜️ **Automatic context compaction for long chats.** Conversations that grow past a configurable token threshold can now be summarized automatically so they stay within a model's context window, with a notification shown while it happens; administrators can enable it, set the threshold, customize the summarization prompt, and lower the threshold per model. It is off by default. [Commit](https://github.com/open-webui/open-webui/commit/3f0c0e0a0ddff841b015f96f9649c6999a435c73), [Commit](https://github.com/open-webui/open-webui/commit/7f08376f0c06e1a7fba983a2fa93deb8dfbe7cb0), [Commit](https://github.com/open-webui/open-webui/commit/8934bfb04bf366aece872028944e280c25e36d3e), [#19594](https://github.com/open-webui/open-webui/issues/19594)
+- 🖥️ **Open WebUI Computer agent support.** Open WebUI can now connect to Open WebUI Computer through its OpenAI-compatible gateway, letting chats run full agent sessions on your own machine with file, terminal, git, and web access. [GitHub](https://github.com/open-webui/computer)
+- 🚀 **Much faster hybrid search on large knowledge bases.** Hybrid search now runs natively in the database on pgvector setups instead of loading an entire collection into memory, so querying large knowledge bases is dramatically faster. [Commit](https://github.com/open-webui/open-webui/commit/223f484ded01d092979693341dc03351a9fa17fa), [#20737](https://github.com/open-webui/open-webui/discussions/20737)
+- 🗂️ **External knowledge bases.** Knowledge bases can now be backed by an external retrieval source through configurable external knowledge connections, so you can search an existing external system from chat instead of only Open WebUI's built-in store. [Commit](https://github.com/open-webui/open-webui/commit/15c7e374384488effc3d6059d09b3a8aa79c618d)
+- 🧠 **Reworked memory system.** Memory has been overhauled with distinct memory types — long-lived personal memories and per-conversation context — managed through a structured add, update, and delete flow, giving models a more reliable way to remember and apply what they've learned about you. [Commit](https://github.com/open-webui/open-webui/commit/dbdcfd8c6080c284024052a482e589de226dbf05), [Commit](https://github.com/open-webui/open-webui/commit/7e13fd7ad19c28ba34502bde6c709665f7a6808c), [Commit](https://github.com/open-webui/open-webui/commit/2560533c1a8a2b2a0f7b47becf3703031053ad30), [Commit](https://github.com/open-webui/open-webui/commit/8977a10a2b1e150393635dc5c24e59660d0f2da9), [Commit](https://github.com/open-webui/open-webui/commit/260f3c3a22c55f15ca8f06c5314b23f2a9eb1739), [Commit](https://github.com/open-webui/open-webui/commit/b0487dd6dd942a757828a25aba92e9feef685275), [Commit](https://github.com/open-webui/open-webui/commit/a285a390c12e27e614d1ba9ffb92d1a0e49e7dfa), [Commit](https://github.com/open-webui/open-webui/commit/70e4ffcc6526c1bc90dbfdc0287283574b12c18b), [Commit](https://github.com/open-webui/open-webui/commit/2c4e1fce8f40b0cb5028f1afcb184b6e58c33041), [Commit](https://github.com/open-webui/open-webui/commit/c7e634776d7e77556d149b6cc884ed64363648a9)
+- 🧩 **New plugin primitive: the Event function.** Where pipe, filter, and action functions all run inside a conversation, the new Event function is the first primitive that hooks into the system itself: it runs your own Python in response to events emitted across the whole application — sign-ups, configuration changes, file uploads, role changes, deletions, startup and shutdown, and more. That makes a new class of behavior possible directly inside Open WebUI, from onboarding and access control to auditing, lifecycle automation, and external integrations. Comes with starter boilerplate in the function editor. [Commit](https://github.com/open-webui/open-webui/commit/e124c2656a4c2092b070e570e35b2f0fb7f584de), [Docs](https://docs.openwebui.com/features/extensibility/plugin/functions/event)
+- 🔔 **New event system with webhooks.** Open WebUI now emits events for a wide range of system activity — sign-ins, configuration changes, startup, and actions across chats, knowledge, files, and more. Administrators can send these as outbound webhooks, route them to specific users or groups, and manage which events go where from a new event settings admin page. [Commit](https://github.com/open-webui/open-webui/commit/b5c43968db0ea1556b228d143ae5946dc4e944ba), [Commit](https://github.com/open-webui/open-webui/commit/745396867888718289a2dfcf0809b3e162e00629), [Commit](https://github.com/open-webui/open-webui/commit/5576e6ed8a80a4032b7c6cb3ee0cda0254019355), [Commit](https://github.com/open-webui/open-webui/commit/7b55a63fc7ee323e9114713ce1d2f3f688aa37e6), [Commit](https://github.com/open-webui/open-webui/commit/1a8e1a993928a28b9d814c77b2aef0361b630f27), [Commit](https://github.com/open-webui/open-webui/commit/ede39d82de05eeb7591329679c8cab97753e5ff0), [Commit](https://github.com/open-webui/open-webui/commit/8f890f0b43aed3d42b9e3d954e25e54e37d526d0), [Commit](https://github.com/open-webui/open-webui/commit/741b64edb6c2ff04c4b787528cfca1c5b66a1a27), [Commit](https://github.com/open-webui/open-webui/commit/303c426c3fffafc2369205021a82659ee8715a85), [#1240](https://github.com/open-webui/open-webui/issues/1240), [#16426](https://github.com/open-webui/open-webui/pull/16426)
+- 🔐 **Configure authentication from the admin panel.** LDAP and OAuth/OIDC settings now have a dedicated Authentication settings page, so providers can be configured from the admin interface. [Commit](https://github.com/open-webui/open-webui/commit/5cdcdbaeec9fc8156721c38c33ec37956962871c), [#12945](https://github.com/open-webui/open-webui/pull/12945)
+- 🏷️ **More custom header variables.** Custom request headers now support "{{USER_MESSAGE_ID}}", "{{USER_MESSAGE_PARENT_ID}}", and "{{TASK}}", letting connected services tell apart real user messages from automated background requests like title, tag, and follow-up generation. [Commit](https://github.com/open-webui/open-webui/commit/f85cb27ef835aa76aff7de6176bf2159ba392061)
+- 📄 **File details forwarded to external document extractors.** External custom document-extraction servers now receive the file's ID, name, and content type, and these are also available as custom header variables, so extraction can be tailored per file. [Commit](https://github.com/open-webui/open-webui/commit/b1c2536ed2f8639efade04618018e6de9b332df2), [#26259](https://github.com/open-webui/open-webui/issues/26259)
+- 🎰 **Last model pre-selected for new slots.** When you add another model to a multi-model chat, the slot now defaults to the model you last picked instead of starting empty. [#25974](https://github.com/open-webui/open-webui/pull/25974)
+- ⚡ **Faster model overview.** The admin model overview now loads its feedback history and tags through batched queries, so it opens noticeably faster on instances with many chats. [Commit](https://github.com/open-webui/open-webui/commit/40c09167cd6de1c853a5dd03c88b4fdcb279dfe1)
+- 🏎️ **Lighter channel profile previews.** Profile previews in channels now load a person's details only when you hover to open one, rather than fetching them for every message up front. [Commit](https://github.com/open-webui/open-webui/commit/4f69c33de0e9a8fde4f16d0b2f1ed8aac8741772)
+- ↩️ **Reset permissions to defaults.** The group and default permission dialogs now include a button to restore all permissions back to their built-in defaults in one step. [#25931](https://github.com/open-webui/open-webui/pull/25931)
+- 📥 **Chat import permission.** Administrators can now control whether users are allowed to import or clone chats, with a new "Allow Chat Import" permission. [Commit](https://github.com/open-webui/open-webui/commit/edf3ae920989b01383be543e7379d3eade03c0b6), [Commit](https://github.com/open-webui/open-webui/commit/9ccda6715c3b2dc2cbc1302d2396b2f0233bdea8), [Commit](https://github.com/open-webui/open-webui/commit/ed4cb358a06fc6962b378f20ef846fd2b0af90bc), [#25927](https://github.com/open-webui/open-webui/pull/25927)
+- 🔔 **Per-group user webhook permission.** Administrators can now control which users may set a personal notification webhook, with a new "User Webhooks" permission. [#25923](https://github.com/open-webui/open-webui/pull/25923)
+- ✍️ **Customizable autocomplete prompt.** Administrators can now set a custom prompt template for autocomplete generation from the admin interface settings. [Commit](https://github.com/open-webui/open-webui/commit/4dbb2f94a66d6e0035e2da857ddb6a841a68f862), [#25879](https://github.com/open-webui/open-webui/pull/25879)
+- 🔑 **Configurable secret key length.** The auto-generated secret key length can now be set with a new environment variable, instead of always using a fixed length. [Commit](https://github.com/open-webui/open-webui/commit/e473ab1231abedcb188c259b42ae7f2390739223), [#25906](https://github.com/open-webui/open-webui/pull/25906)
+- 🏟️ **Arena evaluation models configurable via environment.** Arena evaluation models can now be defined through an environment variable, which previously could not be set that way. [Commit](https://github.com/open-webui/open-webui/commit/fd56086e793a0eceb07a55ff972a5492d8f8a285)
+- ✏️ **Edit prompts from the menu.** The prompts list now has an Edit option in each prompt's menu, taking you straight to its editor. [#25789](https://github.com/open-webui/open-webui/pull/25789)
+- 📋 **Clone automations.** Automations now have a Clone option in their menu, so you can duplicate one as a starting point. [#25790](https://github.com/open-webui/open-webui/pull/25790)
+- 🔁 **Recurring calendar events.** The calendar event editor now includes a repeat option, so events can recur on a schedule. [#25865](https://github.com/open-webui/open-webui/pull/25865)
+- 🧷 **Separate skills import and export permissions.** Administrators can now control importing and exporting skills independently, with new skills import and export permissions. [#25921](https://github.com/open-webui/open-webui/pull/25921)
+- 🏷️ **Filter admin models by tag.** The admin Models settings page now has a tag filter for narrowing the model list by base-model tags. [Commit](https://github.com/open-webui/open-webui/commit/2bdd2ab94eefd3d75dd6511c445e302f82221b5d)
+- 📊 **Sortable analytics chat list.** The model chat list in analytics now has sortable column headers, so you can order it by title, last updated, or user. [Commit](https://github.com/open-webui/open-webui/commit/3730a9eaac68dff60b3ae5b4ed160b91480d66eb), [#26168](https://github.com/open-webui/open-webui/pull/26168)
+- 🔐 **Argon2 password hashing option.** Password hashing can now use Argon2 through a configurable algorithm setting, removing the 72-byte password length limit that came with the previous default. [Commit](https://github.com/open-webui/open-webui/commit/33cd199e6dffddd4ee8974af41ebb894871d74c1), [Commit](https://github.com/open-webui/open-webui/commit/a70a6589afad0b429cdd77afa62163391f406a87), [#25656](https://github.com/open-webui/open-webui/pull/25656)
+- 🔐 **Optional encryption of valve values at rest.** Tool and function valve values can now be encrypted at rest through a new opt-in setting, with existing stored values migrated automatically, so sensitive settings like API keys aren't kept in plaintext. [Commit](https://github.com/open-webui/open-webui/commit/b4073f6378392b23a3954e33031bf5e1d98e090a), [#23721](https://github.com/open-webui/open-webui/pull/23721)
+- 🗄️ **AWS RDS IAM database authentication.** The database connection can now authenticate using AWS RDS IAM tokens through a new opt-in setting, instead of only a static password. [Commit](https://github.com/open-webui/open-webui/commit/c0c6c2181a8dc57b62e8a5eabd550bf89db7ffed), [#23580](https://github.com/open-webui/open-webui/pull/23580)
+- 🔓 **Automatic auth for models with OAuth 2.1 tools.** When a model uses tools that require OAuth 2.1, Open WebUI now initiates the authorization flow automatically instead of failing the request. [Commit](https://github.com/open-webui/open-webui/commit/ae5d23f2267845922c2acb507a4a432908d03b41), [#23325](https://github.com/open-webui/open-webui/pull/23325), [#23272](https://github.com/open-webui/open-webui/issues/23272)
+- 🔤 **Custom tokenizer for token-based text splitting.** Token-based document splitting can now use a configurable Hugging Face tokenizer model, so chunking can match the tokenizer of the model you use. [Commit](https://github.com/open-webui/open-webui/commit/bb6b2db88b1e82395531f67db4f6accd49d8b9eb), [#24139](https://github.com/open-webui/open-webui/pull/24139)
+- 🔒 **Restrict OAuth scopes requested from MCP servers.** A new setting lets administrators limit which OAuth scopes Open WebUI requests when connecting to MCP servers. [Commit](https://github.com/open-webui/open-webui/commit/7be009649a0a94008484335c492ab0af18fde41f), [#25981](https://github.com/open-webui/open-webui/pull/25981), [#25978](https://github.com/open-webui/open-webui/issues/25978)
+- 🧩 **Filter Outlet Hook can now run on API requests and responses.** A filter function's outlet hook now runs for direct API callers, including streaming responses, so response post-processing isn't limited to the web interface; this is controlled by a new setting and on by default. [Commit](https://github.com/open-webui/open-webui/commit/390e200f76877b185002c88b4dda27b123d29e83), [#25650](https://github.com/open-webui/open-webui/pull/25650)
+- 🖥️ **Setting for terminal sidebar auto-open.** A new interface setting controls whether the files sidebar opens automatically when you select a terminal. [Commit](https://github.com/open-webui/open-webui/commit/958237473f8cbde97eb0df8c21f2a4de088c4459), [#25628](https://github.com/open-webui/open-webui/pull/25628)
+- 📌 **Reorder pinned notes by dragging.** Pinned notes in the sidebar can now be dragged to reorder them. [#25677](https://github.com/open-webui/open-webui/pull/25677)
+- 🔎 **Chat actions in search.** The search dialog now offers a context menu on each result, so you can act on a chat directly from search. [#25490](https://github.com/open-webui/open-webui/pull/25490)
+- 🔎 **Snippets in chat search results.** Searching your chats now shows a snippet of the matching content in each result, so you can tell results apart at a glance. [Commit](https://github.com/open-webui/open-webui/commit/0eba3df1199f56e8ac77213772a41313a3237296), [Commit](https://github.com/open-webui/open-webui/commit/67a7b23b85d2e3ce6b094b682ba9f07dc453d355), [Commit](https://github.com/open-webui/open-webui/commit/8927c9bb3d4b04f4fc8e443f42089f2a551276bc), [#25178](https://github.com/open-webui/open-webui/pull/25178)
+- 📝 **Formatted valve descriptions.** Valve descriptions for tools and functions now render Markdown, so they can include formatting and links. [Commit](https://github.com/open-webui/open-webui/commit/7c0b0e42f5afb0e9c39bd2d42e4822d64d0c7b3e)
+- 🔽 **Dropdown inputs for valve options.** Valve and confirmation inputs can now present a set of options as a dropdown instead of free text, making fixed-choice settings easier to configure. [Commit](https://github.com/open-webui/open-webui/commit/422a4768ea7428b5dd6d401ccfba00e1e86eb98a), [#26278](https://github.com/open-webui/open-webui/pull/26278)
+- 🔌 **Control the OAuth resource parameter for MCP connectors.** MCP connectors can now be set to always send, never send, or automatically decide whether to include the OAuth resource parameter, so they work with providers that reject it. [Commit](https://github.com/open-webui/open-webui/commit/5576e6ed8a80a4032b7c6cb3ee0cda0254019355)
+- 🔎 **SERPHouse web search.** SERPHouse can now be used as a web search provider. [Commit](https://github.com/open-webui/open-webui/commit/3a232f5e9a4d31a6b74cb34d007b581feeb2f005), [Commit](https://github.com/open-webui/open-webui/commit/dd4f43bfdb793c4276e6468b65fc32d9b881578a), [#26254](https://github.com/open-webui/open-webui/pull/26254)
+- 🔎 **Microsoft Web IQ web search.** Microsoft Web IQ can now be used as a web search provider, with a matching page-browse loader. [#26178](https://github.com/open-webui/open-webui/pull/26178)
+- ⚠️ **Optional web search confirmation.** Administrators can now require users to confirm before a web search runs, with a banner and message making it clear when search is about to be used. [Commit](https://github.com/open-webui/open-webui/commit/fa76764c3b7f99c5adacd34dabd51ead09542c13), [#24942](https://github.com/open-webui/open-webui/pull/24942)
+- 🪪 **Client User-Agent forwarded to model backends.** The browser's User-Agent is now passed through to all model backends, so upstream services can see the originating client. [#26333](https://github.com/open-webui/open-webui/pull/26333)
+- 🖐️ **Drag items from the sidebar into chat.** Folders, notes, and models — including pinned notes — can now be dragged from the sidebar into the chat input. [#25771](https://github.com/open-webui/open-webui/pull/25771), [Commit](https://github.com/open-webui/open-webui/commit/dc1bc41d2e), [#26384](https://github.com/open-webui/open-webui/pull/26384)
+- 🏷️ **Tag suggestions in the model editor.** The model editor now suggests existing tags as you type, making it easier to reuse a consistent set. [Commit](https://github.com/open-webui/open-webui/commit/b58b0ea7ca849b89d217e1077498a8a3fc92471f), [#25703](https://github.com/open-webui/open-webui/pull/25703)
+- 🗣️ **Voice suggestions in the model editor.** The model editor now offers a dropdown of available text-to-speech voices, making it easier to pick one. [Commit](https://github.com/open-webui/open-webui/commit/a5c945940134b957dbad47790b5baafeecdac6c4), [#25706](https://github.com/open-webui/open-webui/pull/25706)
+- 🎛️ **Unified model picker for workspace base model.** Choosing a base model in the model editor now uses the searchable model selector instead of a plain field, making it easier to find and pick the right model. [Commit](https://github.com/open-webui/open-webui/commit/c89fd237b822877bffbb33a37402622983c7189d), [#24576](https://github.com/open-webui/open-webui/issues/24576)
+- 🔍 **Searchable pickers in the model editor.** Attaching actions, filters, tools, knowledge, and skills to a model now uses type-to-search pickers instead of long checkbox lists, making large libraries easier to manage. [Commit](https://github.com/open-webui/open-webui/commit/61cee42ded4e84e31d7cc9b168994ef165efa8ca)
+- 🖼️ **iPhone images work with OpenAI image editing.** Uploaded images are now normalized before being sent to OpenAI image editing, fixing edits that failed for certain iPhone photo formats, with a new admin toggle to control the behavior. [Commit](https://github.com/open-webui/open-webui/commit/39837e0a3afd17b7ff617d97dddf4c5d6446e42d), [Commit](https://github.com/open-webui/open-webui/commit/2d3035a1122123df471ac9d0a591a9465a8212e2), [#26252](https://github.com/open-webui/open-webui/pull/26252), [#26249](https://github.com/open-webui/open-webui/issues/26249)
+- 🟢 **Loaded-model indicator for llama.cpp.** Models served through llama.cpp now report whether they're currently loaded in memory, including the sleeping state, so the loaded indicator works for them too. [Commit](https://github.com/open-webui/open-webui/commit/b696c5deff15d4c85c84c5bac244f062b1bc879a)
+- 🧱 **Structured model output rendered on the client.** Reasoning, tool calls, and server-side tool steps such as web and file search are now rendered in the browser from the model's structured output instead of being flattened into the message text on the server, giving more accurate and editable rendering of these items. [Commit](https://github.com/open-webui/open-webui/commit/0443ab3a61492799f1aaa449f89cbd8aa5912f57), [Commit](https://github.com/open-webui/open-webui/commit/c33fadc26671190c94d86485e6e2ef2f6fd486a3)
+- 📜 **Custom CA bundle for outbound connections.** A new environment variable lets you point Open WebUI at a custom CA certificate bundle, and the per-connection SSL settings now accept a bundle path, so deployments behind a corporate or internal CA can keep certificate verification on instead of disabling it. [Commit](https://github.com/open-webui/open-webui/commit/a54878b14f044d4aa1d8cf5be6f8ce9fc4285438), [Commit](https://github.com/open-webui/open-webui/commit/8b9e28b50354307a314111262f6737b6d8aa4685)
+- 🖥️ **More terminal server orchestrator controls.** Admins connecting an orchestrator terminal server can now configure session lifecycle policies and refresh or reset running terminal sessions, including targeting only idle ones, from the connection settings. [Commit](https://github.com/open-webui/open-webui/commit/7e8153e889a59afe4cf77261ea5e1ef5a66665f1)
+- 📁 **Terminal file browser can stay within a root folder.** The terminal file navigator now anchors to a defined root and home directory, so users can be kept within their workspace instead of browsing into system folders by accident. [Commit](https://github.com/open-webui/open-webui/commit/a0c2ec3d2cf8d696ede479211330eec2da360d39)
+- 🧠 **Memory toggle follows the server default.** When a user hasn't set their own memory preference, it now follows the admin's global memory setting instead of defaulting to off. [#25909](https://github.com/open-webui/open-webui/pull/25909)
+- 🧹 **Unshare all shared chats at once.** The Shared Chats dialog now has a button to stop sharing every shared chat in one action. [#25848](https://github.com/open-webui/open-webui/pull/25848)
+- 📈 **Richer analytics with a date picker.** The analytics dashboard now lets you choose a date range and shows additional columns. [#25922](https://github.com/open-webui/open-webui/pull/25922), [#25919](https://github.com/open-webui/open-webui/issues/25919)
+- 🔢 **Chat and file counts in their dialogs.** The Chats and Files dialogs now show the total number of chats and files in their titles. [#25872](https://github.com/open-webui/open-webui/pull/25872), [#25873](https://github.com/open-webui/open-webui/pull/25873)
+- ⚡ **Faster math rendering.** Rendered math is now cached and reused, so messages with repeated or unchanged math expressions render more efficiently. [#25847](https://github.com/open-webui/open-webui/pull/25847)
+- ⚡ **Lighter Markdown setup.** Markdown extension setup now runs once instead of on every render, avoiding repeated work and extension stacking. [#25837](https://github.com/open-webui/open-webui/pull/25837)
+- ⚡ **Snappier read-only code blocks.** Read-only code blocks now skip language auto-detection, so they render faster. [#25824](https://github.com/open-webui/open-webui/pull/25824)
+- ⚡ **Non-blocking audio model loading.** Loading speech models no longer blocks the server, keeping it responsive while they initialize. [#25806](https://github.com/open-webui/open-webui/pull/25806)
+- ⚡ **Faster URL safety checks.** The safety check on fetched URLs now resolves addresses off the main loop, so it no longer blocks other work. [#25825](https://github.com/open-webui/open-webui/pull/25825)
+- ⚡ **Fewer queries for channel reactions and replies.** Channel reactions and thread replies now load through batched queries, reducing database load on busy channels. [#25831](https://github.com/open-webui/open-webui/pull/25831)
+- ⚡ **Lighter streaming.** Streaming responses now skip re-processing message content that hasn't changed, reducing work on every update. [#26325](https://github.com/open-webui/open-webui/pull/26325), [#26326](https://github.com/open-webui/open-webui/pull/26326)
+- ⚡ **Smoother tool-call rendering.** Displaying tool calls now parses their content iteratively, avoiding slowdowns on deeply nested data. [#26146](https://github.com/open-webui/open-webui/pull/26146)
+- ⚡ **Hidden tool-call details cost nothing.** When tool-call arguments are collapsed, they are no longer rendered behind the scenes, noticeably speeding up chats with heavy tool use. [Commit](https://github.com/open-webui/open-webui/commit/b7934e918223ec0a9e972accd647a8654e503156), [#26147](https://github.com/open-webui/open-webui/pull/26147)
+- ⚡ **Leaner knowledge-file reading for agents.** The built-in tools that let a model read knowledge files now return output in bounded, paginated chunks with a default and a hard cap, instead of potentially returning an entire large file at once, sharply reducing token usage. [Commit](https://github.com/open-webui/open-webui/commit/a285a390c12e27e614d1ba9ffb92d1a0e49e7dfa), [#26139](https://github.com/open-webui/open-webui/issues/26139)
+- ⚡ **Lighter, faster file search on large knowledge bases.** Listing and searching files no longer returns each file's full extracted text by default, and content matching is now length-bounded, so these requests are far lighter and searching across very large knowledge bases is dramatically faster. [Commit](https://github.com/open-webui/open-webui/commit/36d08fa2a7), [Commit](https://github.com/open-webui/open-webui/commit/46c1d6591badb6ab567ba1b8fae23475d5da105a), [Commit](https://github.com/open-webui/open-webui/commit/ab84bbf08c5935f1a19044ef581986f83311da8b), [#25774](https://github.com/open-webui/open-webui/pull/25774), [#25741](https://github.com/open-webui/open-webui/issues/25741), [#26145](https://github.com/open-webui/open-webui/pull/26145), [#25867](https://github.com/open-webui/open-webui/issues/25867)
+- ⚡ **Faster password hashing and bulk user import.** Password hashing and verification no longer block the server, and importing users from a CSV is now processed in a single batch, keeping large imports and sign-ins responsive. [Commit](https://github.com/open-webui/open-webui/commit/6fdf9b4340), [#25804](https://github.com/open-webui/open-webui/pull/25804), [#25805](https://github.com/open-webui/open-webui/pull/25805)
+- ⚡ **Non-blocking model downloads.** Downloading large Ollama models no longer blocks the server on file reads and checksums, keeping it responsive during big downloads. [#25829](https://github.com/open-webui/open-webui/pull/25829)
+- ⚡ **Non-blocking uploads and link fetches.** Hashing uploaded files and fetching URLs now run off the main loop, so large uploads and link previews don't hold up other requests. [#25822](https://github.com/open-webui/open-webui/pull/25822)
+- ⚡ **More blocking work moved off the main loop.** Additional blocking operations in audio, pipelines, and plugin handling now run in worker threads, keeping the server responsive under load. [#26381](https://github.com/open-webui/open-webui/pull/26381)
+- ⚡ **Unreachable backends don't stall model loading.** Loading models and tool servers no longer blocks on backends that are down or slow to respond, so the model list stays responsive when one connection is unreachable. [#26289](https://github.com/open-webui/open-webui/pull/26289)
+- ⚡ **Batched streaming updates.** Streaming responses now group small updates of the same type before sending them, reducing overhead during fast token streams and tool-call output. [Commit](https://github.com/open-webui/open-webui/commit/7240517807a8b0097065f7cbbb384d34084f90fd), [#26202](https://github.com/open-webui/open-webui/pull/26202)
+- 🔄 **General improvements.** Various improvements were implemented across the application to enhance performance, stability, and security.
+- 🌐 **Updated translations.** Catalan, Brazilian Portuguese (pt-BR), Irish, German (de-DE), and Spanish (es-ES) translations were updated.
+
+### Fixed
+
+- 🛡️ **Security Advisory**: This release includes security and access-control fixes. We recommend updating production deployments at your earliest convenience. Not all security fixes in this version may be enumerated in the fixed section — some may be withheld for a short time to give administrators time to upgrade. [Advisories](https://github.com/open-webui/open-webui/security)
+- 🔐 **Knowledge base write access enforced on upload.** Attaching an uploaded file to a knowledge base now requires the same write access as the rest of the knowledge API, so users without write access can no longer add files to a collection by referencing its ID. [#26001](https://github.com/open-webui/open-webui/pull/26001)
+- 🗝️ **API key permission enforced on all key endpoints.** Viewing and deleting API keys now respects the API keys permission, matching the protection already applied to key creation. [#25992](https://github.com/open-webui/open-webui/pull/25992)
+- 🔊 **Text-to-speech permission enforced on the speech endpoint.** The OpenAI speech proxy now honors the text-to-speech permission, so it can no longer be used by people who are not allowed to use that feature. [#25993](https://github.com/open-webui/open-webui/pull/25993)
+- 🎲 **Model access enforced on arena fallback.** Reaching a model indirectly through an arena model on background and task requests now enforces that model's access rules, closing a path that could otherwise bypass them. [#26046](https://github.com/open-webui/open-webui/pull/26046)
+- ⏰ **Scheduled automations stop for deactivated accounts.** Scheduled automations now re-check the owner's account status and permissions before each run, so they stop when an account is deactivated or has automations access revoked. [#26047](https://github.com/open-webui/open-webui/pull/26047)
+- 🚧 **Heavily encoded paths rejected behind the proxy.** Request paths that remain encoded after repeated decoding are now rejected instead of forwarded, preventing a path traversal that could otherwise slip through. [#26050](https://github.com/open-webui/open-webui/pull/26050)
+- 🌐 **Image URL fetches hardened against DNS rebinding.** Fetching user-supplied image URLs now re-checks the destination address at connection time, closing a path that could be used to reach internal addresses behind a public hostname. [#25960](https://github.com/open-webui/open-webui/pull/25960)
+- 🛂 **Web fetch blocklist matches on hostname.** The web fetch filter now matches entries against the request's hostname on domain boundaries, so blocked hosts can no longer slip through with an added path and lookalike domains are no longer mistaken for allowed ones. [#25949](https://github.com/open-webui/open-webui/pull/25949)
+- 🪪 **MCP connectors request least-privilege scopes.** MCP connectors that register dynamically over OAuth now request only the scopes for the specific resource rather than the authorization server's full catalog. [#25958](https://github.com/open-webui/open-webui/pull/25958)
+- 🙈 **Channel member lists no longer expose private data.** Viewing a channel's members now returns only basic profile details, instead of also exposing other members' settings, linked-account data, and personal information. [Commit](https://github.com/open-webui/open-webui/commit/fbcdcf146b99b5002705060a8243eee769108f9e)
+- 🛟 **SCIM sync can't demote an admin.** A SCIM provisioning sync that marks a user inactive can no longer strip an existing administrator's role, preventing an instance from being locked out of its own administration. [#25948](https://github.com/open-webui/open-webui/pull/25948)
+- 👻 **Collaborative notes reject unauthenticated presence events.** The remaining real-time note-collaboration events now require an authenticated session, so presence and cursors can no longer be spoofed by someone who only knows a note's ID. [#25946](https://github.com/open-webui/open-webui/pull/25946)
+- ⏱️ **Login timing no longer reveals which accounts exist.** Sign-in now takes the same amount of time whether or not an account exists, removing a timing difference that could be used to discover valid accounts. [Commit](https://github.com/open-webui/open-webui/commit/993e74912199c66c522f08ec81abe31d76985e39), [Commit](https://github.com/open-webui/open-webui/commit/7b29834d4216e5db70b68f3598fa1ad654d3512b)
+- 🔌 **Terminal connections can't be redirected to another user.** Terminal session identifiers are now safely encoded before being passed upstream, closing a way to tamper with the connection's user identity. [#26042](https://github.com/open-webui/open-webui/pull/26042)
+- 📡 **Real-time events only reach your own session.** The server now verifies that a real-time event is delivered only to the requesting user's own active session, instead of trusting a client-supplied session identifier. [#25763](https://github.com/open-webui/open-webui/pull/25763)
+- 🔓 **Revoked sessions are rejected on real-time connections.** Real-time and terminal WebSocket connections now honor token revocation and expiry, so a signed-out or expired session can no longer keep a live connection open. [Commit](https://github.com/open-webui/open-webui/commit/33b91bd8ae8a100a5a306c91441a7d0b422c4cde), [#25764](https://github.com/open-webui/open-webui/pull/25764), [#25686](https://github.com/open-webui/open-webui/pull/25686)
+- 🕳️ **Another DNS-rebinding gap closed in URL fetching.** Fetching a URL's content now re-checks the destination address at connection time, closing another path that could reach internal addresses behind a public hostname. [#25775](https://github.com/open-webui/open-webui/pull/25775)
+- 🗣️ **Azure speech input is escaped.** Voice and language values are now escaped when building Azure text-to-speech requests, preventing malformed or injected markup. [#25776](https://github.com/open-webui/open-webui/pull/25776)
+- ⚙️ **Interface settings update respects its permission.** Saving interface settings now enforces the interface permission, so users without it can no longer change those settings through the API. [#25996](https://github.com/open-webui/open-webui/pull/25996)
+- 🗄️ **Unknown knowledge collections are denied by default.** Retrieval now rejects unknown or unscoped collection names by default, closing a legacy path that could be used to reach collections outside the normal access checks. [Commit](https://github.com/open-webui/open-webui/commit/d99ac7d3f83b25161ca775229150c8f7c74cceee)
+- 🙈 **Error responses no longer leak internals.** Server error responses now return sanitized messages instead of raw exception text, so internal details aren't exposed to signed-in users. [Commit](https://github.com/open-webui/open-webui/commit/ee5de69e374aabf5631da18a5bbc1c285ee6f7a1), [Commit](https://github.com/open-webui/open-webui/commit/0cc331d1c60341bb06b78ceeecfb2db86179c93e), [Commit](https://github.com/open-webui/open-webui/commit/396d9ac18193d43e40fb9d068075d4b780e971d7), [Commit](https://github.com/open-webui/open-webui/commit/0883638027a9b3cb7c9851f031c4f5fc1af1f25d), [#26375](https://github.com/open-webui/open-webui/pull/26375), [#26374](https://github.com/open-webui/open-webui/issues/26374)
+- 📏 **Upload size limit enforced on the server.** The maximum upload size is now enforced server-side, so it can't be bypassed by a client that ignores the limit. [Commit](https://github.com/open-webui/open-webui/commit/f8ec63203c4408c46bb06698ae624d17b01b9301), [Commit](https://github.com/open-webui/open-webui/commit/d3676b4f71bfdbaf4e4d76943c51117e18932ccf), [#25869](https://github.com/open-webui/open-webui/pull/25869)
+- 🖼️ **OAuth profile pictures are validated.** Profile picture URLs from OAuth providers are now validated and their type checked when stored, preventing unsafe image sources. [Commit](https://github.com/open-webui/open-webui/commit/eb53281c9acb3660e09554a8dbde0a0b42646b70), [#24548](https://github.com/open-webui/open-webui/pull/24548)
+- 📦 **Security updates to frontend dependencies.** Several frontend dependencies were updated to patch known security vulnerabilities. [#26281](https://github.com/open-webui/open-webui/pull/26281)
+- 🤝 **Chat sharing respects the user-sharing permission.** The share-chat dialog now hides the option to share with specific users from people who lack that permission, matching the access rules enforced elsewhere. [#25915](https://github.com/open-webui/open-webui/pull/25915)
+- 📤 **Chat export respects its permission everywhere.** Every chat export menu now checks the export permission, so users without it can no longer export chats through one of the dropdown menus. [#25914](https://github.com/open-webui/open-webui/pull/25914)
+- 📂 **File write access requires real ownership.** Editing or deleting a file through a knowledge base or workspace model now requires that the object's owner actually owns the file, so a read-only file can no longer gain write access by being referenced from an object you control. [#26032](https://github.com/open-webui/open-webui/pull/26032)
+- 🖌️ **Image edit endpoint enforces permission.** The image-edit endpoint now checks the image-edit switch and the image-generation permission, matching image generation, so it can't be called by users who lack access. [#26009](https://github.com/open-webui/open-webui/pull/26009)
+- 📁 **Folder permission enforced on all folder actions.** Every folder operation now checks the folders permission, so the setting is respected consistently instead of only when listing folders. [Commit](https://github.com/open-webui/open-webui/commit/19a176fd36bea15c49d7f2d1539b4832e57a8bc2)
+- 🧩 **Code Execution settings collapse when off.** The Code Execution settings section now collapses when the toggle is disabled, keeping the settings page tidy. [#25970](https://github.com/open-webui/open-webui/pull/25970)
+- 📅 **German date format in Notes.** Dates in the Notes view now display correctly for German, where they previously failed to render. [#25985](https://github.com/open-webui/open-webui/pull/25985)
+- 🎙️ **ElevenLabs speech keeps working when voices can't load.** Text-to-speech through ElevenLabs no longer fails when the available-voice list can't be fetched, instead of rejecting every voice. [Commit](https://github.com/open-webui/open-webui/commit/bb1419328b11b801b4c939dfc112700ba6f6fdab), [#26075](https://github.com/open-webui/open-webui/issues/26075)
+- 🪟 **Default Permissions modal resets on close.** Closing the Default Permissions dialog without saving now discards unsaved edits instead of keeping them around the next time you open it. [Commit](https://github.com/open-webui/open-webui/commit/78a5015846a9e55ff2bc9d6cc98f880437abe8ed)
+- 👯 **Side-by-side chat with the same model.** Running two panes with the same model no longer leaves one pane stuck waiting or showing the other pane's reply after a reload, since each pane's messages are now tracked separately. [Commit](https://github.com/open-webui/open-webui/commit/56ae99e96a845289b5787d2dd26a3d828f2295e7), [#25982](https://github.com/open-webui/open-webui/issues/25982)
+- 💾 **Model edits no longer lost when changing access.** Adjusting a model's access no longer auto-saves on its own and discards your other unsaved changes to that model. [#26004](https://github.com/open-webui/open-webui/pull/26004)
+- 🔧 **Parallel tool calls over the Anthropic-compatible API.** External Anthropic-compatible clients calling Open WebUI's messages endpoint now receive tool calls reliably when a model issues several at once or returns them in its final message. [Commit](https://github.com/open-webui/open-webui/commit/4210cae68e30173d7902582d32128dd699d5628a), [#25963](https://github.com/open-webui/open-webui/pull/25963), [#25964](https://github.com/open-webui/open-webui/discussions/25964)
+- 🗃️ **Prompt caching preserved over the Anthropic-compatible API.** Requests through the Anthropic-compatible API now keep their prompt-caching markers instead of having them stripped, so clients that rely on caching work as intended. [Commit](https://github.com/open-webui/open-webui/commit/caedcbae4988ef59ea7052b2a3198e2da4b5291a), [#25998](https://github.com/open-webui/open-webui/pull/25998), [#25964](https://github.com/open-webui/open-webui/discussions/25964)
+- 🔁 **Fewer redundant data loads.** Several views no longer fire duplicate background fetches at once, avoiding occasional glitches from overlapping requests. [#25943](https://github.com/open-webui/open-webui/pull/25943), [#25942](https://github.com/open-webui/open-webui/pull/25942), [#25934](https://github.com/open-webui/open-webui/pull/25934), [#25935](https://github.com/open-webui/open-webui/pull/25935), [#25838](https://github.com/open-webui/open-webui/pull/25838), [Commit](https://github.com/open-webui/open-webui/commit/e8d55c0a8beac9de0b2a0fe90f0bc0f9b64c1c1f)
+- 🔎 **Steadier search boxes across admin and workspace.** Search fields for users, knowledge, prompts, tools, and similar lists now run only as you type and reset to the first page correctly, instead of occasionally re-searching on their own. [Commit](https://github.com/open-webui/open-webui/commit/fc9c2ea1915accd1f6edca467e965283dff71cd7), [#25938](https://github.com/open-webui/open-webui/pull/25938)
+- 📊 **Admin feedback list loads again on PostgreSQL.** The admin feedback list no longer fails to load on PostgreSQL setups, where it previously returned a server error. [Commit](https://github.com/open-webui/open-webui/commit/7ee75a0c04a31528954903e88c9213d5fbb31aa7), [#25953](https://github.com/open-webui/open-webui/issues/25953)
+- 🗂️ **Deleting nested folders checks chats correctly.** Deleting a folder that contains subfolders now accounts for the chats inside those subfolders when applying the delete-permission check, instead of only the top-level folder's chats. [Commit](https://github.com/open-webui/open-webui/commit/232421f40b84590e6d6fdecab4e43274aac37add), [#25920](https://github.com/open-webui/open-webui/issues/25920)
+- 🖱️ **Dragging chats into folders is more reliable.** Dragging a chat into a folder no longer throws an error in cases where the chat couldn't be resolved. [#25928](https://github.com/open-webui/open-webui/pull/25928)
+- 🛠️ **Workspace menu shows for the skills permission.** Users who only have the skills permission now see the Workspace entry in their menu, which previously appeared only for other workspace permissions. [#25925](https://github.com/open-webui/open-webui/pull/25925)
+- 🧠 **Admins can always reach memories.** Administrators can now use the memories endpoints regardless of the memories permission toggle, matching how admin access works for other features. [#25924](https://github.com/open-webui/open-webui/pull/25924)
+- 🖼️ **Image settings page survives a config load failure.** The admin image settings page no longer crashes when its configuration fails to load, showing the page instead. [#25933](https://github.com/open-webui/open-webui/pull/25933)
+- 🧵 **Code blocks render in channel threads.** Code blocks now display correctly in a channel's thread view, where duplicated message identifiers previously broke their rendering. [Commit](https://github.com/open-webui/open-webui/commit/7d1f9415807a47e0da4f862327e9802a3b839753), [#25917](https://github.com/open-webui/open-webui/pull/25917)
+- 🔵 **No more false unread badges on chats.** Chats no longer show an unread indicator after automatic changes like title generation or pinning, archiving, and moving them between folders, and newly created chats are marked read correctly so they don't appear unread after a refresh. [#25912](https://github.com/open-webui/open-webui/pull/25912), [#25782](https://github.com/open-webui/open-webui/pull/25782), [#25108](https://github.com/open-webui/open-webui/issues/25108)
+- 📌 **Pinned notes stay in sync.** Pinning, unpinning, or deleting a note now updates the sidebar's pinned list consistently, instead of showing a stale pin state. [#25918](https://github.com/open-webui/open-webui/pull/25918), [#25640](https://github.com/open-webui/open-webui/pull/25640)
+- 📅 **All-day calendar events keep their date.** Saving an all-day calendar event no longer shifts it by a day for users in certain time zones. [#25864](https://github.com/open-webui/open-webui/pull/25864)
+- 🧷 **Damaged chat history recovers more reliably.** When a chat's current position is missing or points at a malformed message, Open WebUI now repairs it from the latest valid message — on both the client and the server — instead of risking a broken history view. [Commit](https://github.com/open-webui/open-webui/commit/2308b59f135e4c2da11eabdf2306e55a5dd4e9fb), [Commit](https://github.com/open-webui/open-webui/commit/a146e17bdcaf94fee3a98aa36b4f401e1f06c1d4), [#26298](https://github.com/open-webui/open-webui/pull/26298), [#26258](https://github.com/open-webui/open-webui/pull/26258), [#26257](https://github.com/open-webui/open-webui/issues/26257)
+- 💾 **Saving a chat no longer drops messages.** Chat updates are now merged with the existing history on the server, with explicit tracking of deleted messages, instead of overwriting it, preventing message loss from concurrent or partial saves. [Commit](https://github.com/open-webui/open-webui/commit/22a44e67a8ba781feb8f2a267fed0c40213d8432), [Commit](https://github.com/open-webui/open-webui/commit/3319b6410e1b600b7a885a5fb78573e9a2061c22), [Commit](https://github.com/open-webui/open-webui/commit/24b8619f64731788ac38813768abbb64405effa4), [#25657](https://github.com/open-webui/open-webui/pull/25657)
+- 📺 **Channel message updates stay in their channel.** Streaming updates to a channel message are now skipped if the message no longer exists or belongs to a different channel, preventing stray updates. [Commit](https://github.com/open-webui/open-webui/commit/ac3449cac91e62b08a7c28e54fcd044d14dea791)
+- 📌 **Pinned channel messages update for everyone.** Pinning or unpinning a channel message now updates live for all members and works from thread views, instead of only changing for the person who pinned it. [Commit](https://github.com/open-webui/open-webui/commit/7ea7680f563da30b121258e5a7d7123185c4da2a)
+- 📄 **Mistral OCR uploads work again.** Document OCR through Mistral has been repaired after an upstream library change broke its file uploads. [#25779](https://github.com/open-webui/open-webui/pull/25779)
+- 🗂️ **Chroma collection detection fixed.** Open WebUI now correctly detects existing Chroma collections, fixing a case where it always reported them as missing. [#25780](https://github.com/open-webui/open-webui/pull/25780)
+- 📊 **Vega-Lite charts render reliably.** Vega-Lite charts in chat are now detected by their code block language tag, so they render correctly. [#25843](https://github.com/open-webui/open-webui/pull/25843)
+- 🏷️ **Long chat tag lists scroll.** The tags section in the chat menu now scrolls instead of overflowing when a chat has many tags. [#26031](https://github.com/open-webui/open-webui/pull/26031)
+- ⌨️ **Enter key shows correctly on iOS.** The Enter key symbol in the keyboard shortcuts list no longer renders as an emoji on iOS. [#26173](https://github.com/open-webui/open-webui/pull/26173)
+- 🔗 **Whitespace in names no longer breaks MCP connections.** User name and info headers are now trimmed before being forwarded, fixing MCP connection failures when a display name contained leading or trailing whitespace. [#26182](https://github.com/open-webui/open-webui/pull/26182), [#26181](https://github.com/open-webui/open-webui/issues/26181)
+- 🈳 **Search no longer fires mid-composition.** Typing in search with an input method editor (such as Japanese, Chinese, or Korean) no longer triggers a search when you press Enter to confirm a composition. [#26238](https://github.com/open-webui/open-webui/pull/26238), [#26285](https://github.com/open-webui/open-webui/pull/26285), [#26172](https://github.com/open-webui/open-webui/issues/26172)
+- 🧰 **Valves icon stays visible.** The icon for configuring valves no longer disappears, so user-configurable tool and function settings remain reachable. [#26256](https://github.com/open-webui/open-webui/pull/26256)
+- 🎛️ **Chat controls persist across navigation.** Edits to chat controls are now kept when navigating between chats, and reverting a control to the chat's saved value persists correctly, instead of being lost. [#26336](https://github.com/open-webui/open-webui/pull/26336), [#25793](https://github.com/open-webui/open-webui/pull/25793)
+- 🔍 **Chat search tool handles empty queries.** The built-in chat search tool no longer crashes when called with an empty query. [Commit](https://github.com/open-webui/open-webui/commit/b854eb09b13216f914ce5fd07ab717b8f752882b), [#26310](https://github.com/open-webui/open-webui/issues/26310)
+- 📑 **More robust MinerU document processing.** Document processing through MinerU now handles its ZIP results more safely, including very large outputs. [Commit](https://github.com/open-webui/open-webui/commit/23d03d6aaebcced6c1e39e98dfff76ab73df8804), [#26263](https://github.com/open-webui/open-webui/pull/26263)
+- ⏰ **Scheduled automations with session-auth tools work.** Automations that use session-authenticated tools or terminals now authenticate correctly when running on a schedule, instead of failing. [Commit](https://github.com/open-webui/open-webui/commit/5b1c42e81a3ef3ad5ce5852dbf84020cb5e2498c), [#26247](https://github.com/open-webui/open-webui/pull/26247), [#26137](https://github.com/open-webui/open-webui/issues/26137)
+- 📝 **Model system prompt preserved with knowledge.** A model's system prompt is no longer dropped when knowledge retrieval runs with native tool calling. [Commit](https://github.com/open-webui/open-webui/commit/cfb49c4c181a96d5df07fbfacd819639baef0bab), [#26217](https://github.com/open-webui/open-webui/pull/26217)
+- 🔑 **Expired sessions return you to sign-in.** When a request fails because your session has expired, Open WebUI now redirects you to the sign-in page instead of leaving you on a broken view. [Commit](https://github.com/open-webui/open-webui/commit/5922727402593900758d84004f950071c701f6de), [#26237](https://github.com/open-webui/open-webui/pull/26237)
+- 🎯 **Ejecting a workspace model unloads the right model.** Unloading a workspace model now resolves to its underlying base model, so the correct model is freed from memory. [Commit](https://github.com/open-webui/open-webui/commit/464e703e4716812d015966582152ddd8a2c71572), [#26269](https://github.com/open-webui/open-webui/pull/26269)
+- 🔄 **Edited models refresh in the admin list.** After editing a model in the admin settings, the models list now updates right away instead of needing a manual reload. [Commit](https://github.com/open-webui/open-webui/commit/b34d6c836ee43d0e9721fa4fd6457d934e3e2a17)
+- 🗂️ **Workspace model bulk actions and search work across pages.** Bulk actions on workspace models now apply across all of them, and search results paginate correctly. [#26274](https://github.com/open-webui/open-webui/pull/26274)
+- 🧩 **MCP resource results come through.** Tool results that return resource content — including binary blobs and URI references — are no longer silently dropped, and image results are attached as files. [#25260](https://github.com/open-webui/open-webui/pull/25260), [#24038](https://github.com/open-webui/open-webui/issues/24038), [Commit](https://github.com/open-webui/open-webui/commit/783205a965c556815fae84b64d74f26a2e5e5729)
+- 🔗 **Broader MCP server compatibility for OAuth.** Open WebUI now discovers an MCP server's protected resource metadata even when the server doesn't advertise it, and recognizes more OAuth preflight variations, so more MCP servers connect. [#25980](https://github.com/open-webui/open-webui/pull/25980), [#25954](https://github.com/open-webui/open-webui/issues/25954), [Commit](https://github.com/open-webui/open-webui/commit/45fea34bd0c8ce54b0822499c40e3e6964220354), [#26068](https://github.com/open-webui/open-webui/pull/26068)
+- 📤 **Clearer upload error messages.** Failed uploads now show a readable explanation instead of an opaque error stub. [#25961](https://github.com/open-webui/open-webui/pull/25961)
+- 📋 **Cloned prompts get a proper title.** Cloning a prompt now adds the clone suffix to the correct field, so the duplicate is named as expected. [#25800](https://github.com/open-webui/open-webui/pull/25800)
+- 📐 **Long default group names don't overflow.** A long default group name no longer overflows its row in the admin authentication settings. [#25685](https://github.com/open-webui/open-webui/pull/25685)
+- 🖐️ **Sidebar drags don't trigger uploads.** Dragging a chat item in the sidebar no longer shows the file-upload overlay. [#25675](https://github.com/open-webui/open-webui/pull/25675)
+- 🔁 **Recovers from a stuck streaming response.** If the signal that a response finished is missed — for example after a mobile app is backgrounded mid-stream — Open WebUI now recovers the chat instead of leaving it stuck in a streaming state. [Commit](https://github.com/open-webui/open-webui/commit/aa851d93c63e7da6e94292d0b7586674339d47e5), [Commit](https://github.com/open-webui/open-webui/commit/edf2c6c8f76e7f6a5917e991f371a604adc34c5f), [Commit](https://github.com/open-webui/open-webui/commit/2856def6c05b2fb8c55b4e7170f05db0c4f956f1), [#26320](https://github.com/open-webui/open-webui/pull/26320), [#26315](https://github.com/open-webui/open-webui/issues/26315)
+- 🧠 **Model skills load on demand instead of filling the prompt.** A model's attached skills are now presented to the model as a manifest it can load when needed, rather than having their full content inserted into the system prompt; skills you mention inline still get their content included directly. [Commit](https://github.com/open-webui/open-webui/commit/e6d35fc4cca4f4b1e5cad97d7b7e3089ef832018), [Commit](https://github.com/open-webui/open-webui/commit/44b9463498085741669e6f5d92e21b5ecc5fd795), [#25592](https://github.com/open-webui/open-webui/issues/25592), [#25599](https://github.com/open-webui/open-webui/pull/25599)
+- 🗂️ **Empty metadata no longer breaks Chroma indexing.** Document metadata with empty values is now filtered out before indexing, fixing a case that could fail on Chroma. [Commit](https://github.com/open-webui/open-webui/commit/118549caf3), [#26342](https://github.com/open-webui/open-webui/pull/26342), [#26339](https://github.com/open-webui/open-webui/issues/26339)
+- 🔁 **Updating a knowledge file won't break the knowledge base.** When a file's content is updated, its new embeddings are now added before the old ones are removed, so a failed reindex leaves the knowledge base intact and usable instead of empty. [Commit](https://github.com/open-webui/open-webui/commit/248315de14d4537e0f2ec3f94dee8a7334cad248), [#23789](https://github.com/open-webui/open-webui/pull/23789), [#23787](https://github.com/open-webui/open-webui/issues/23787)
+- 🔤 **Documents with special tokens index correctly.** Measuring chunk sizes no longer fails when a document contains text that looks like a special token. [#26210](https://github.com/open-webui/open-webui/pull/26210)
+- 📝 **Note file attachments stay in sync.** Updating the files attached to a note now keeps the editor and saved note in sync. [Commit](https://github.com/open-webui/open-webui/commit/5055fb85aa8c8d5ef785daea7438498e36ddf33f)
+- 📱 **Better banner layout on mobile.** Notification banners now lay out correctly on small screens. [Commit](https://github.com/open-webui/open-webui/commit/4ed45ce84394c435405f93d03f07dabd797cdec3), [#24912](https://github.com/open-webui/open-webui/pull/24912)
+- 📂 **Knowledge file listing includes attached files.** Listing files through the knowledge tools now also shows files attached directly to a model, not only those inside a knowledge base, fixing cases where listing returned no results for a model with a single attached file. [Commit](https://github.com/open-webui/open-webui/commit/40b655e99e2c6dd802654ec0cdac38a4bcda08b3), [#26301](https://github.com/open-webui/open-webui/issues/26301)
+- 🏷️ **Chat titles generate after long first responses.** A new chat now gets its title even when the first response takes a long time, such as one with extensive reasoning or many tool calls, instead of staying "New Chat". [Commit](https://github.com/open-webui/open-webui/commit/754787f43dffad3dce2c90e4fd0417b1f9dbb3c0), [#26240](https://github.com/open-webui/open-webui/issues/26240)
+- 🔌 **Cancelling an MCP request no longer errors.** Stopping a response that was using MCP tools now shuts the connection down cleanly instead of surfacing a server error. [Commit](https://github.com/open-webui/open-webui/commit/ff5cec43bd360829cfdcc6a5253d1ba63f236b7f)
+- 🧠 **Reasoning details preserved across turns.** Models that return structured or encrypted reasoning data, such as Gemini, no longer have their assistant message split mid-stream, keeping reasoning continuity across turns. [Commit](https://github.com/open-webui/open-webui/commit/75db531c1238af113bb2b211882713e5e2f459cf), [#23852](https://github.com/open-webui/open-webui/pull/23852)
+- 📡 **Error messages show for non-standard streaming responses.** Providers that send errors over non-standard server-sent events now surface a readable error instead of nothing. [#23228](https://github.com/open-webui/open-webui/pull/23228)
+- 🔑 **Whitespace in terminal server keys no longer breaks auth.** Terminal server API keys are now trimmed before use, so a key with stray leading or trailing whitespace still authenticates. [Commit](https://github.com/open-webui/open-webui/commit/fe3300bd6581aa469c2cdf757700ecc65a200df4), [Commit](https://github.com/open-webui/open-webui/commit/d6cda4a04b2e3a48855fc91abb2376cfd3a0378d)
+- 🔥 **One bad URL no longer fails Firecrawl scraping.** When fetching multiple pages through Firecrawl, a single failing URL is now skipped instead of aborting the whole batch, and rate limits are respected between requests. [Commit](https://github.com/open-webui/open-webui/commit/6f8221df58b17334233ac6bfe069b8f837f677d6), [#24183](https://github.com/open-webui/open-webui/pull/24183)
+- 📱 **Usable chat input on mobile with many tools.** When skills, tools, terminal, web search, and image generation buttons fill the chat input, the row of buttons now scrolls horizontally while the menu, voice, and send controls stay reachable, instead of pushing them off-screen. [Commit](https://github.com/open-webui/open-webui/commit/6f8221df58b17334233ac6bfe069b8f837f677d6), [#26142](https://github.com/open-webui/open-webui/issues/26142)
+- 👤 **Owner avatars only show on shared folders.** Chat owner avatars in a folder's chat list now appear only when the folder is actually shared, instead of showing whenever owner information happened to be present. [Commit](https://github.com/open-webui/open-webui/commit/9802b0d13563b3535b86a350bab000d84686b1e9)
+- 📜 **No stray scrollbar on the About page.** Extra spacing that caused an unnecessary scrollbar on the About settings page has been removed. [#25802](https://github.com/open-webui/open-webui/pull/25802)
+- 🚪 **Sign out works from the Account Pending page.** Signing out while your account is pending now goes through the proper sign-out flow, so single sign-on sessions are ended and you are no longer left stuck on the pending screen. [#25681](https://github.com/open-webui/open-webui/pull/25681), [#25644](https://github.com/open-webui/open-webui/issues/25644)
+- 🔢 **Built-in tools accept numeric arguments.** Built-in tools no longer crash when a model passes a number or a string where a specific scalar type is expected; values are now coerced to the declared type. [Commit](https://github.com/open-webui/open-webui/commit/c4688b958d7f7929f5f4303493ca311c2c121683), [#25638](https://github.com/open-webui/open-webui/pull/25638), [#25731](https://github.com/open-webui/open-webui/pull/25731), [#25641](https://github.com/open-webui/open-webui/issues/25641)
+- ⏱️ **MinerU timeout saves.** The MinerU API timeout can now be saved from the admin settings, accepting a numeric value. [Commit](https://github.com/open-webui/open-webui/commit/3fd0384ffcd0eddd6f4c688475f8ad5d3b4de510), [#25604](https://github.com/open-webui/open-webui/pull/25604), [#25603](https://github.com/open-webui/open-webui/issues/25603)
+- 🔧 **Background completion no longer clears active tasks.** Finishing a chat in the background no longer wipes the set of active tasks, fixing a case where ongoing task indicators could be lost. [Commit](https://github.com/open-webui/open-webui/commit/388f62f8a002b789887d016892a1bf152c9d90af), [#25217](https://github.com/open-webui/open-webui/issues/25217)
+- 👁️ **Workspace base model selector respects visibility.** The base model selector in the workspace now hides models you don't have access to, matching their visibility settings. [#25668](https://github.com/open-webui/open-webui/pull/25668)
+- 🧵 **Channel threads bind to the right channel.** A channel thread's parent and replies are now tied to the channel in the URL, preventing mismatches when switching channels. [#25766](https://github.com/open-webui/open-webui/pull/25766)
+- 🗑️ **Unsharing cleans up orphaned rows.** Unsharing a chat now handles leftover shared-chat records, avoiding stale entries. [#25632](https://github.com/open-webui/open-webui/pull/25632)
+- 🔎 **Web search results reach the model with retrieval on.** Web search results are now passed to the model even when embedding and retrieval are enabled, instead of being left out. [#25600](https://github.com/open-webui/open-webui/pull/25600)
+- 🔢 **Group count follows search.** The groups count now reflects the filtered search results instead of the full list. [#25689](https://github.com/open-webui/open-webui/pull/25689)
+- ␣ **Space key works when renaming.** Pressing space while renaming a file or folder no longer opens it, so spaces can be typed in names. [#25627](https://github.com/open-webui/open-webui/pull/25627)
+- 🩹 **Missing local embedding model no longer blocks startup.** A missing local embedding model now surfaces as a deferred error instead of preventing the server from starting. [#25683](https://github.com/open-webui/open-webui/pull/25683)
+- 🔤 **Consistent settings label capitalization.** Toggle labels in settings now use consistent title casing. [#25765](https://github.com/open-webui/open-webui/pull/25765)
+- ♿ **Better screen-reader labels on toggles.** Integration and switch toggles now expose proper accessibility labels and pressed state for screen readers. [#25258](https://github.com/open-webui/open-webui/pull/25258), [#25230](https://github.com/open-webui/open-webui/pull/25230)
+- 📜 **Long dropdowns scroll.** Dropdown selects now scroll when their list is long, so all options stay reachable. [Commit](https://github.com/open-webui/open-webui/commit/4bc463072185d0d7c1c9218cd4487501090eccfe), [#25608](https://github.com/open-webui/open-webui/pull/25608)
+- 🔽 **Collapsible sections don't misfire on load.** Collapsible sections no longer trigger their change action when first rendered, avoiding unintended toggles on page load. [Commit](https://github.com/open-webui/open-webui/commit/c93d4f04aad1b0d4f8a8bda7ac403b2c8ee35f38), [#25229](https://github.com/open-webui/open-webui/pull/25229)
+- ➗ **Large math expressions no longer crash rendering.** Parsing math delimiters no longer overflows on very large or deeply nested input, so messages with heavy math render instead of failing. [#25845](https://github.com/open-webui/open-webui/pull/25845)
+- 🗄️ **Oversized chunks no longer break Milvus indexing.** Overly long text chunks are now trimmed before being sent to Milvus, so a single large chunk can no longer fail the whole batch and leave a file with no embeddings. [#25857](https://github.com/open-webui/open-webui/pull/25857), [#25858](https://github.com/open-webui/open-webui/pull/25858)
+- 📝 **Code editor stays open when empty.** The code editor drawer no longer collapses when its content is empty. [#25855](https://github.com/open-webui/open-webui/pull/25855)
+- 💽 **Settings no longer lost after a restart.** Admin configuration is now stored more reliably, fixing cases where external connections and model parameters could be lost after restarting the server. [Commit](https://github.com/open-webui/open-webui/commit/5cdcdbaeec9fc8156721c38c33ec37956962871c), [Commit](https://github.com/open-webui/open-webui/commit/21f9e5295bf484169d72f4538f7c926b5519723c), [Commit](https://github.com/open-webui/open-webui/commit/8958b64b5a7e96cd8c2260571b54324ca3bfe127), [#24743](https://github.com/open-webui/open-webui/issues/24743), [#25911](https://github.com/open-webui/open-webui/pull/25911), [#25959](https://github.com/open-webui/open-webui/pull/25959)
+- 📜 **Visible chat scrollbar.** The chat area now shows a scrollbar, making it easier to scroll through long responses. [Commit](https://github.com/open-webui/open-webui/commit/d56e1cb0b9), [#25833](https://github.com/open-webui/open-webui/issues/25833)
+- 🎚️ **Default model parameters apply to requests.** Default model parameters are now applied to outbound requests, so settings like temperature and the context window take effect as configured. [Commit](https://github.com/open-webui/open-webui/commit/cd6cc39c6d), [Commit](https://github.com/open-webui/open-webui/commit/19db873603215773f9a64e03785a1a076dc6c8a8), [#24930](https://github.com/open-webui/open-webui/issues/24930), [#26209](https://github.com/open-webui/open-webui/issues/26209)
+- 🟢 **Ollama loaded-model indicator restored.** The indicator showing which Ollama model is loaded in VRAM works again after recent changes. [#25586](https://github.com/open-webui/open-webui/issues/25586), [#25732](https://github.com/open-webui/open-webui/issues/25732)
+- 🪪 **Static MCP connectors recover missing OAuth details.** MCP connectors configured with static OAuth credentials now fill in a missing scope or resource from the server's published metadata, so they connect correctly instead of failing when those values were left out. [Commit](https://github.com/open-webui/open-webui/commit/88901bfa041ddcceab1cd4a97f08f0b43835eb05), [#25898](https://github.com/open-webui/open-webui/issues/25898)
+- 📊 **Token usage and cost stats no longer wiped by background tasks.** A response's token usage and cost are now preserved when background tasks like title, tag, and follow-up generation run on the same chat, instead of being overwritten. [Commit](https://github.com/open-webui/open-webui/commit/95391221dfabbfcd9090ab472b1c02a4c75c0387)
+- 🔗 **Model share link updated.** Sharing a model now opens the current community post page, fixing the link that pointed at the old endpoint. [#25801](https://github.com/open-webui/open-webui/pull/25801)
+
+### Changed
+
+- ⚠️ **Database Migrations**: This update contains database migrations. Please be sure to back up your database before updating, as downgrading after the migration is not supported.
+- 🔔 **System events now fire automatically.** With the new event system, Open WebUI emits events for activity like startup, sign-ins, and configuration changes, so any webhook you already have configured may begin receiving calls for these newly emitted events after upgrading. Review your event and webhook settings after updating so you only receive the events you want. [Commit](https://github.com/open-webui/open-webui/commit/b5c43968db0ea1556b228d143ae5946dc4e944ba)
+- 🔀 **Native tool calling is now the default.** Every chat and model that had not explicitly chosen a tool-calling mode now runs Native, which relies on a model's built-in tool support, while the old behavior has been renamed "Legacy" and made the explicit opt-out; if your models depend on the previous approach you must switch them back to "Legacy" per chat, per model, or globally in your default model parameters to preserve their behavior. [Commit](https://github.com/open-webui/open-webui/commit/b1d40f340921c27eb9a965b9feeb2563856e25e2)
+- 🗂️ **Authentication settings moved to their own page.** LDAP, OAuth, and related authentication settings have moved out of the General settings page into a dedicated Authentication page in the admin panel. [Commit](https://github.com/open-webui/open-webui/commit/5cdcdbaeec9fc8156721c38c33ec37956962871c)
+- 🎓 **Several features are no longer beta.** Memories, Notes, Channels, and High Contrast Mode have graduated out of beta and no longer carry a beta label. [Commit](https://github.com/open-webui/open-webui/commit/7b55a63fc7ee323e9114713ce1d2f3f688aa37e6)
+- 🔧 **Local web fetch setting renamed.** The "ENABLE_RAG_LOCAL_WEB_FETCH" environment variable is now "ENABLE_LOCAL_WEB_FETCH", reflecting that it applies beyond retrieval; the old name still works as a deprecated alias. [Commit](https://github.com/open-webui/open-webui/commit/e3ba6984534898695b47ee4fc3d6b746e2865abc)
+- 🔧 **You.com search key renamed.** You.com web search now prefers the "YDC_API_KEY" environment variable, with the previous "YOUCOM_API_KEY" still accepted as a fallback. [Commit](https://github.com/open-webui/open-webui/commit/df634bb64f5043b0292e43c69bd1d31676c89328), [#26316](https://github.com/open-webui/open-webui/pull/26316)
+- 🧪 **Client-side Python now runs sandboxed.** Client-side Python (Pyodide) now runs in a sandboxed, opaque-origin iframe by default, isolating executed code from your session, cookies, local storage, and the app's own endpoints, while full Python, JavaScript, and external network access keep working. Code that relied on reaching same-origin Open WebUI endpoints from Pyodide will no longer be able to, and Pyodide is now marked legacy in the admin Code Execution settings. [Commit](https://github.com/open-webui/open-webui/commit/516051304e1b1f250c34438746ade673a79bd40c), [Commit](https://github.com/open-webui/open-webui/commit/c7be66626fd10c75ec35f662a709129ba1b020ec), [Commit](https://github.com/open-webui/open-webui/commit/62ae2069183109d878d72b9444a0e7c4f6c66caa), [Commit](https://github.com/open-webui/open-webui/commit/518702caae5a6484e71aa79e8ab908ec398290a7), [Commit](https://github.com/open-webui/open-webui/commit/03a8363583b7e0e04760d49f1e8d28dbbfefee4d)
+
 ## [0.9.6] - 2026-06-01
 
 ### Added

Dockerfile

@@ -126,7 +126,7 @@ RUN chown -R $UID:$GID /app $HOME
 # Install common system dependencies
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
-    git build-essential pandoc gcc netcat-openbsd curl jq \
+    git build-essential pandoc gcc netcat-openbsd curl jq ca-certificates \
     libmariadb-dev \
     python3-dev \
     ffmpeg libsm6 libxext6 zstd \
@@ -184,6 +184,17 @@ COPY --chown=$UID:$GID --from=build /app/package.json /app/package.json
 # copy backend files
 COPY --chown=$UID:$GID ./backend .
 
+# The backend rewrites its bundled static assets (favicons, splash, manifest,
+# loader.js, ...) under open_webui/static at startup. Make that directory
+# writable by an arbitrary UID -- which under OpenShift's restricted SCC is
+# always a member of GID 0 -- so those writes don't fail with EACCES and crash
+# the boot log with "[Errno 13] Permission denied". `chmod -R g=u` mirrors the
+# owner bits onto the group (the Red Hat arbitrary-UID idiom). This is applied
+# unconditionally because it targets a directory the app writes on every start;
+# the broader, opt-in USE_PERMISSION_HARDENING below covers the rest of /app.
+RUN chgrp -R 0 /app/backend/open_webui/static && \
+    chmod -R g=u /app/backend/open_webui/static
+
 EXPOSE 8080
 
 HEALTHCHECK CMD curl --silent --fail http://localhost:${PORT:-8080}/health | jq -ne 'input.status == true' || exit 1

README.md

@@ -8,7 +8,7 @@
 ![GitHub top language](https://img.shields.io/github/languages/top/open-webui/open-webui)
 ![GitHub last commit](https://img.shields.io/github/last-commit/open-webui/open-webui?color=red)
 [![Discord](https://img.shields.io/badge/Discord-Open_WebUI-blue?logo=discord&logoColor=white)](https://discord.gg/5rJgQTnV4s)
-[![](https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&color=%23fe8e86)](https://github.com/sponsors/tjbck)
+[![](https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&color=%23fe8e86)](https://github.com/sponsors/open-webui)
 
 ![Open WebUI Banner](./banner.png)
 
@@ -27,58 +27,82 @@ For more information, be sure to check out our [Open WebUI Documentation](https:
 
 ## Key Features of Open WebUI ⭐
 
-- 🚀 **Effortless Setup**: Install seamlessly using Docker or Kubernetes (kubectl, kustomize or helm) for a hassle-free experience with support for both `:ollama` and `:cuda` tagged images.
+- 🚀 **Effortless Setup**: Install seamlessly via pip, uv, Docker, or Kubernetes (kubectl, kustomize, or helm), with `:ollama` and `:cuda` tagged images available for container deployments.
 
-- 🤝 **Ollama/OpenAI API Integration**: Effortlessly integrate OpenAI-compatible APIs for versatile conversations alongside Ollama models. Customize the OpenAI API URL to link with **LMStudio, GroqCloud, Mistral, OpenRouter, and more**.
+- 🤝 **Broad Model & API Integration**: Connect any OpenAI-compatible API alongside local Ollama models. Point the API URL at **LMStudio, GroqCloud, Mistral, OpenRouter, vLLM, and more** to mix and match providers freely.
 
-- 🛡️ **Granular Permissions and User Groups**: By allowing administrators to create detailed user roles and permissions, we ensure a secure user environment. This granularity not only enhances security but also allows for customized user experiences, fostering a sense of ownership and responsibility amongst users.
+- 🔐 **Granular RBAC & User Groups**: Administrators define detailed roles, groups, and permissions, giving each user exactly the access they need. Secure by default, with tailored experiences per group.
 
-- 📱 **Responsive Design**: Enjoy a seamless experience across Desktop PC, Laptop, and Mobile devices.
+- 🧩 **Plugin Support**: Extend Open WebUI with **Filters**, **Actions**, **Pipes**, **Tools**, and **Skills**. Connect external services through **MCP**, **MCPO**, and **OpenAPI tool servers**. Build custom integrations, rate limits, approval flows, data connections, and more.
 
-- 📱 **Progressive Web App (PWA) for Mobile**: Enjoy a native app-like experience on your mobile device with our PWA, providing offline access on localhost and a seamless user interface.
+- 🤖 **Models & Agents**: Wrap any base model with custom instructions, tools, and knowledge to build specialized agents. Supports dynamic variables, per-user/group access control, and community preset imports via [Open WebUI Community](https://openwebui.com/).
 
-- ✒️🔢 **Full Markdown and LaTeX Support**: Elevate your LLM experience with comprehensive Markdown and LaTeX capabilities for enriched interaction.
+- 📝 **Notes**: A dedicated workspace for content outside conversations. Draft with a rich editor, use AI to rewrite selected text, and attach notes to any chat for full-context injection.
 
-- 🎤📹 **Hands-Free Voice/Video Call**: Experience seamless communication with integrated hands-free voice and video call features using multiple Speech-to-Text providers (Local Whisper, OpenAI, Deepgram, Azure) and Text-to-Speech engines (Azure, ElevenLabs, OpenAI, Transformers, WebAPI), allowing for dynamic and interactive chat environments.
+- 📢 **Channels**: Real-time shared spaces where your team and AI models collaborate in one timeline. Tag models to draft or critique, with threads, reactions, pins, and access control.
 
-- 🛠️ **Model Builder**: Easily create Ollama models via the Web UI. Create and add custom characters/agents, customize chat elements, and import models effortlessly through [Open WebUI Community](https://openwebui.com/) integration.
+- 🧠 **Persistent Memory**: The AI remembers facts about you across conversations, carrying context from one chat to the next.
 
-- 🐍 **Native Python Function Calling Tool**: Enhance your LLMs with built-in code editor support in the tools workspace. Bring Your Own Function (BYOF) by simply adding your pure Python functions, enabling seamless integration with LLMs.
+- ✅ **Live Workflow & Message Flow**: Watch the AI build and work through checklists in real time. Queue messages while the AI is still responding; they send automatically when it's ready.
 
-- 💾 **Persistent Artifact Storage**: Built-in key-value storage API for artifacts, enabling features like journals, trackers, leaderboards, and collaborative tools with both personal and shared data scopes across sessions.
+- 📅 **Calendar & AI Scheduling**: Built-in personal and shared calendars with month/week/day views, recurring events, color coding, attendees, and reminders. Models manage your schedule conversationally through native function calling.
 
-- 📚 **Local RAG Integration**: Dive into the future of chat interactions with groundbreaking Retrieval Augmented Generation (RAG) support using your choice of 9 vector databases and multiple content extraction engines (Tika, Docling, Document Intelligence, Mistral OCR, PaddleOCR-vl, External loaders). Load documents directly into chat or add files to your document library, effortlessly accessing them using the `#` command before a query.
+- ⏱️ **Automations**: Schedule prompts to run on recurring schedules, with runs surfaced on your calendar and each completed run linking back to the chat it produced.
 
-- 🔍 **Web Search for RAG**: Perform web searches using 15+ providers including `SearXNG`, `Google PSE`, `Brave Search`, `Kagi`, `Mojeek`, `Tavily`, `Perplexity`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `SearchApi`, `SerpApi`, `Bing`, `Jina`, `Exa`, `Sougou`, `Azure AI Search`, and `Ollama Cloud`, injecting results directly into your chat experience.
+- 📱 **Responsive Design & PWA**: Seamless experience across desktop, laptop, and mobile, with a Progressive Web App for native app-like feel and offline access on localhost.
 
-- 🌐 **Web Browsing Capability**: Seamlessly integrate websites into your chat experience using the `#` command followed by a URL. This feature allows you to incorporate web content directly into your conversations, enhancing the richness and depth of your interactions.
+- ✒️🔢 **Full Markdown and LaTeX Support**: Comprehensive Markdown and LaTeX capabilities for enriched interaction.
 
-- 🎨 **Image Generation & Editing Integration**: Create and edit images using multiple engines including OpenAI's DALL-E, Gemini, ComfyUI (local), and AUTOMATIC1111 (local), with support for both generation and prompt-based editing workflows.
+- 🎤📹 **Hands-Free Voice/Video Call**: Integrated voice and video calls with multiple Speech-to-Text providers (Local Whisper, OpenAI, Deepgram, Azure) and Text-to-Speech engines (Azure, ElevenLabs, OpenAI, Transformers, WebAPI).
 
-- ⚙️ **Many Models Conversations**: Effortlessly engage with various models simultaneously, harnessing their unique strengths for optimal responses. Enhance your experience by leveraging a diverse set of models in parallel.
+- 💾 **Persistent Artifact Storage**: Built-in key-value storage API for artifacts, enabling journals, trackers, leaderboards, and collaborative tools with personal and shared data scopes.
 
-- 🔐 **Role-Based Access Control (RBAC)**: Ensure secure access with restricted permissions; only authorized individuals can access your Ollama, and exclusive model creation/pulling rights are reserved for administrators.
+- 📚 **Local RAG Integration**: Retrieval Augmented Generation backed by 9 vector databases and multiple content-extraction engines (Tika, Docling, Document Intelligence, Mistral OCR, PaddleOCR-vl, external loaders). Supports hybrid search (BM25 + vector) with reranking and full-context mode. Load documents into chat or pull them from your library with the `#` command.
 
-- 🗄️ **Flexible Database & Storage Options**: Choose from SQLite (with optional encryption), PostgreSQL, or configure cloud storage backends (S3, Google Cloud Storage, Azure Blob Storage) for scalable deployments.
+- 🔍 **Web Search for RAG**: Search the web through dozens of providers including `SearXNG`, `Google PSE`, `Brave Search`, `Kagi`, `Mojeek`, `Tavily`, `Perplexity`, `Firecrawl`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `SearchApi`, `SerpApi`, `Bing`, `Jina`, `Exa`, `Sougou`, `Azure AI Search`, and `Ollama Cloud`, injecting results directly into the conversation.
 
-- 🔍 **Advanced Vector Database Support**: Select from 9 vector database options including ChromaDB, PGVector, Qdrant, Milvus, Elasticsearch, OpenSearch, Pinecone, S3Vector, and Oracle 23ai for optimal RAG performance.
+- 🌐 **Web Browsing Capability**: Pull websites into chat with the `#` command followed by a URL, or let the model fetch them on its own when needed.
 
-- 🔐 **Enterprise Authentication**: Full support for LDAP/Active Directory integration, SCIM 2.0 automated provisioning, and SSO via trusted headers alongside OAuth providers. Enterprise-grade user and group provisioning through SCIM 2.0 protocol, enabling seamless integration with identity providers like Okta, Azure AD, and Google Workspace for automated user lifecycle management.
+- 🎨 **Image Generation & Editing**: Create and edit images with multiple engines including OpenAI DALL·E, Gemini, ComfyUI (local), and AUTOMATIC1111 (local), supporting both generation and prompt-based editing.
 
-- ☁️ **Cloud-Native Integration**: Native support for Google Drive and OneDrive/SharePoint file picking, enabling seamless document import from enterprise cloud storage.
+- ⚙️ **Multi-Model Conversations**: Engage several models at once, harnessing their individual strengths in parallel for the best possible responses.
 
-- 📊 **Production Observability**: Built-in OpenTelemetry support for traces, metrics, and logs, enabling comprehensive monitoring with your existing observability stack.
+- 📊 **Usage Analytics & Model Evaluation**: Admin dashboards track message volume, token consumption, and cost across users and models. Evaluate models with a built-in arena, A/B testing, and ELO-based leaderboards.
 
-- ⚖️ **Horizontal Scalability**: Redis-backed session management and WebSocket support for multi-worker and multi-node deployments behind load balancers.
+- 🗄️ **Flexible Database & Storage**: Choose SQLite (with optional encryption) or PostgreSQL, and store files locally or on S3, Google Cloud Storage, or Azure Blob Storage.
 
-- 🌐🌍 **Multilingual Support**: Experience Open WebUI in your preferred language with our internationalization (i18n) support. Join us in expanding our supported languages! We're actively seeking contributors!
+- 🧬 **Advanced Vector Database Support**: Pick from 9 vector databases: ChromaDB, PGVector, Qdrant, Milvus, Elasticsearch, OpenSearch, Pinecone, S3Vector, and Oracle 23ai.
 
-- 🧩 **Pipelines, Open WebUI Plugin Support**: Seamlessly integrate custom logic and Python libraries into Open WebUI using [Pipelines Plugin Framework](https://github.com/open-webui/pipelines). Launch your Pipelines instance, set the OpenAI URL to the Pipelines URL, and explore endless possibilities. [Examples](https://github.com/open-webui/pipelines/tree/main/examples) include **Function Calling**, User **Rate Limiting** to control access, **Usage Monitoring** with tools like Langfuse, **Live Translation with LibreTranslate** for multilingual support, **Toxic Message Filtering** and much more.
+- 🪪 **Enterprise Authentication & Provisioning**: Full LDAP/Active Directory integration, SSO via trusted headers and OAuth providers, and SCIM 2.0 automated provisioning for identity providers like Okta, Azure AD, and Google Workspace.
 
-- 🌟 **Continuous Updates**: We are committed to improving Open WebUI with regular updates, fixes, and new features.
+- ☁️ **Cloud-Native File Integration**: Native Google Drive and OneDrive/SharePoint file picking for seamless document import from enterprise cloud storage.
+
+- 🔭 **Production Observability**: Built-in OpenTelemetry support for traces, metrics, and logs, plugging into your existing monitoring stack.
+
+- ⚖️ **Horizontal Scalability**: Redis-backed session management and WebSocket support for multi-worker, multi-node deployments behind load balancers.
+
+- 🌐🌍 **Multilingual Support**: Use Open WebUI in your preferred language with i18n support. We're actively seeking contributors to expand language coverage!
+
+- 🌟 **Continuous Updates**: We're committed to improving Open WebUI with regular updates, fixes, and new features.
+
+- 🛡️ **Transparent Security Process**: Security reports are triaged, fixed, and published as open advisories through a documented responsible-disclosure process. See our [Security Policy](https://github.com/open-webui/open-webui/security).
 
 Want to learn more about Open WebUI's features? Check out our [Open WebUI documentation](https://docs.openwebui.com/features) for a comprehensive overview!
 
+## The Open WebUI Ecosystem 🌐
+
+Open WebUI is the core, surrounded by companion apps and infrastructure that extend what your AI can do, where it can reach, and how you run it:
+
+- 💻 **Open WebUI Computer** ([open-webui/computer](https://github.com/open-webui/computer)): A standalone, mobile-first computer and coding agent that runs on the machine you own. Files, terminal, and git in a browser tab, reachable from your phone. Connect it into Open WebUI as a model, or reach it from Telegram, WhatsApp, and more.
+
+- ⚡ **Open Terminal** and **Terminals (Enterprise)** ([open-webui/open-terminal](https://github.com/open-webui/open-terminal) & [open-webui/terminals](https://github.com/open-webui/terminals)): A self-hosted computing environment that plugs into Open WebUI, giving the AI a place to write code, run it, read output, fix errors, and iterate inside the chat. Terminals gives you per-user isolated containers with separate credentials, resource limits, and network rules. Automatic lifecycle management on Docker or Kubernetes.
+
+- 🔄 **oikb** ([open-webui/oikb](https://github.com/open-webui/oikb)): Feed your Knowledge Bases from 45+ sources (GitHub, Confluence, ServiceNow, Salesforce, Jira, Slack, SharePoint, Notion, and more), keeping the tools your team already uses continuously in sync.
+
+- 🖥️ **Native Desktop App** ([open-webui/desktop](https://github.com/open-webui/desktop)): Run Open WebUI as a native app on macOS, Windows, and Linux. System-wide Spotlight chat bar with screenshot capture, push-to-talk voice, and optional fully-local inference via a built-in llama.cpp engine.
+
+Want to learn more? Check out our [Open WebUI documentation](https://docs.openwebui.com) for more details!
+
 ---
 
 We are incredibly grateful for the generous support of our sponsors. Their contributions help us to maintain and improve our project, ensuring we can continue to deliver quality work to our community. Thank you!
@@ -222,6 +246,10 @@ This project contains code under multiple licenses. The current codebase include
 If you have any questions, suggestions, or need assistance, please open an issue or join our
 [Open WebUI Discord community](https://discord.gg/5rJgQTnV4s) to connect with us! 🤝
 
+## Security 🛡️
+
+If you believe you've found a security vulnerability, or something that shouldn't be disclosed publicly, please [reach out confidentially through our responsible disclosure program on GitHub](https://github.com/open-webui/open-webui/security). We accept reports only through GitHub, not through any other platform. Thank you for helping us keep Open WebUI secure!
+
 ## Star History
 
 <a href="https://star-history.com/#open-webui/open-webui&Date">

backend/open_webui/__init__.py

@@ -11,6 +11,7 @@ import uvicorn
 app = typer.Typer()
 
 KEY_FILE = Path.cwd() / '.webui_secret_key'
+DEFAULT_SECRET_KEY_LENGTH = 24
 
 
 def version_callback(value: bool) -> None:
@@ -37,8 +38,11 @@ def serve(
     if os.getenv('WEBUI_SECRET_KEY') is None:
         typer.echo('Loading WEBUI_SECRET_KEY from file, not provided as an environment variable.')
         if not KEY_FILE.exists():
+            key_length = int(os.getenv('WEBUI_SECRET_KEY_LENGTH', DEFAULT_SECRET_KEY_LENGTH))
+            if key_length < 1:
+                raise ValueError('WEBUI_SECRET_KEY_LENGTH must be a positive integer')
             typer.echo(f'Generating a new secret key and saving it to {KEY_FILE}')
-            KEY_FILE.write_bytes(base64.b64encode(random.randbytes(12)))
+            KEY_FILE.write_bytes(base64.b64encode(random.randbytes(key_length)))
         typer.echo(f'Loading WEBUI_SECRET_KEY from {KEY_FILE}')
         os.environ['WEBUI_SECRET_KEY'] = KEY_FILE.read_text()
 

backend/open_webui/constants.py

@@ -1,8 +1,29 @@
 from __future__ import annotations
 
+import errno
 from enum import Enum
 
 
+_ERRNO_MESSAGES = {
+    errno.ENAMETOOLONG: 'File name is too long.',
+    errno.ENOSPC: 'The server is out of storage space.',
+    errno.EDQUOT: 'Server storage quota exceeded.',
+    errno.EACCES: 'Server storage is not writable.',
+    errno.EPERM: 'Server storage is not writable.',
+    errno.EROFS: 'Server storage is not writable.',
+}
+
+
+def _error_message(err='', fallback='') -> str:
+    if not err:
+        return 'Something went wrong :/'
+    if isinstance(err, OSError) and err.errno in _ERRNO_MESSAGES:
+        return f'[ERROR: {_ERRNO_MESSAGES[err.errno]}]'
+    if isinstance(err, Exception):
+        return f'[ERROR: {fallback}]' if fallback else 'Something went wrong :/'
+    return f'[ERROR: {err}]'
+
+
 class MESSAGES(str, Enum):
     DEFAULT = lambda msg='': f'{msg if msg else ""}'
     MODEL_ADDED = lambda model='': f"The model '{model}' has been added successfully."
@@ -18,7 +39,7 @@ class ERROR_MESSAGES(str, Enum):
     def __str__(self) -> str:
         return super().__str__()
 
-    DEFAULT = lambda err='': f'{"Something went wrong :/" if err == "" else "[ERROR: " + str(err) + "]"}'
+    DEFAULT = _error_message
     ENV_VAR_NOT_FOUND = 'Required environment variable not found. Terminating now.'
     CREATE_USER_ERROR = 'Oops! Something went wrong while creating your account. Please try again later. If the issue persists, contact support for assistance.'
     DELETE_USER_ERROR = 'Oops! Something went wrong. We encountered an issue while trying to delete the user. Please give it another shot.'

backend/open_webui/env.py

@@ -102,6 +102,7 @@ class JSONFormatter(logging.Formatter):
 
 
 LOG_FORMAT = os.getenv('LOG_FORMAT', '').lower()
+LOGURU_DIAGNOSE = os.getenv('LOGURU_DIAGNOSE', 'False').lower() == 'true'
 
 GLOBAL_LOG_LEVEL = os.getenv('GLOBAL_LOG_LEVEL', '').upper()
 if GLOBAL_LOG_LEVEL in logging.getLevelNamesMapping():
@@ -149,6 +150,11 @@ INSTANCE_ID = os.getenv('INSTANCE_ID', str(uuid4()))
 
 ENABLE_DB_MIGRATIONS = os.getenv('ENABLE_DB_MIGRATIONS', 'True').lower() == 'true'
 
+# Swap the JSON encoder/decoder used across the app (HTTP request bodies, JSONResponse
+# bodies, upstream provider responses, socket.io payloads) from the stdlib `json` module
+# to orjson. Faster, but stricter: see open_webui/utils/json_codec.py for the differences.
+ENABLE_ORJSON = os.getenv('ENABLE_ORJSON', 'False').lower() == 'true'
+
 
 # Function to parse each section
 def parse_section(section):
@@ -291,6 +297,7 @@ if 'postgres://' in DATABASE_URL:
     DATABASE_URL = DATABASE_URL.replace('postgres://', 'postgresql://')
 
 DATABASE_SCHEMA = os.getenv('DATABASE_SCHEMA', None)
+DATABASE_ENABLE_IAM_TOKEN_AUTH = os.getenv('DATABASE_ENABLE_IAM_TOKEN_AUTH', 'False').lower() == 'true'
 
 _pool_size_raw = os.getenv('DATABASE_POOL_SIZE')
 try:
@@ -388,6 +395,12 @@ try:
 except ValueError:
     REDIS_SOCKET_CONNECT_TIMEOUT = None
 
+REDIS_SOCKET_TIMEOUT = os.getenv('REDIS_SOCKET_TIMEOUT', '')
+try:
+    REDIS_SOCKET_TIMEOUT = float(REDIS_SOCKET_TIMEOUT)
+except ValueError:
+    REDIS_SOCKET_TIMEOUT = None
+
 # Whether to enable TCP SO_KEEPALIVE on Redis client sockets. Opt-in:
 # defaults to off so behavior is unchanged for existing deployments. When
 # enabled, the kernel sends TCP keepalive probes on idle connections so
@@ -443,17 +456,16 @@ WEBSOCKET_REDIS_OPTIONS = os.getenv('WEBSOCKET_REDIS_OPTIONS', '')
 
 
 if WEBSOCKET_REDIS_OPTIONS == '':
+    WEBSOCKET_REDIS_OPTIONS = {'socket_timeout': None}
     if REDIS_SOCKET_CONNECT_TIMEOUT:
-        WEBSOCKET_REDIS_OPTIONS = {'socket_connect_timeout': REDIS_SOCKET_CONNECT_TIMEOUT}
-    else:
-        log.debug('No WEBSOCKET_REDIS_OPTIONS provided, defaulting to None')
-        WEBSOCKET_REDIS_OPTIONS = None
+        WEBSOCKET_REDIS_OPTIONS['socket_connect_timeout'] = REDIS_SOCKET_CONNECT_TIMEOUT
 else:
     try:
         WEBSOCKET_REDIS_OPTIONS = json.loads(WEBSOCKET_REDIS_OPTIONS)
+        WEBSOCKET_REDIS_OPTIONS.setdefault('socket_timeout', None)
     except Exception:
-        log.warning('Invalid WEBSOCKET_REDIS_OPTIONS, defaulting to None')
-        WEBSOCKET_REDIS_OPTIONS = None
+        log.warning('Invalid WEBSOCKET_REDIS_OPTIONS, defaulting to socket_timeout=None')
+        WEBSOCKET_REDIS_OPTIONS = {'socket_timeout': None}
 
 WEBSOCKET_REDIS_URL = os.getenv('WEBSOCKET_REDIS_URL', REDIS_URL)
 WEBSOCKET_REDIS_CLUSTER = os.getenv('WEBSOCKET_REDIS_CLUSTER', str(REDIS_CLUSTER)).lower() == 'true'
@@ -498,6 +510,65 @@ else:
         WEBSOCKET_EVENT_CALLER_TIMEOUT = 300
 
 
+import ssl as _ssl
+
+
+# Dedicated env var for a custom CA bundle file path.  When set, this is
+# used as the default CA bundle for all outbound HTTPS connections that
+# have SSL verification enabled (i.e. when their per-connection SSL env
+# var is ``"True"``).  Per-connection overrides (setting the SSL env var
+# to a path directly) take precedence over this global fallback.
+#
+# This follows the industry convention of ``SSL_CERT_FILE`` / ``REQUESTS_CA_BUNDLE``
+# but is scoped to Open WebUI to avoid interfering with system-level settings.
+AIOHTTP_CLIENT_SSL_CERT_FILE = os.getenv('AIOHTTP_CLIENT_SSL_CERT_FILE', '').strip()
+
+
+def _build_ssl_context_from_file(path: str) -> '_ssl.SSLContext | None':
+    """Create an SSLContext from a CA bundle file, or None if invalid."""
+    if not path:
+        return None
+    if not os.path.isfile(path):
+        log.warning(
+            'SSL CA bundle path does not exist: %r, ignoring',
+            path,
+        )
+        return None
+    ctx = _ssl.create_default_context(cafile=path)
+    log.info('Using custom SSL CA bundle: %s', path)
+    return ctx
+
+
+# Pre-built SSLContext from the dedicated env var (cached once at startup).
+_GLOBAL_SSL_CONTEXT = _build_ssl_context_from_file(AIOHTTP_CLIENT_SSL_CERT_FILE)
+
+
+def _parse_ssl_env(value: str) -> 'bool | _ssl.SSLContext':
+    """Parse an SSL env var into a bool or SSLContext.
+
+    - ``"true"``  → uses ``AIOHTTP_CLIENT_SSL_CERT_FILE`` context if set,
+      otherwise ``True``  (default SSL verification via certifi)
+    - ``"false"`` → ``False`` (no verification)
+    - ``"/path/to/ca-bundle.crt"`` → ``SSLContext`` loading that CA file
+      (takes precedence over ``AIOHTTP_CLIENT_SSL_CERT_FILE``)
+
+    This allows users with corporate or internal CAs to point Open WebUI
+    at a custom CA bundle without disabling verification entirely.
+    """
+    lower = value.strip().lower()
+    if lower == 'true':
+        # Use the global dedicated CA bundle if configured, otherwise default
+        return _GLOBAL_SSL_CONTEXT if _GLOBAL_SSL_CONTEXT is not None else True
+    if lower == 'false':
+        return False
+    # Treat as a file path to a CA bundle (per-connection override)
+    ctx = _build_ssl_context_from_file(value.strip())
+    if ctx is not None:
+        return ctx
+    # Path was invalid — fall back to default
+    return _GLOBAL_SSL_CONTEXT if _GLOBAL_SSL_CONTEXT is not None else True
+
+
 REQUESTS_VERIFY = os.getenv('REQUESTS_VERIFY', 'True').lower() == 'true'
 
 _aiohttp_timeout_raw = os.getenv('AIOHTTP_CLIENT_TIMEOUT', '')
@@ -506,8 +577,27 @@ try:
 except (ValueError, TypeError):
     AIOHTTP_CLIENT_TIMEOUT = 300
 
+# Optional between-chunks idle cap for streaming aiohttp requests.
+AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT = os.getenv('AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT', '')
+if AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT == '':
+    AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT = None
+else:
+    try:
+        AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT = int(AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT)
+    except (ValueError, TypeError):
+        AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT = None
 
-AIOHTTP_CLIENT_SESSION_SSL = os.getenv('AIOHTTP_CLIENT_SESSION_SSL', 'True').lower() == 'true'
+if AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT is not None and AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT <= 0:
+    AIOHTTP_CLIENT_STREAM_IDLE_TIMEOUT = None
+
+
+# SSL verification for general outbound requests (OpenAI, OAuth, etc.).
+# Accepts "True", "False", or a path to a CA bundle file.
+# When "True", falls back to AIOHTTP_CLIENT_SSL_CERT_FILE if set.
+AIOHTTP_CLIENT_SESSION_SSL = _parse_ssl_env(os.getenv('AIOHTTP_CLIENT_SESSION_SSL', 'True'))
+
+SEARXNG_CLIENT_CERT_FILE = os.getenv('SEARXNG_CLIENT_CERT_FILE', '').strip()
+SEARXNG_CLIENT_KEY_FILE = os.getenv('SEARXNG_CLIENT_KEY_FILE', '').strip()
 
 # When False (default), outbound HTTP requests do not follow 3xx redirects.
 AIOHTTP_CLIENT_ALLOW_REDIRECTS = os.getenv('AIOHTTP_CLIENT_ALLOW_REDIRECTS', 'False').lower() == 'true'
@@ -532,8 +622,20 @@ try:
 except (ValueError, TypeError):
     AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER_DATA = 10
 
+AIOHTTP_FILE_STREAM_CHUNK_SIZE = os.getenv('AIOHTTP_FILE_STREAM_CHUNK_SIZE', str(1024 * 1024))
+try:
+    AIOHTTP_FILE_STREAM_CHUNK_SIZE = int(AIOHTTP_FILE_STREAM_CHUNK_SIZE)
+except Exception:
+    AIOHTTP_FILE_STREAM_CHUNK_SIZE = 1024 * 1024
 
-AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL = os.getenv('AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL', 'True').lower() == 'true'
+if AIOHTTP_FILE_STREAM_CHUNK_SIZE <= 0:
+    AIOHTTP_FILE_STREAM_CHUNK_SIZE = 1024 * 1024
+
+
+# SSL verification for tool server connections specifically.
+# Accepts "True", "False", or a path to a CA bundle file.
+# When "True", falls back to AIOHTTP_CLIENT_SSL_CERT_FILE if set.
+AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL = _parse_ssl_env(os.getenv('AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL', 'True'))
 
 AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER = os.getenv('AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER', '')
 
@@ -616,6 +718,8 @@ WEBUI_SECRET_KEY = os.getenv(
     os.getenv('WEBUI_JWT_SECRET_KEY', ''),
 )
 
+ENABLE_VALVE_ENCRYPTION = os.getenv('ENABLE_VALVE_ENCRYPTION', 'False').lower() == 'true'
+
 WEBUI_SESSION_COOKIE_SAME_SITE = os.getenv('WEBUI_SESSION_COOKIE_SAME_SITE', 'lax')
 WEBUI_SESSION_COOKIE_SECURE = os.getenv('WEBUI_SESSION_COOKIE_SECURE', 'false').lower() == 'true'
 WEBUI_AUTH_COOKIE_SAME_SITE = os.getenv('WEBUI_AUTH_COOKIE_SAME_SITE', WEBUI_SESSION_COOKIE_SAME_SITE)
@@ -662,6 +766,7 @@ WEBUI_AUTH_TRUSTED_ROLE_HEADER = os.getenv('WEBUI_AUTH_TRUSTED_ROLE_HEADER', Non
 CUSTOM_API_KEY_HEADER = os.getenv('CUSTOM_API_KEY_HEADER', 'x-api-key')
 
 ENABLE_PASSWORD_VALIDATION = os.getenv('ENABLE_PASSWORD_VALIDATION', 'False').lower() == 'true'
+PASSWORD_HASH_ALGORITHM = os.getenv('PASSWORD_HASH_ALGORITHM', 'bcrypt').lower()
 PASSWORD_VALIDATION_REGEX_PATTERN = os.getenv(
     'PASSWORD_VALIDATION_REGEX_PATTERN',
     r'^(?=.*[a-z])(?=.*[A-Z])(?=.*\d)(?=.*[^\w\s]).{8,}$',
@@ -686,6 +791,9 @@ BYPASS_RETRIEVAL_ACCESS_CONTROL = os.getenv('BYPASS_RETRIEVAL_ACCESS_CONTROL', '
 # for non-admin users.  When False (default), unknown collection names are
 # denied — closing the legacy unscoped namespace.
 ENABLE_RETRIEVAL_UNSCOPED_COLLECTIONS = os.getenv('ENABLE_RETRIEVAL_UNSCOPED_COLLECTIONS', 'False').lower() == 'true'
+MINERU_MAX_MARKDOWN_BYTES = (
+    int(os.getenv('MINERU_MAX_MARKDOWN_BYTES')) if os.getenv('MINERU_MAX_MARKDOWN_BYTES') else None
+)
 
 # When enabled, skips pydub-based preprocessing (format conversion, compression,
 # and chunked splitting) before sending files to processing engines. Useful when
@@ -717,6 +825,18 @@ OAUTH_MAX_SESSIONS_PER_USER = int(os.getenv('OAUTH_MAX_SESSIONS_PER_USER', '10')
 # Token Exchange Configuration
 # Allows external apps to exchange OAuth tokens for OpenWebUI tokens
 ENABLE_OAUTH_TOKEN_EXCHANGE = os.getenv('ENABLE_OAUTH_TOKEN_EXCHANGE', 'False').lower() == 'true'
+_oauth_token_exchange_rate_limit = (os.getenv('OAUTH_TOKEN_EXCHANGE_RATE_LIMIT') or '').strip()
+OAUTH_TOKEN_EXCHANGE_RATE_LIMIT = (
+    int(_oauth_token_exchange_rate_limit)
+    if _oauth_token_exchange_rate_limit and _oauth_token_exchange_rate_limit.lower() != 'none'
+    else None
+)
+OAUTH_TOKEN_EXCHANGE_RATE_LIMIT_WINDOW = int(os.getenv('OAUTH_TOKEN_EXCHANGE_RATE_LIMIT_WINDOW', str(60 * 3)))
+OAUTH_TOKEN_EXCHANGE_TRUSTED_CLIENT_IDS = [
+    client_id.strip()
+    for client_id in os.getenv('OAUTH_TOKEN_EXCHANGE_TRUSTED_CLIENT_IDS', '').split(',')
+    if client_id.strip()
+]
 
 # Back-Channel Logout Configuration
 # When enabled, exposes POST /oauth/backchannel-logout for IdP-initiated logout
@@ -862,6 +982,7 @@ else:
 ENABLE_CHAT_RESPONSE_BASE64_IMAGE_URL_CONVERSION = (
     os.getenv('ENABLE_CHAT_RESPONSE_BASE64_IMAGE_URL_CONVERSION', 'False').lower() == 'true'
 )
+ENABLE_API_OUTLET_FILTERS = os.getenv('ENABLE_API_OUTLET_FILTERS', 'True').lower() == 'true'
 
 # When enabled, uses a hardcoded extension-to-MIME dictionary as a last-resort
 # fallback when both mimetypes.guess_type() and file.meta.content_type fail to
@@ -962,10 +1083,34 @@ SENTENCE_TRANSFORMERS_CROSS_ENCODER_SIGMOID_ACTIVATION_FUNCTION = (
     os.getenv('SENTENCE_TRANSFORMERS_CROSS_ENCODER_SIGMOID_ACTIVATION_FUNCTION', 'True').lower() == 'true'
 )
 
+####################################
+# KNOWLEDGE TOOLS
+####################################
+
+
+def _int_env(name: str, default: int) -> int:
+    try:
+        return max(int(os.getenv(name) or default), 1)
+    except (ValueError, TypeError):
+        return default
+
+
+# Total output of a single kb_exec call, whatever the command.
+KB_EXEC_MAX_OUTPUT_CHARS = _int_env('KB_EXEC_MAX_OUTPUT_CHARS', 30_000)
+# Files a single kb_exec grep may scan before it asks for a narrower scope.
+KB_EXEC_MAX_GREP_FILES = _int_env('KB_EXEC_MAX_GREP_FILES', 200)
+# Matching lines returned by kb_exec grep and grep_knowledge_files.
+KNOWLEDGE_GREP_MAX_MATCHES = _int_env('KNOWLEDGE_GREP_MAX_MATCHES', 50)
+# Characters returned by view_file / view_knowledge_file.
+VIEW_FILE_MAX_CHARS = _int_env('VIEW_FILE_MAX_CHARS', 100_000)
+VIEW_FILE_DEFAULT_MAX_CHARS = _int_env('VIEW_FILE_DEFAULT_MAX_CHARS', 10_000)
+
 ####################################
 # TOOLS/FUNCTIONS PIP OPTIONS
 ####################################
 
+ENABLE_PLUGINS = os.getenv('ENABLE_PLUGINS', 'True').lower() == 'true'
+
 ENABLE_PIP_INSTALL_FRONTMATTER_REQUIREMENTS = (
     os.getenv('ENABLE_PIP_INSTALL_FRONTMATTER_REQUIREMENTS', 'True').lower() == 'true'
 )
@@ -985,6 +1130,12 @@ if OFFLINE_MODE:
     os.environ['HF_HUB_OFFLINE'] = '1'
     ENABLE_VERSION_UPDATE_CHECK = False
 
+####################################
+# Pyodide file persistence
+####################################
+
+ENABLE_PYODIDE_FILE_PERSISTENCE = os.getenv('ENABLE_PYODIDE_FILE_PERSISTENCE', 'false').lower() == 'true'
+
 ####################################
 # Audit logging
 ####################################
@@ -1013,15 +1164,19 @@ except ValueError:
     MAX_BODY_LOG_SIZE = 2048
 
 # Comma separated list for urls to exclude from audit
-AUDIT_EXCLUDED_PATHS = os.getenv('AUDIT_EXCLUDED_PATHS', '/chats,/chat,/folders').split(',')
-AUDIT_EXCLUDED_PATHS = [path.strip() for path in AUDIT_EXCLUDED_PATHS]
-AUDIT_EXCLUDED_PATHS = [path.lstrip('/') for path in AUDIT_EXCLUDED_PATHS]
+AUDIT_EXCLUDED_PATHS = [
+    path
+    for path in (
+        path.strip().lstrip('/') for path in os.getenv('AUDIT_EXCLUDED_PATHS', '/chats,/chat,/folders').split(',')
+    )
+    if path
+]
 
 # Comma separated list of urls to include in audit (whitelist mode)
 # When set, only these paths are audited and AUDIT_EXCLUDED_PATHS is ignored
-AUDIT_INCLUDED_PATHS = os.getenv('AUDIT_INCLUDED_PATHS', '').split(',')
-AUDIT_INCLUDED_PATHS = [path.strip() for path in AUDIT_INCLUDED_PATHS]
-AUDIT_INCLUDED_PATHS = [path.lstrip('/') for path in AUDIT_INCLUDED_PATHS if path]
+AUDIT_INCLUDED_PATHS = [
+    path for path in (path.strip().lstrip('/') for path in os.getenv('AUDIT_INCLUDED_PATHS', '').split(',')) if path
+]
 
 # When enabled, GET requests are also audited (disabled by default to avoid log noise)
 ENABLE_AUDIT_GET_REQUESTS = os.getenv('ENABLE_AUDIT_GET_REQUESTS', 'False').lower() == 'true'

backend/open_webui/functions.py

@@ -20,7 +20,7 @@ from starlette.responses import Response, StreamingResponse
 
 from open_webui.config import BYPASS_ADMIN_ACCESS_CONTROL
 from open_webui.constants import ERROR_MESSAGES
-from open_webui.env import BYPASS_MODEL_ACCESS_CONTROL, GLOBAL_LOG_LEVEL
+from open_webui.env import BYPASS_MODEL_ACCESS_CONTROL, ENABLE_PLUGINS, GLOBAL_LOG_LEVEL
 from open_webui.models.functions import Functions
 from open_webui.models.models import Models
 from open_webui.models.users import UserModel
@@ -69,6 +69,9 @@ async def get_function_module_by_id(request: Request, pipe_id: str):
 
 
 async def get_function_models(request):
+    if not ENABLE_PLUGINS:
+        return []
+
     pipes = await Functions.get_functions_by_type('pipe', active_only=True)
     pipe_models = []
 
@@ -144,7 +147,10 @@ async def get_function_models(request):
     return pipe_models
 
 
-async def generate_function_chat_completion(request, form_data, user, models: dict = {}):
+async def generate_function_chat_completion(request, form_data, user, models: dict | None = None):
+    if models is None:
+        models = {}
+
     async def execute_pipe(pipe, params):
         if inspect.iscoroutinefunction(pipe):
             return await pipe(**params)
@@ -203,6 +209,10 @@ async def generate_function_chat_completion(request, form_data, user, models: di
 
         return params
 
+    # Copy so the base-model substitution below doesn't leak into the caller's
+    # payload, which the tool-call continuation re-submits. Mirrors the routers.
+    form_data = {**form_data}
+
     model_id = form_data.get('model')
     model_info = await Models.get_model_by_id(model_id)
 

backend/open_webui/internal/config.py

@@ -1,265 +0,0 @@
-"""Database-backed configuration with environment variable defaults."""
-
-from __future__ import annotations
-
-import asyncio
-import json
-import logging
-from datetime import datetime
-from functools import reduce
-from typing import Any, Optional, Union
-
-import redis
-from open_webui.internal.db import Base, get_async_db, get_db
-from open_webui.utils.redis import get_redis_connection
-from sqlalchemy import JSON, Column, DateTime, Integer, func, select
-
-log = logging.getLogger(__name__)
-
-
-# ── Model ────────────────────────────────────────────────────────────────────
-
-
-class ConfigTable(Base):
-    __tablename__ = 'config'
-
-    id = Column(Integer, primary_key=True)
-    data = Column(JSON, nullable=False)
-    version = Column(Integer, nullable=False, default=0)
-    created_at = Column(DateTime, nullable=False, server_default=func.now())
-    updated_at = Column(DateTime, nullable=True, onupdate=func.now())
-
-
-# ── Blob ─────────────────────────────────────────────────────────────────────
-
-
-class ConfigState:
-    """In-memory mirror of the single-row config JSON blob."""
-
-    __slots__ = ('_data',)
-
-    def __init__(self) -> None:
-        self._data: dict[str, Any] = {}
-
-    @property
-    def snapshot(self) -> dict:
-        return self._data
-
-    def read(self, path: str) -> Any:
-        return reduce(
-            lambda n, k: n.get(k) if isinstance(n, dict) else None,
-            path.split('.'),
-            self._data,
-        )
-
-    def write(self, path: str, value: Any) -> None:
-        keys = path.split('.')
-        reduce(lambda d, k: d.setdefault(k, {}), keys[:-1], self._data)[keys[-1]] = value
-
-    def replace(self, data: dict) -> None:
-        self._data = data
-
-    def load(self) -> dict:
-        with get_db() as db:
-            row = db.query(ConfigTable).order_by(ConfigTable.id.desc()).first()
-            self._data = row.data if row else {'version': 0, 'ui': {}}
-        return self._data
-
-    def persist(self, data: dict | None = None) -> None:
-        if data is not None:
-            self._data = data
-        with get_db() as db:
-            row = db.query(ConfigTable).first()
-            if row is None:
-                db.add(ConfigTable(data=self._data, version=0))
-            else:
-                row.data, row.updated_at = self._data, datetime.now()
-                db.add(row)
-            db.commit()
-
-    async def persist_async(self, data: dict | None = None) -> None:
-        if data is not None:
-            self._data = data
-        async with get_async_db() as db:
-            result = await db.execute(select(ConfigTable).limit(1))
-            row = result.scalars().first()
-            if row is None:
-                db.add(ConfigTable(data=self._data, version=0))
-            else:
-                row.data, row.updated_at = self._data, datetime.now()
-                db.add(row)
-            await db.commit()
-
-    def clear(self) -> None:
-        with get_db() as db:
-            db.query(ConfigTable).delete()
-            db.commit()
-
-    async def clear_async(self) -> None:
-        from sqlalchemy import delete as sa_delete
-
-        async with get_async_db() as db:
-            await db.execute(sa_delete(ConfigTable))
-            await db.commit()
-
-
-STATE = ConfigState()
-
-
-# ── ConfigVar ──────────────────────────────────────────────────────────────────
-
-
-_persist_enabled: bool = True
-_oauth_persist_enabled: bool = False
-_all_configs: list[ConfigVar] = []
-
-
-def initialize(*, enable_persistent: bool = True, enable_oauth_persistent: bool = False) -> dict:
-    global _persist_enabled, _oauth_persist_enabled
-    _persist_enabled = enable_persistent
-    _oauth_persist_enabled = enable_oauth_persistent
-    return STATE.load()
-
-
-class ConfigVar:
-    __slots__ = ('env_name', 'config_path', 'env_value', 'config_value', 'value')
-
-    def __init__(self, env_name: str, config_path: str, env_value: Any) -> None:
-        self.env_name = env_name
-        self.config_path = config_path
-        self.env_value = env_value
-        self.config_value = STATE.read(config_path)
-
-        if self.config_value is not None and _persist_enabled:
-            if config_path.startswith('oauth.') and not _oauth_persist_enabled:
-                log.info("Skipping DB value for '%s' (OAuth persistence disabled)", env_name)
-                self.value = env_value
-            else:
-                log.info("'%s' loaded from database", env_name)
-                self.value = self.config_value
-        else:
-            self.value = env_value
-
-        _all_configs.append(self)
-
-    def __str__(self) -> str:
-        return str(self.value)
-
-    def __repr__(self) -> str:
-        return f'<ConfigVar {self.env_name}={self.value!r}>'
-
-    @property
-    def __dict__(self):  # type: ignore[override]
-        raise TypeError(f"ConfigVar('{self.env_name}') cannot be cast to dict; use .value")
-
-    def __getattribute__(self, item: str):
-        if item == '__dict__':
-            raise TypeError('ConfigVar cannot be cast to dict; use .value')
-        return super().__getattribute__(item)
-
-    def refresh(self) -> None:
-        current = STATE.read(self.config_path)
-        if current is not None:
-            self.value = current
-            log.info('Refreshed %s → %s', self.env_name, self.value)
-
-    def commit(self) -> None:
-        log.info("Persisting '%s'", self.env_name)
-        STATE.write(self.config_path, self.value)
-        self.config_value = self.value
-        STATE.persist()
-
-    async def commit_async(self) -> None:
-        log.info("Persisting '%s'", self.env_name)
-        STATE.write(self.config_path, self.value)
-        self.config_value = self.value
-        await STATE.persist_async()
-
-
-# ── AppConfig ──────────────────────────────────────────────────────────
-
-
-class AppConfig:
-    """Attribute-style container for ConfigVars with optional Redis sync."""
-
-    def __init__(
-        self,
-        *,
-        redis_url: Optional[str] = None,
-        redis_sentinels: Optional[list] = None,
-        redis_cluster: bool = False,
-        redis_key_prefix: str = 'open-webui',
-    ) -> None:
-        super().__setattr__('_entries', {})
-        super().__setattr__('_key_prefix', redis_key_prefix)
-
-        # If sentinels weren't explicitly provided, read from env.
-        if redis_sentinels is None:
-            from open_webui.env import REDIS_SENTINEL_HOSTS, REDIS_SENTINEL_PORT
-            from open_webui.utils.redis import get_sentinels_from_env
-
-            redis_sentinels = get_sentinels_from_env(REDIS_SENTINEL_HOSTS, REDIS_SENTINEL_PORT)
-
-        rc: Union[redis.Redis, redis.cluster.RedisCluster, None] = None
-        if redis_url:
-            rc = get_redis_connection(redis_url, redis_sentinels or [], redis_cluster, decode_responses=True)
-        super().__setattr__('_rc', rc)
-
-    def __setattr__(self, name: str, value: Any) -> None:
-        entries: dict = super().__getattribute__('_entries')
-
-        if isinstance(value, ConfigVar):
-            entries[name] = value
-            return
-
-        entries[name].value = value
-
-        try:
-            asyncio.get_running_loop().create_task(self._write_async(name))
-        except RuntimeError:
-            entries[name].commit()
-
-        rc = super().__getattribute__('_rc')
-        if rc and _persist_enabled:
-            prefix = super().__getattribute__('_key_prefix')
-            try:
-                rc.set(f'{prefix}:config:{name}', json.dumps(entries[name].value))
-            except Exception as exc:
-                log.error("Redis write failed for '%s': %s", name, exc)
-
-    async def _write_async(self, name: str) -> None:
-        try:
-            await self._entries[name].commit_async()
-        except Exception as exc:
-            log.error("Async persist failed for '%s': %s", name, exc)
-
-    def __getattr__(self, name: str) -> Any:
-        entries = super().__getattribute__('_entries')
-        if name not in entries:
-            raise AttributeError(f"No config key '{name}'")
-
-        rc = super().__getattribute__('_rc')
-        if rc and _persist_enabled:
-            prefix = super().__getattribute__('_key_prefix')
-            try:
-                raw = rc.get(f'{prefix}:config:{name}')
-                if raw is not None:
-                    decoded = json.loads(raw)
-                    if entries[name].value != decoded:
-                        entries[name].value = decoded
-                        log.info("Updated '%s' from Redis", name)
-            except Exception as exc:
-                log.error("Redis read failed for '%s': %s", name, exc)
-
-        return entries[name].value
-
-    def _sync_to_redis(self) -> None:
-        rc = super().__getattribute__('_rc')
-        if not rc or not _persist_enabled:
-            return
-        prefix = super().__getattribute__('_key_prefix')
-        for name, s in super().__getattribute__('_entries').items():
-            try:
-                rc.set(f'{prefix}:config:{name}', json.dumps(s.value))
-            except Exception as exc:
-                log.error("Redis sync failed for '%s': %s", name, exc)

backend/open_webui/internal/db.py

@@ -5,10 +5,12 @@ import logging
 import os
 import sys
 from contextlib import asynccontextmanager, contextmanager
+from datetime import datetime, timedelta, timezone
 from typing import Any, Optional
 from urllib.parse import parse_qs, urlencode, urlparse, urlunparse
 
 from open_webui.env import (
+    DATABASE_ENABLE_IAM_TOKEN_AUTH,
     DATABASE_ENABLE_SESSION_SHARING,
     DATABASE_ENABLE_SQLITE_WAL,
     DATABASE_POOL_MAX_OVERFLOW,
@@ -27,6 +29,7 @@ from open_webui.env import (
     OPEN_WEBUI_DIR,
 )
 from sqlalchemy import Dialect, MetaData, create_engine, event, types
+from sqlalchemy.engine.url import make_url
 from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker, create_async_engine
 from sqlalchemy.ext.declarative import declarative_base
 from sqlalchemy.orm import Session, scoped_session, sessionmaker
@@ -146,6 +149,63 @@ _url_without_ssl, _ssl_dict = extract_ssl_params_from_url(DATABASE_URL)
 SQLALCHEMY_DATABASE_URL = reattach_ssl_params_to_url(_url_without_ssl, _ssl_dict) if _ssl_dict else DATABASE_URL
 
 
+class RDSIAMTokenAuth:
+    _refresh_after = timedelta(minutes=14)
+
+    def __init__(self, database_url: str) -> None:
+        url = make_url(database_url)
+        if not url.drivername.startswith(('postgresql', 'postgres')):
+            raise ValueError('DATABASE_ENABLE_IAM_TOKEN_AUTH is only supported for PostgreSQL databases')
+        if not url.host or not url.username:
+            raise ValueError('DATABASE_ENABLE_IAM_TOKEN_AUTH requires a database host and user')
+
+        self.host = url.host
+        self.port = url.port or 5432
+        self.username = url.username
+        self._client = None
+        self._token: str | None = None
+        self._expires_at = datetime.min.replace(tzinfo=timezone.utc)
+
+    @property
+    def client(self):
+        if self._client is None:
+            import boto3
+
+            self._client = boto3.client('rds')
+        return self._client
+
+    def get_password(self) -> str:
+        now = datetime.now(timezone.utc)
+        if self._token and now < self._expires_at:
+            return self._token
+
+        self._token = self.client.generate_db_auth_token(
+            DBHostname=self.host,
+            Port=self.port,
+            DBUsername=self.username,
+        )
+        self._expires_at = now + self._refresh_after
+        log.info('AWS RDS IAM database token refreshed; next refresh after %s', self._expires_at.isoformat())
+        return self._token
+
+
+_rds_iam_token_auth = RDSIAMTokenAuth(SQLALCHEMY_DATABASE_URL) if DATABASE_ENABLE_IAM_TOKEN_AUTH else None
+
+
+def _set_iam_token_password(dialect, conn_rec, cargs, cparams):
+    if _rds_iam_token_auth is not None:
+        cparams['password'] = _rds_iam_token_auth.get_password()
+
+
+def enable_iam_token_auth(connectable) -> None:
+    if _rds_iam_token_auth is None:
+        return
+
+    engine = getattr(connectable, 'sync_engine', connectable)
+    if not event.contains(engine, 'do_connect', _set_iam_token_password):
+        event.listen(engine, 'do_connect', _set_iam_token_password)
+
+
 def _make_async_url(url: str) -> str:
     """Convert a sync database URL to its async driver equivalent.
 
@@ -268,6 +328,8 @@ else:
     else:
         engine = create_engine(SQLALCHEMY_DATABASE_URL, pool_pre_ping=True)
 
+enable_iam_token_auth(engine)
+
 
 # Sync session — used ONLY for startup config loading (config.py runs at import time)
 SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine, expire_on_commit=False)
@@ -308,6 +370,8 @@ if sys.platform == 'win32' and _is_postgres_url(DATABASE_URL):
 
 if 'sqlite' in ASYNC_SQLALCHEMY_DATABASE_URL:
     # Generous default — async coroutines + no session sharing = high connection demand.
+    # No pool_pre_ping: a local SQLite file cannot drop connections, and the
+    # ping costs a worker-thread hop plus a SELECT 1 on every checkout.
     _sqlite_pool_size = DATABASE_POOL_SIZE if isinstance(DATABASE_POOL_SIZE, int) and DATABASE_POOL_SIZE > 0 else 512
     async_engine = create_async_engine(
         ASYNC_SQLALCHEMY_DATABASE_URL,
@@ -315,7 +379,6 @@ if 'sqlite' in ASYNC_SQLALCHEMY_DATABASE_URL:
         pool_size=_sqlite_pool_size,
         pool_timeout=DATABASE_POOL_TIMEOUT,
         pool_recycle=DATABASE_POOL_RECYCLE,
-        pool_pre_ping=True,
     )
 
     @event.listens_for(async_engine.sync_engine, 'connect')
@@ -344,6 +407,8 @@ else:
             pool_pre_ping=True,
         )
 
+enable_iam_token_auth(async_engine)
+
 
 AsyncSessionLocal = async_sessionmaker(
     bind=async_engine,

backend/open_webui/migrations/env.py

@@ -6,9 +6,11 @@ import logging.config
 import logging
 import alembic.context
 from open_webui.env import DATABASE_PASSWORD, DATABASE_URL, LOG_FORMAT
-from open_webui.internal.db import extract_ssl_params_from_url, reattach_ssl_params_to_url
+from open_webui.internal.db import enable_iam_token_auth, extract_ssl_params_from_url, reattach_ssl_params_to_url
 from open_webui.models.auths import Auth
 from open_webui.models.calendar import Calendar, CalendarEvent, CalendarEventAttendee  # noqa: F401
+from open_webui.models.chat_messages import ChatMessage  # noqa: F401
+from open_webui.models.chats import Chat  # noqa: F401
 from sqlalchemy import create_engine, engine_from_config, pool
 
 alembic_config = alembic.context.config
@@ -68,6 +70,7 @@ def _get_engine_connectable():
 def run_migrations_online() -> None:
     """Execute migrations against a live database connection."""
     live_connectable = _get_engine_connectable()
+    enable_iam_token_auth(live_connectable)
     with live_connectable.connect() as live_connection:
         alembic.context.configure(
             connection=live_connection,

backend/open_webui/migrations/versions/242a2047eae0_update_chat_table.py

@@ -49,7 +49,7 @@ def upgrade():
 
     # Step 3: Migrate data from 'old_chat' to 'chat' (only if old_chat exists)
     # Re-check columns after potential rename above
-    current_cols = {c['name'] for c in inspector.get_columns('chat')}
+    current_cols = {c['name'] for c in sa.inspect(conn).get_columns('chat')}
     if 'old_chat' in current_cols:
         chat_table = table(
             'chat',
@@ -76,8 +76,12 @@ def upgrade():
 
 
 def downgrade():
+    conn = op.get_bind()
+    columns = {col['name'] for col in sa.inspect(conn).get_columns('chat')}
+
     # Step 1: Add 'old_chat' column back as Text
-    op.add_column('chat', sa.Column('old_chat', sa.Text(), nullable=True))
+    if 'old_chat' not in columns:
+        op.add_column('chat', sa.Column('old_chat', sa.Text(), nullable=True))
 
     # Step 2: Convert 'chat' JSON data back to text and store in 'old_chat'
     chat_table = table(
@@ -87,14 +91,14 @@ def downgrade():
         sa.Column('old_chat', sa.Text()),
     )
 
-    connection = op.get_bind()
-    results = connection.execute(select(chat_table.c.id, chat_table.c.chat))
-    for row in results:
-        text_data = json.dumps(row.chat) if row.chat is not None else None
-        connection.execute(sa.update(chat_table).where(chat_table.c.id == row.id).values(old_chat=text_data))
+    if 'chat' in columns:
+        results = conn.execute(select(chat_table.c.id, chat_table.c.chat))
+        for row in results:
+            text_data = json.dumps(row.chat) if row.chat is not None else None
+            conn.execute(sa.update(chat_table).where(chat_table.c.id == row.id).values(old_chat=text_data))
 
-    # Step 3: Remove the new 'chat' JSON column
-    op.drop_column('chat', 'chat')
+        # Step 3: Remove the new 'chat' JSON column
+        op.drop_column('chat', 'chat')
 
     # Step 4: Rename 'old_chat' back to 'chat'
     op.alter_column('chat', 'old_chat', new_column_name='chat', existing_type=sa.Text())

backend/open_webui/migrations/versions/3ff2c63645b8_reshape_config_to_per_key_rows.py

@@ -0,0 +1,584 @@
+"""reshape config to per key rows
+
+Revision ID: 3ff2c63645b8
+Revises: 461111b60977
+Create Date: 2026-06-17 00:50:51.477073
+
+"""
+
+import json
+import time
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision: str = '3ff2c63645b8'
+down_revision: Union[str, None] = '461111b60977'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+# Maps every dot-notation blob path to its legacy env/config key name.
+# Built from the legacy persistent config declarations in config.py.
+BLOB_PATH_TO_KEY = {
+    'audio.stt.allowed_extensions': 'AUDIO_STT_ALLOWED_EXTENSIONS',
+    'audio.stt.azure.api_key': 'AUDIO_STT_AZURE_API_KEY',
+    'audio.stt.azure.base_url': 'AUDIO_STT_AZURE_BASE_URL',
+    'audio.stt.azure.locales': 'AUDIO_STT_AZURE_LOCALES',
+    'audio.stt.azure.max_speakers': 'AUDIO_STT_AZURE_MAX_SPEAKERS',
+    'audio.stt.azure.region': 'AUDIO_STT_AZURE_REGION',
+    'audio.stt.deepgram.api_key': 'DEEPGRAM_API_KEY',
+    'audio.stt.engine': 'AUDIO_STT_ENGINE',
+    'audio.stt.mistral.api_base_url': 'AUDIO_STT_MISTRAL_API_BASE_URL',
+    'audio.stt.mistral.api_key': 'AUDIO_STT_MISTRAL_API_KEY',
+    'audio.stt.mistral.use_chat_completions': 'AUDIO_STT_MISTRAL_USE_CHAT_COMPLETIONS',
+    'audio.stt.model': 'AUDIO_STT_MODEL',
+    'audio.stt.openai.api_base_url': 'AUDIO_STT_OPENAI_API_BASE_URL',
+    'audio.stt.openai.api_key': 'AUDIO_STT_OPENAI_API_KEY',
+    'audio.stt.supported_content_types': 'AUDIO_STT_SUPPORTED_CONTENT_TYPES',
+    'audio.stt.whisper_model': 'WHISPER_MODEL',
+    'audio.tts.api_key': 'AUDIO_TTS_API_KEY',
+    'audio.tts.azure.speech_base_url': 'AUDIO_TTS_AZURE_SPEECH_BASE_URL',
+    'audio.tts.azure.speech_output_format': 'AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT',
+    'audio.tts.azure.speech_region': 'AUDIO_TTS_AZURE_SPEECH_REGION',
+    'audio.tts.engine': 'AUDIO_TTS_ENGINE',
+    'audio.tts.mistral.api_base_url': 'AUDIO_TTS_MISTRAL_API_BASE_URL',
+    'audio.tts.mistral.api_key': 'AUDIO_TTS_MISTRAL_API_KEY',
+    'audio.tts.model': 'AUDIO_TTS_MODEL',
+    'audio.tts.openai.api_base_url': 'AUDIO_TTS_OPENAI_API_BASE_URL',
+    'audio.tts.openai.api_key': 'AUDIO_TTS_OPENAI_API_KEY',
+    'audio.tts.openai.params': 'AUDIO_TTS_OPENAI_PARAMS',
+    'audio.tts.split_on': 'AUDIO_TTS_SPLIT_ON',
+    'audio.tts.voice': 'AUDIO_TTS_VOICE',
+    'auth.admin.email': 'ADMIN_EMAIL',
+    'auth.admin.show': 'SHOW_ADMIN_DETAILS',
+    'auth.api_key.allowed_endpoints': 'API_KEYS_ALLOWED_ENDPOINTS',
+    'auth.api_key.endpoint_restrictions': 'ENABLE_API_KEYS_ENDPOINT_RESTRICTIONS',
+    'auth.enable_api_keys': 'ENABLE_API_KEYS',
+    'auth.jwt_expiry': 'JWT_EXPIRES_IN',
+    'automations.enable': 'ENABLE_AUTOMATIONS',
+    'automations.max_count': 'AUTOMATION_MAX_COUNT',
+    'automations.min_interval': 'AUTOMATION_MIN_INTERVAL',
+    'calendar.enable': 'ENABLE_CALENDAR',
+    'channels.enable': 'ENABLE_CHANNELS',
+    'code_execution.enable': 'ENABLE_CODE_EXECUTION',
+    'code_execution.engine': 'CODE_EXECUTION_ENGINE',
+    'code_execution.jupyter.auth': 'CODE_EXECUTION_JUPYTER_AUTH',
+    'code_execution.jupyter.auth_password': 'CODE_EXECUTION_JUPYTER_AUTH_PASSWORD',
+    'code_execution.jupyter.auth_token': 'CODE_EXECUTION_JUPYTER_AUTH_TOKEN',
+    'code_execution.jupyter.timeout': 'CODE_EXECUTION_JUPYTER_TIMEOUT',
+    'code_execution.jupyter.url': 'CODE_EXECUTION_JUPYTER_URL',
+    'code_interpreter.enable': 'ENABLE_CODE_INTERPRETER',
+    'code_interpreter.engine': 'CODE_INTERPRETER_ENGINE',
+    'code_interpreter.jupyter.auth': 'CODE_INTERPRETER_JUPYTER_AUTH',
+    'code_interpreter.jupyter.auth_password': 'CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD',
+    'code_interpreter.jupyter.auth_token': 'CODE_INTERPRETER_JUPYTER_AUTH_TOKEN',
+    'code_interpreter.jupyter.timeout': 'CODE_INTERPRETER_JUPYTER_TIMEOUT',
+    'code_interpreter.jupyter.url': 'CODE_INTERPRETER_JUPYTER_URL',
+    'code_interpreter.prompt_template': 'CODE_INTERPRETER_PROMPT_TEMPLATE',
+    'direct.enable': 'ENABLE_DIRECT_CONNECTIONS',
+    'evaluation.arena.enable': 'ENABLE_EVALUATION_ARENA_MODELS',
+    'evaluation.arena.models': 'EVALUATION_ARENA_MODELS',
+    'file.image_compression_height': 'FILE_IMAGE_COMPRESSION_HEIGHT',
+    'file.image_compression_width': 'FILE_IMAGE_COMPRESSION_WIDTH',
+    'folders.enable': 'ENABLE_FOLDERS',
+    'folders.max_file_count': 'FOLDER_MAX_FILE_COUNT',
+    'google_drive.api_key': 'GOOGLE_DRIVE_API_KEY',
+    'google_drive.client_id': 'GOOGLE_DRIVE_CLIENT_ID',
+    'google_drive.enable': 'ENABLE_GOOGLE_DRIVE_INTEGRATION',
+    'image_generation.automatic1111.api_auth': 'AUTOMATIC1111_API_AUTH',
+    'image_generation.automatic1111.api_params': 'AUTOMATIC1111_PARAMS',
+    'image_generation.automatic1111.base_url': 'AUTOMATIC1111_BASE_URL',
+    'image_generation.comfyui.api_key': 'COMFYUI_API_KEY',
+    'image_generation.comfyui.base_url': 'COMFYUI_BASE_URL',
+    'image_generation.comfyui.nodes': 'COMFYUI_WORKFLOW_NODES',
+    'image_generation.comfyui.workflow': 'COMFYUI_WORKFLOW',
+    'image_generation.enable': 'ENABLE_IMAGE_GENERATION',
+    'image_generation.engine': 'IMAGE_GENERATION_ENGINE',
+    'image_generation.gemini.api_base_url': 'IMAGES_GEMINI_API_BASE_URL',
+    'image_generation.gemini.api_key': 'IMAGES_GEMINI_API_KEY',
+    'image_generation.gemini.endpoint_method': 'IMAGES_GEMINI_ENDPOINT_METHOD',
+    'image_generation.model': 'IMAGE_GENERATION_MODEL',
+    'image_generation.openai.api_base_url': 'IMAGES_OPENAI_API_BASE_URL',
+    'image_generation.openai.api_key': 'IMAGES_OPENAI_API_KEY',
+    'image_generation.openai.api_version': 'IMAGES_OPENAI_API_VERSION',
+    'image_generation.openai.params': 'IMAGES_OPENAI_API_PARAMS',
+    'image_generation.prompt.enable': 'ENABLE_IMAGE_PROMPT_GENERATION',
+    'image_generation.size': 'IMAGE_SIZE',
+    'image_generation.steps': 'IMAGE_STEPS',
+    'images.edit.comfyui.api_key': 'IMAGES_EDIT_COMFYUI_API_KEY',
+    'images.edit.comfyui.base_url': 'IMAGES_EDIT_COMFYUI_BASE_URL',
+    'images.edit.comfyui.nodes': 'IMAGES_EDIT_COMFYUI_WORKFLOW_NODES',
+    'images.edit.comfyui.workflow': 'IMAGES_EDIT_COMFYUI_WORKFLOW',
+    'images.edit.enable': 'ENABLE_IMAGE_EDIT',
+    'images.edit.engine': 'IMAGE_EDIT_ENGINE',
+    'images.edit.gemini.api_base_url': 'IMAGES_EDIT_GEMINI_API_BASE_URL',
+    'images.edit.gemini.api_key': 'IMAGES_EDIT_GEMINI_API_KEY',
+    'images.edit.model': 'IMAGE_EDIT_MODEL',
+    'images.edit.openai.api_base_url': 'IMAGES_EDIT_OPENAI_API_BASE_URL',
+    'images.edit.openai.api_key': 'IMAGES_EDIT_OPENAI_API_KEY',
+    'images.edit.openai.api_version': 'IMAGES_EDIT_OPENAI_API_VERSION',
+    'images.edit.size': 'IMAGE_EDIT_SIZE',
+    'ldap.enable': 'ENABLE_LDAP',
+    'ldap.group.enable_creation': 'ENABLE_LDAP_GROUP_CREATION',
+    'ldap.group.enable_management': 'ENABLE_LDAP_GROUP_MANAGEMENT',
+    'ldap.server.app_dn': 'LDAP_APP_DN',
+    'ldap.server.app_password': 'LDAP_APP_PASSWORD',
+    'ldap.server.attribute_for_groups': 'LDAP_ATTRIBUTE_FOR_GROUPS',
+    'ldap.server.attribute_for_mail': 'LDAP_ATTRIBUTE_FOR_MAIL',
+    'ldap.server.attribute_for_username': 'LDAP_ATTRIBUTE_FOR_USERNAME',
+    'ldap.server.ca_cert_file': 'LDAP_CA_CERT_FILE',
+    'ldap.server.ciphers': 'LDAP_CIPHERS',
+    'ldap.server.host': 'LDAP_SERVER_HOST',
+    'ldap.server.label': 'LDAP_SERVER_LABEL',
+    'ldap.server.port': 'LDAP_SERVER_PORT',
+    'ldap.server.search_filter': 'LDAP_SEARCH_FILTER',
+    'ldap.server.use_tls': 'LDAP_USE_TLS',
+    'ldap.server.users_dn': 'LDAP_SEARCH_BASE',
+    'ldap.server.validate_cert': 'LDAP_VALIDATE_CERT',
+    'memories.enable': 'ENABLE_MEMORIES',
+    'models.base_models_cache': 'ENABLE_BASE_MODELS_CACHE',
+    'models.default_metadata': 'DEFAULT_MODEL_METADATA',
+    'models.default_params': 'DEFAULT_MODEL_PARAMS',
+    'notes.enable': 'ENABLE_NOTES',
+    # OAuth — direct paths
+    'oauth.admin_roles': 'OAUTH_ADMIN_ROLES',
+    'oauth.allowed_domains': 'OAUTH_ALLOWED_DOMAINS',
+    'oauth.allowed_roles': 'OAUTH_ALLOWED_ROLES',
+    'oauth.audience': 'OAUTH_AUDIENCE',
+    'oauth.auto_redirect': 'OAUTH_AUTO_REDIRECT',
+    'oauth.blocked_groups': 'OAUTH_BLOCKED_GROUPS',
+    'oauth.client.timeout': 'OAUTH_CLIENT_TIMEOUT',
+    'oauth.enable_group_creation': 'ENABLE_OAUTH_GROUP_CREATION',
+    'oauth.enable_group_mapping': 'ENABLE_OAUTH_GROUP_MANAGEMENT',
+    'oauth.enable_role_mapping': 'ENABLE_OAUTH_ROLE_MANAGEMENT',
+    'oauth.enable_signup': 'ENABLE_OAUTH_SIGNUP',
+    'oauth.group_default_share': 'OAUTH_GROUP_DEFAULT_SHARE',
+    'oauth.merge_accounts_by_email': 'OAUTH_MERGE_ACCOUNTS_BY_EMAIL',
+    'oauth.refresh_token_include_scope': 'OAUTH_REFRESH_TOKEN_INCLUDE_SCOPE',
+    'oauth.roles_claim': 'OAUTH_ROLES_CLAIM',
+    'oauth.update_email_on_login': 'OAUTH_UPDATE_EMAIL_ON_LOGIN',
+    'oauth.update_name_on_login': 'OAUTH_UPDATE_NAME_ON_LOGIN',
+    'oauth.update_picture_on_login': 'OAUTH_UPDATE_PICTURE_ON_LOGIN',
+    # OAuth — generic provider paths
+    'oauth.client_id': 'OAUTH_CLIENT_ID',
+    'oauth.client_secret': 'OAUTH_CLIENT_SECRET',
+    'oauth.code_challenge_method': 'OAUTH_CODE_CHALLENGE_METHOD',
+    'oauth.email_claim': 'OAUTH_EMAIL_CLAIM',
+    'oauth.end_session_endpoint': 'OPENID_END_SESSION_ENDPOINT',
+    'oauth.group_claim': 'OAUTH_GROUP_CLAIM',
+    'oauth.picture_claim': 'OAUTH_PICTURE_CLAIM',
+    'oauth.provider_name': 'OAUTH_PROVIDER_NAME',
+    'oauth.provider_url': 'OPENID_PROVIDER_URL',
+    'oauth.redirect_uri': 'OPENID_REDIRECT_URI',
+    'oauth.scopes': 'OAUTH_SCOPES',
+    'oauth.sub_claim': 'OAUTH_SUB_CLAIM',
+    'oauth.timeout': 'OAUTH_TIMEOUT',
+    'oauth.token_endpoint_auth_method': 'OAUTH_TOKEN_ENDPOINT_AUTH_METHOD',
+    'oauth.username_claim': 'OAUTH_USERNAME_CLAIM',
+    # OAuth — OIDC nested paths (flattened)
+    'oauth.oidc.avatar_claim': 'OAUTH_PICTURE_CLAIM',
+    'oauth.oidc.client_id': 'OAUTH_CLIENT_ID',
+    'oauth.oidc.client_secret': 'OAUTH_CLIENT_SECRET',
+    'oauth.oidc.code_challenge_method': 'OAUTH_CODE_CHALLENGE_METHOD',
+    'oauth.oidc.email_claim': 'OAUTH_EMAIL_CLAIM',
+    'oauth.oidc.end_session_endpoint': 'OPENID_END_SESSION_ENDPOINT',
+    'oauth.oidc.group_claim': 'OAUTH_GROUP_CLAIM',  # renamed from OAUTH_GROUPS_CLAIM
+    'oauth.oidc.oauth_timeout': 'OAUTH_TIMEOUT',
+    'oauth.oidc.provider_name': 'OAUTH_PROVIDER_NAME',
+    'oauth.oidc.provider_url': 'OPENID_PROVIDER_URL',
+    'oauth.oidc.redirect_uri': 'OPENID_REDIRECT_URI',
+    'oauth.oidc.scopes': 'OAUTH_SCOPES',
+    'oauth.oidc.sub_claim': 'OAUTH_SUB_CLAIM',
+    'oauth.oidc.token_endpoint_auth_method': 'OAUTH_TOKEN_ENDPOINT_AUTH_METHOD',
+    'oauth.oidc.username_claim': 'OAUTH_USERNAME_CLAIM',
+    # OAuth — provider-specific
+    'oauth.feishu.client_id': 'FEISHU_CLIENT_ID',
+    'oauth.feishu.client_secret': 'FEISHU_CLIENT_SECRET',
+    'oauth.feishu.redirect_uri': 'FEISHU_REDIRECT_URI',
+    'oauth.feishu.scope': 'FEISHU_OAUTH_SCOPE',
+    'oauth.github.client_id': 'GITHUB_CLIENT_ID',
+    'oauth.github.client_secret': 'GITHUB_CLIENT_SECRET',
+    'oauth.github.redirect_uri': 'GITHUB_CLIENT_REDIRECT_URI',
+    'oauth.github.scope': 'GITHUB_CLIENT_SCOPE',
+    'oauth.google.client_id': 'GOOGLE_CLIENT_ID',
+    'oauth.google.client_secret': 'GOOGLE_CLIENT_SECRET',
+    'oauth.google.redirect_uri': 'GOOGLE_REDIRECT_URI',
+    'oauth.google.scope': 'GOOGLE_OAUTH_SCOPE',
+    'oauth.microsoft.client_id': 'MICROSOFT_CLIENT_ID',
+    'oauth.microsoft.client_secret': 'MICROSOFT_CLIENT_SECRET',
+    'oauth.microsoft.login_base_url': 'MICROSOFT_CLIENT_LOGIN_BASE_URL',
+    'oauth.microsoft.picture_url': 'MICROSOFT_CLIENT_PICTURE_URL',
+    'oauth.microsoft.redirect_uri': 'MICROSOFT_REDIRECT_URI',
+    'oauth.microsoft.scope': 'MICROSOFT_OAUTH_SCOPE',
+    'oauth.microsoft.tenant_id': 'MICROSOFT_CLIENT_TENANT_ID',
+    # Ollama / OpenAI
+    'ollama.api_configs': 'OLLAMA_API_CONFIGS',
+    'ollama.base_urls': 'OLLAMA_BASE_URLS',
+    'ollama.enable': 'ENABLE_OLLAMA_API',
+    'onedrive.enable': 'ENABLE_ONEDRIVE_INTEGRATION',
+    'onedrive.sharepoint_tenant_id': 'ONEDRIVE_SHAREPOINT_TENANT_ID',
+    'onedrive.sharepoint_url': 'ONEDRIVE_SHAREPOINT_URL',
+    'openai.api_base_urls': 'OPENAI_API_BASE_URLS',
+    'openai.api_configs': 'OPENAI_API_CONFIGS',
+    'openai.api_keys': 'OPENAI_API_KEYS',
+    'openai.enable': 'ENABLE_OPENAI_API',
+    # RAG
+    'rag.content_extraction_engine': 'CONTENT_EXTRACTION_ENGINE',
+    'rag.datalab_marker_use_llm': 'DATALAB_MARKER_USE_LLM',
+    'rag.mistral_ocr_api_base_url': 'MISTRAL_OCR_API_BASE_URL',
+    'rag.azure_openai.api_key': 'RAG_AZURE_OPENAI_API_KEY',
+    'rag.azure_openai.api_version': 'RAG_AZURE_OPENAI_API_VERSION',
+    'rag.azure_openai.base_url': 'RAG_AZURE_OPENAI_BASE_URL',
+    'rag.bypass_embedding_and_retrieval': 'BYPASS_EMBEDDING_AND_RETRIEVAL',
+    'rag.chunk_min_size_target': 'CHUNK_MIN_SIZE_TARGET',
+    'rag.chunk_overlap': 'CHUNK_OVERLAP',
+    'rag.chunk_size': 'CHUNK_SIZE',
+    'rag.datalab_marker_additional_config': 'DATALAB_MARKER_ADDITIONAL_CONFIG',
+    'rag.datalab_marker_api_base_url': 'DATALAB_MARKER_API_BASE_URL',
+    'rag.datalab_marker_api_key': 'DATALAB_MARKER_API_KEY',
+    'rag.datalab_marker_disable_image_extraction': 'DATALAB_MARKER_DISABLE_IMAGE_EXTRACTION',
+    'rag.datalab_marker_force_ocr': 'DATALAB_MARKER_FORCE_OCR',
+    'rag.datalab_marker_format_lines': 'DATALAB_MARKER_FORMAT_LINES',
+    'rag.datalab_marker_output_format': 'DATALAB_MARKER_OUTPUT_FORMAT',
+    'rag.datalab_marker_paginate': 'DATALAB_MARKER_PAGINATE',
+    'rag.datalab_marker_skip_cache': 'DATALAB_MARKER_SKIP_CACHE',
+    'rag.datalab_marker_strip_existing_ocr': 'DATALAB_MARKER_STRIP_EXISTING_OCR',
+    'rag.docling_api_key': 'DOCLING_API_KEY',
+    'rag.docling_params': 'DOCLING_PARAMS',
+    'rag.docling_server_url': 'DOCLING_SERVER_URL',
+    'rag.document_intelligence_endpoint': 'DOCUMENT_INTELLIGENCE_ENDPOINT',
+    'rag.document_intelligence_key': 'DOCUMENT_INTELLIGENCE_KEY',
+    'rag.document_intelligence_model': 'DOCUMENT_INTELLIGENCE_MODEL',
+    'rag.embedding_batch_size': 'RAG_EMBEDDING_BATCH_SIZE',
+    'rag.embedding_concurrent_requests': 'RAG_EMBEDDING_CONCURRENT_REQUESTS',
+    'rag.embedding_engine': 'RAG_EMBEDDING_ENGINE',
+    'rag.embedding_model': 'RAG_EMBEDDING_MODEL',
+    'rag.enable_async_embedding': 'ENABLE_ASYNC_EMBEDDING',
+    'rag.enable_hybrid_search': 'ENABLE_RAG_HYBRID_SEARCH',
+    'rag.enable_hybrid_search_enriched_texts': 'ENABLE_RAG_HYBRID_SEARCH_ENRICHED_TEXTS',
+    'rag.enable_markdown_header_text_splitter': 'ENABLE_MARKDOWN_HEADER_TEXT_SPLITTER',
+    'rag.external_document_loader_api_key': 'EXTERNAL_DOCUMENT_LOADER_API_KEY',
+    'rag.external_document_loader_url': 'EXTERNAL_DOCUMENT_LOADER_URL',
+    'rag.external_reranker_api_key': 'RAG_EXTERNAL_RERANKER_API_KEY',
+    'rag.external_reranker_timeout': 'RAG_EXTERNAL_RERANKER_TIMEOUT',
+    'rag.external_reranker_url': 'RAG_EXTERNAL_RERANKER_URL',
+    'rag.file.allowed_extensions': 'RAG_ALLOWED_FILE_EXTENSIONS',
+    'rag.file.max_count': 'RAG_FILE_MAX_COUNT',
+    'rag.file.max_size': 'RAG_FILE_MAX_SIZE',
+    'rag.full_context': 'RAG_FULL_CONTEXT',
+    'rag.hybrid_bm25_weight': 'RAG_HYBRID_BM25_WEIGHT',
+    'rag.mineru_api_key': 'MINERU_API_KEY',
+    'rag.mineru_api_mode': 'MINERU_API_MODE',
+    'rag.mineru_api_timeout': 'MINERU_API_TIMEOUT',
+    'rag.mineru_api_url': 'MINERU_API_URL',
+    'rag.mineru_file_extensions': 'MINERU_FILE_EXTENSIONS',
+    'rag.mineru_params': 'MINERU_PARAMS',
+    'rag.mistral_ocr_api_key': 'MISTRAL_OCR_API_KEY',
+    'rag.ollama.key': 'RAG_OLLAMA_API_KEY',
+    'rag.ollama.url': 'RAG_OLLAMA_BASE_URL',
+    'rag.openai_api_base_url': 'RAG_OPENAI_API_BASE_URL',
+    'rag.openai_api_key': 'RAG_OPENAI_API_KEY',
+    'rag.paddleocr_vl_base_url': 'PADDLEOCR_VL_BASE_URL',
+    'rag.paddleocr_vl_token': 'PADDLEOCR_VL_TOKEN',
+    'rag.pdf_extract_images': 'PDF_EXTRACT_IMAGES',
+    'rag.pdf_loader_mode': 'PDF_LOADER_MODE',
+    'rag.relevance_threshold': 'RAG_RELEVANCE_THRESHOLD',
+    'rag.reranking_batch_size': 'RAG_RERANKING_BATCH_SIZE',
+    'rag.reranking_engine': 'RAG_RERANKING_ENGINE',
+    'rag.reranking_model': 'RAG_RERANKING_MODEL',
+    'rag.template': 'RAG_TEMPLATE',
+    'rag.text_splitter': 'RAG_TEXT_SPLITTER',
+    'rag.tika_server_url': 'TIKA_SERVER_URL',
+    'rag.tiktoken_encoding_name': 'TIKTOKEN_ENCODING_NAME',
+    'rag.top_k': 'RAG_TOP_K',
+    'rag.top_k_reranker': 'RAG_TOP_K_RERANKER',
+    # RAG — Web
+    'rag.web.fetch.max_content_length': 'WEB_FETCH_MAX_CONTENT_LENGTH',
+    'rag.web.loader.concurrent_requests': 'WEB_LOADER_CONCURRENT_REQUESTS',
+    'rag.web.loader.engine': 'WEB_LOADER_ENGINE',
+    'rag.web.loader.external_web_loader_api_key': 'EXTERNAL_WEB_LOADER_API_KEY',
+    'rag.web.loader.external_web_loader_url': 'EXTERNAL_WEB_LOADER_URL',
+    'rag.web.loader.firecrawl_api_key': 'FIRECRAWL_API_KEY',
+    'rag.web.loader.firecrawl_api_url': 'FIRECRAWL_API_BASE_URL',
+    'rag.web.loader.firecrawl_timeout': 'FIRECRAWL_TIMEOUT',
+    'rag.web.loader.playwright_timeout': 'PLAYWRIGHT_TIMEOUT',
+    'rag.web.loader.playwright_ws_url': 'PLAYWRIGHT_WS_URL',
+    'rag.web.loader.ssl_verification': 'ENABLE_WEB_LOADER_SSL_VERIFICATION',
+    'rag.web.loader.timeout': 'WEB_LOADER_TIMEOUT',
+    'rag.web.search.azure_ai_search_api_key': 'AZURE_AI_SEARCH_API_KEY',
+    'rag.web.search.azure_ai_search_endpoint': 'AZURE_AI_SEARCH_ENDPOINT',
+    'rag.web.search.azure_ai_search_index_name': 'AZURE_AI_SEARCH_INDEX_NAME',
+    'rag.web.search.bing_search_v7_endpoint': 'BING_SEARCH_V7_ENDPOINT',
+    'rag.web.search.bing_search_v7_subscription_key': 'BING_SEARCH_V7_SUBSCRIPTION_KEY',
+    'rag.web.search.bocha_search_api_key': 'BOCHA_SEARCH_API_KEY',
+    'rag.web.search.brave_search_api_key': 'BRAVE_SEARCH_API_KEY',
+    'rag.web.search.brave_search_context_tokens': 'BRAVE_SEARCH_CONTEXT_TOKENS',
+    'rag.web.search.bypass_embedding_and_retrieval': 'BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL',
+    'rag.web.search.bypass_web_loader': 'BYPASS_WEB_SEARCH_WEB_LOADER',
+    'rag.web.search.concurrent_requests': 'WEB_SEARCH_CONCURRENT_REQUESTS',
+    'rag.web.search.ddgs_backend': 'DDGS_BACKEND',
+    'rag.web.search.domain.filter_list': 'WEB_SEARCH_DOMAIN_FILTER_LIST',
+    'rag.web.search.enable': 'ENABLE_WEB_SEARCH',
+    'rag.web.search.engine': 'WEB_SEARCH_ENGINE',
+    'rag.web.search.exa_api_key': 'EXA_API_KEY',
+    'rag.web.search.external_web_search_api_key': 'EXTERNAL_WEB_SEARCH_API_KEY',
+    'rag.web.search.external_web_search_url': 'EXTERNAL_WEB_SEARCH_URL',
+    'rag.web.search.google_pse_api_key': 'GOOGLE_PSE_API_KEY',
+    'rag.web.search.google_pse_engine_id': 'GOOGLE_PSE_ENGINE_ID',
+    'rag.web.search.jina_api_base_url': 'JINA_API_BASE_URL',
+    'rag.web.search.jina_api_key': 'JINA_API_KEY',
+    'rag.web.search.kagi_search_api_key': 'KAGI_SEARCH_API_KEY',
+    'rag.web.search.linkup_api_key': 'LINKUP_API_KEY',
+    'rag.web.search.linkup_search_params': 'LINKUP_SEARCH_PARAMS',
+    'rag.web.search.mojeek_search_api_key': 'MOJEEK_SEARCH_API_KEY',
+    'rag.web.search.ollama_cloud_api_key': 'OLLAMA_CLOUD_WEB_SEARCH_API_KEY',
+    'rag.web.search.perplexity_api_key': 'PERPLEXITY_API_KEY',
+    'rag.web.search.perplexity_model': 'PERPLEXITY_MODEL',
+    'rag.web.search.perplexity_search_api_url': 'PERPLEXITY_SEARCH_API_URL',
+    'rag.web.search.perplexity_search_context_usage': 'PERPLEXITY_SEARCH_CONTEXT_USAGE',
+    'rag.web.search.result_count': 'WEB_SEARCH_RESULT_COUNT',
+    'rag.web.search.searchapi_api_key': 'SEARCHAPI_API_KEY',
+    'rag.web.search.searchapi_engine': 'SEARCHAPI_ENGINE',
+    'rag.web.search.searxng_language': 'SEARXNG_LANGUAGE',
+    'rag.web.search.searxng_query_url': 'SEARXNG_QUERY_URL',
+    'rag.web.search.serpapi_api_key': 'SERPAPI_API_KEY',
+    'rag.web.search.serpapi_engine': 'SERPAPI_ENGINE',
+    'rag.web.search.serper_api_key': 'SERPER_API_KEY',
+    'rag.web.search.serply_api_key': 'SERPLY_API_KEY',
+    'rag.web.search.serpstack_api_key': 'SERPSTACK_API_KEY',
+    'rag.web.search.serpstack_https': 'SERPSTACK_HTTPS',
+    'rag.web.search.sougou_api_sid': 'SOUGOU_API_SID',
+    'rag.web.search.sougou_api_sk': 'SOUGOU_API_SK',
+    'rag.web.search.tavily_api_key': 'TAVILY_API_KEY',
+    'rag.web.search.tavily_extract_depth': 'TAVILY_EXTRACT_DEPTH',
+    'rag.web.search.trust_env': 'WEB_SEARCH_TRUST_ENV',
+    'rag.web.search.yacy_password': 'YACY_PASSWORD',
+    'rag.web.search.yacy_query_url': 'YACY_QUERY_URL',
+    'rag.web.search.yacy_username': 'YACY_USERNAME',
+    'rag.web.search.yandex_web_search_api_key': 'YANDEX_WEB_SEARCH_API_KEY',
+    'rag.web.search.yandex_web_search_config': 'YANDEX_WEB_SEARCH_CONFIG',
+    'rag.web.search.yandex_web_search_url': 'YANDEX_WEB_SEARCH_URL',
+    'rag.web.search.youcom_api_key': 'YOUCOM_API_KEY',
+    'rag.youtube_loader_language': 'YOUTUBE_LOADER_LANGUAGE',
+    'rag.youtube_loader_proxy_url': 'YOUTUBE_LOADER_PROXY_URL',
+    # Tasks
+    'task.autocomplete.enable': 'ENABLE_AUTOCOMPLETE_GENERATION',
+    'task.autocomplete.input_max_length': 'AUTOCOMPLETE_GENERATION_INPUT_MAX_LENGTH',
+    'task.autocomplete.prompt_template': 'AUTOCOMPLETE_GENERATION_PROMPT_TEMPLATE',
+    'task.follow_up.enable': 'ENABLE_FOLLOW_UP_GENERATION',
+    'task.follow_up.prompt_template': 'FOLLOW_UP_GENERATION_PROMPT_TEMPLATE',
+    'task.image.prompt_template': 'IMAGE_PROMPT_GENERATION_PROMPT_TEMPLATE',
+    'task.model.default': 'TASK_MODEL',
+    'task.model.external': 'TASK_MODEL_EXTERNAL',
+    'task.query.prompt_template': 'QUERY_GENERATION_PROMPT_TEMPLATE',
+    'task.query.retrieval.enable': 'ENABLE_RETRIEVAL_QUERY_GENERATION',
+    'task.query.search.enable': 'ENABLE_SEARCH_QUERY_GENERATION',
+    'task.tags.enable': 'ENABLE_TAGS_GENERATION',
+    'task.tags.prompt_template': 'TAGS_GENERATION_PROMPT_TEMPLATE',
+    'task.title.enable': 'ENABLE_TITLE_GENERATION',
+    'task.title.prompt_template': 'TITLE_GENERATION_PROMPT_TEMPLATE',
+    'task.tools.prompt_template': 'TOOLS_FUNCTION_CALLING_PROMPT_TEMPLATE',
+    'task.voice.prompt.enable': 'ENABLE_VOICE_MODE_PROMPT',
+    'task.voice.prompt_template': 'VOICE_MODE_PROMPT_TEMPLATE',
+    # Misc
+    'terminal_server.connections': 'TERMINAL_SERVER_CONNECTIONS',
+    'tool_server.connections': 'TOOL_SERVER_CONNECTIONS',
+    'ui.banners': 'WEBUI_BANNERS',
+    'ui.default_group_id': 'DEFAULT_GROUP_ID',
+    'ui.default_locale': 'DEFAULT_LOCALE',
+    'ui.default_models': 'DEFAULT_MODELS',
+    'ui.default_pinned_models': 'DEFAULT_PINNED_MODELS',
+    'ui.default_user_role': 'DEFAULT_USER_ROLE',
+    'ui.enable_community_sharing': 'ENABLE_COMMUNITY_SHARING',
+    'ui.enable_login_form': 'ENABLE_LOGIN_FORM',
+    'ui.enable_message_rating': 'ENABLE_MESSAGE_RATING',
+    'ui.enable_password_change_form': 'ENABLE_PASSWORD_CHANGE_FORM',
+    'ui.enable_signup': 'ENABLE_SIGNUP',
+    'ui.enable_user_webhooks': 'ENABLE_USER_WEBHOOKS',
+    'ui.model_order_list': 'MODEL_ORDER_LIST',
+    'ui.pending_user_overlay_content': 'PENDING_USER_OVERLAY_CONTENT',
+    'ui.pending_user_overlay_title': 'PENDING_USER_OVERLAY_TITLE',
+    'ui.prompt_suggestions': 'DEFAULT_PROMPT_SUGGESTIONS',
+    'ui.watermark': 'RESPONSE_WATERMARK',
+    'user.permissions': 'USER_PERMISSIONS',
+    'users.enable_status': 'ENABLE_USER_STATUS',
+    'webhook_url': 'WEBHOOK_URL',
+    'webui.url': 'WEBUI_URL',
+}
+
+
+STORAGE_KEY_REWRITES = {
+    'oauth.refresh_token_include_scope': 'oauth.refresh_token.include_scope',
+    'rag.openai_api_base_url': 'rag.openai.api_base_url',
+    'rag.openai_api_key': 'rag.openai.api_key',
+    'rag.ollama.url': 'rag.ollama.base_url',
+    'rag.ollama.key': 'rag.ollama.api_key',
+    'oauth.oidc.avatar_claim': 'oauth.picture_claim',
+    'oauth.oidc.client_id': 'oauth.client_id',
+    'oauth.oidc.client_secret': 'oauth.client_secret',
+    'oauth.oidc.code_challenge_method': 'oauth.code_challenge_method',
+    'oauth.oidc.email_claim': 'oauth.email_claim',
+    'oauth.oidc.end_session_endpoint': 'oauth.end_session_endpoint',
+    'oauth.oidc.group_claim': 'oauth.group_claim',
+    'oauth.oidc.oauth_timeout': 'oauth.timeout',
+    'oauth.oidc.provider_name': 'oauth.provider_name',
+    'oauth.oidc.provider_url': 'oauth.provider_url',
+    'oauth.oidc.redirect_uri': 'oauth.redirect_uri',
+    'oauth.oidc.scopes': 'oauth.scopes',
+    'oauth.oidc.sub_claim': 'oauth.sub_claim',
+    'oauth.oidc.token_endpoint_auth_method': 'oauth.token_endpoint_auth_method',
+    'oauth.oidc.username_claim': 'oauth.username_claim',
+}
+
+
+LEGACY_KEY_TO_STORAGE_KEY = {
+    legacy_key: STORAGE_KEY_REWRITES.get(blob_path, blob_path) for blob_path, legacy_key in BLOB_PATH_TO_KEY.items()
+}
+
+
+def _walk_blob(data: dict, prefix: str = '') -> dict:
+    """Recursively walk a nested config blob, preserving known config values.
+
+    Some config values are intentionally dictionaries, e.g. OPENAI_API_CONFIGS
+    and OLLAMA_API_CONFIGS. Once the current path is a known config key, keep
+    that value intact instead of flattening its internals into orphaned rows.
+    """
+    result = {}
+    for key, value in data.items():
+        path = f'{prefix}{key}' if not prefix else f'{prefix}.{key}'
+        if path in BLOB_PATH_TO_KEY or path in LEGACY_KEY_TO_STORAGE_KEY:
+            result[path] = value
+        elif isinstance(value, dict):
+            result.update(_walk_blob(value, path))
+        else:
+            result[path] = value
+    return result
+
+
+def upgrade() -> None:
+    """Reshape config from single-row JSON blob to per-key rows."""
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    table_names = set(inspector.get_table_names())
+    config_columns = (
+        {column['name'] for column in inspector.get_columns('config')} if 'config' in table_names else set()
+    )
+    has_old_config = {'id', 'data'}.issubset(config_columns)
+    has_new_config = {'key', 'value'}.issubset(config_columns)
+
+    # Ad-hoc table reference for reading the old schema
+    old_config = sa.table(
+        'config',
+        sa.column('id', sa.Integer),
+        sa.column('data', sa.JSON),
+    )
+
+    # 1. Read existing blob
+    blob_data = {}
+    if has_old_config:
+        try:
+            result = conn.execute(sa.select(old_config.c.data).order_by(old_config.c.id.desc()).limit(1))
+            row = result.fetchone()
+            if row and row[0]:
+                raw = row[0]
+                blob_data = json.loads(raw) if isinstance(raw, str) else raw
+        except Exception:
+            pass  # Table might be partially migrated or empty
+
+    # 2. Preserve old blob table for rollback/inspection, then create per-key table.
+    if has_old_config:
+        if 'config_old' in table_names:
+            op.drop_table('config_old')
+        op.rename_table('config', 'config_old')
+
+    # 3. Create new per-key table
+    new_config = (
+        sa.table(
+            'config',
+            sa.column('key', sa.Text),
+            sa.column('value', sa.JSON()),
+            sa.column('updated_at', sa.BigInteger),
+        )
+        if has_new_config
+        else op.create_table(
+            'config',
+            sa.Column('key', sa.Text(), primary_key=True),
+            sa.Column('value', sa.JSON(), nullable=False),
+            sa.Column('updated_at', sa.BigInteger(), nullable=True),
+        )
+    )
+
+    # 4. Flatten blob and insert per-key rows
+    if blob_data:
+        flat = _walk_blob(blob_data)
+
+        # Keep stable dot-notation paths as the database keys.
+        # Known legacy env-style keys are rewritten to their dotted keys; unknown
+        # keys are still copied so custom/future config is not silently lost.
+        rows = {}
+        for blob_path, value in flat.items():
+            if blob_path in BLOB_PATH_TO_KEY:
+                storage_key = STORAGE_KEY_REWRITES.get(blob_path, blob_path)
+            elif blob_path in LEGACY_KEY_TO_STORAGE_KEY:
+                storage_key = LEGACY_KEY_TO_STORAGE_KEY[blob_path]
+            else:
+                storage_key = STORAGE_KEY_REWRITES.get(blob_path, blob_path)
+
+            if storage_key not in rows:
+                rows[storage_key] = value
+
+        # Batch insert via SQLAlchemy table reference
+        if rows:
+            now = int(time.time())
+            op.bulk_insert(
+                new_config,
+                [{'key': k, 'value': v, 'updated_at': now} for k, v in rows.items()],
+            )
+
+
+def downgrade() -> None:
+    """Restore preserved old single-row config table when available."""
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    table_names = set(inspector.get_table_names())
+
+    if 'config_old' in table_names:
+        if 'config' in table_names:
+            op.drop_table('config')
+        op.rename_table('config_old', 'config')
+        return
+
+    config_columns = (
+        {column['name'] for column in inspector.get_columns('config')} if 'config' in table_names else set()
+    )
+    has_per_key_config = {'key', 'value'}.issubset(config_columns)
+
+    blob_data = {}
+    if has_per_key_config:
+        config = sa.table(
+            'config',
+            sa.column('key', sa.Text),
+            sa.column('value', sa.JSON),
+        )
+        for key, value in conn.execute(sa.select(config.c.key, config.c.value)):
+            blob_data[key] = json.loads(value) if isinstance(value, str) else value
+        op.drop_table('config')
+
+    if 'config' in table_names and not has_per_key_config:
+        return
+
+    old_config = op.create_table(
+        'config',
+        sa.Column('id', sa.Integer(), primary_key=True),
+        sa.Column('data', sa.JSON(), nullable=False),
+        sa.Column('version', sa.Integer(), nullable=False, server_default='0'),
+        sa.Column('created_at', sa.DateTime(), nullable=False, server_default=sa.func.now()),
+        sa.Column('updated_at', sa.DateTime(), nullable=True),
+    )
+
+    if blob_data:
+        op.bulk_insert(old_config, [{'data': blob_data, 'version': 0}])

backend/open_webui/migrations/versions/42e2978c7933_add_memory_path_and_meta.py

@@ -0,0 +1,40 @@
+"""add memory path and meta
+
+Revision ID: 42e2978c7933
+Revises: 7b3f2a9c1d4e
+Create Date: 2026-06-29 05:35:50.565887
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+revision: str = '42e2978c7933'
+down_revision: Union[str, None] = '7b3f2a9c1d4e'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {column['name'] for column in inspector.get_columns('memory')}
+
+    if 'path' not in columns:
+        op.add_column('memory', sa.Column('path', sa.Text(), nullable=True))
+    if 'meta' not in columns:
+        op.add_column('memory', sa.Column('meta', sa.JSON(), nullable=True))
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {column['name'] for column in inspector.get_columns('memory')}
+
+    if 'meta' in columns:
+        op.drop_column('memory', 'meta')
+    if 'path' in columns:
+        op.drop_column('memory', 'path')

backend/open_webui/migrations/versions/4c5ce3d2f27f_add_context_summary_to_chat_message.py

@@ -0,0 +1,37 @@
+"""add context summary to chat message
+
+Revision ID: 4c5ce3d2f27f
+Revises: 3ff2c63645b8
+Create Date: 2026-06-18 23:48:08.310063
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '4c5ce3d2f27f'
+down_revision: Union[str, None] = '3ff2c63645b8'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {column['name'] for column in inspector.get_columns('chat_message')}
+
+    if 'context_summary' not in columns:
+        op.add_column('chat_message', sa.Column('context_summary', sa.Text(), nullable=True))
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {column['name'] for column in inspector.get_columns('chat_message')}
+
+    if 'context_summary' in columns:
+        op.drop_column('chat_message', 'context_summary')

backend/open_webui/migrations/versions/55f1302ac17c_add_memory_id_user_id_covering_index.py

@@ -0,0 +1,36 @@
+"""Add memory (id, user_id) covering index
+
+Revision ID: 55f1302ac17c
+Revises: b0018471bbbe
+Create Date: 2026-07-24 00:00:00.000000
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+
+revision: str = '55f1302ac17c'
+down_revision: Union[str, None] = 'b0018471bbbe'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    indexes = {index['name'] for index in inspector.get_indexes('memory')}
+
+    if 'ix_memory_id_user_id' not in indexes:
+        op.create_index('ix_memory_id_user_id', 'memory', ['id', 'user_id'])
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    indexes = {index['name'] for index in inspector.get_indexes('memory')}
+
+    if 'ix_memory_id_user_id' in indexes:
+        op.drop_index('ix_memory_id_user_id', table_name='memory')

backend/open_webui/migrations/versions/7b3f2a9c1d4e_add_memory_type.py

@@ -0,0 +1,44 @@
+"""add memory type
+
+Revision ID: 7b3f2a9c1d4e
+Revises: 4c5ce3d2f27f
+Create Date: 2026-06-25 00:00:00.000000
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+revision: str = '7b3f2a9c1d4e'
+down_revision: Union[str, None] = '4c5ce3d2f27f'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {column['name'] for column in inspector.get_columns('memory')}
+    indexes = {index['name'] for index in inspector.get_indexes('memory')}
+
+    if 'type' not in columns:
+        op.add_column('memory', sa.Column('type', sa.String(), server_default='context', nullable=False))
+
+    if 'ix_memory_type' not in indexes:
+        op.create_index('ix_memory_type', 'memory', ['type'])
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {column['name'] for column in inspector.get_columns('memory')}
+    indexes = {index['name'] for index in inspector.get_indexes('memory')}
+
+    if 'ix_memory_type' in indexes:
+        op.drop_index('ix_memory_type', table_name='memory')
+
+    if 'type' in columns:
+        op.drop_column('memory', 'type')

backend/open_webui/migrations/versions/856c5b02fb54_add_chat_message_meta.py

@@ -0,0 +1,25 @@
+"""add chat message meta
+
+Revision ID: 856c5b02fb54
+Revises: 42e2978c7933
+Create Date: 2026-07-16 01:39:39.291935
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = '856c5b02fb54'
+down_revision: Union[str, None] = '42e2978c7933'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.add_column('chat_message', sa.Column('meta', sa.JSON(), nullable=True))
+
+
+def downgrade() -> None:
+    op.drop_column('chat_message', 'meta')

backend/open_webui/migrations/versions/959eaac8f909_add_automation_folder_id.py

@@ -0,0 +1,54 @@
+"""add automation folder id
+
+Revision ID: 959eaac8f909
+Revises: 55f1302ac17c
+Create Date: 2026-07-26 19:19:31.345756
+
+"""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+from alembic import context, op
+
+# revision identifiers, used by Alembic.
+revision: str = '959eaac8f909'
+down_revision: str | None = '55f1302ac17c'
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+
+def upgrade() -> None:
+    if context.is_offline_mode():
+        op.add_column('automation', sa.Column('folder_id', sa.Text(), nullable=True))
+        op.create_index('ix_automation_user_folder', 'automation', ['user_id', 'folder_id'])
+        return
+
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {col['name'] for col in inspector.get_columns('automation')}
+    indexes = {index['name'] for index in inspector.get_indexes('automation')}
+
+    if 'folder_id' not in columns:
+        op.add_column('automation', sa.Column('folder_id', sa.Text(), nullable=True))
+
+    if 'ix_automation_user_folder' not in indexes:
+        op.create_index('ix_automation_user_folder', 'automation', ['user_id', 'folder_id'])
+
+
+def downgrade() -> None:
+    if context.is_offline_mode():
+        op.drop_index('ix_automation_user_folder', table_name='automation')
+        op.drop_column('automation', 'folder_id')
+        return
+
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = {col['name'] for col in inspector.get_columns('automation')}
+    indexes = {index['name'] for index in inspector.get_indexes('automation')}
+
+    if 'ix_automation_user_folder' in indexes:
+        op.drop_index('ix_automation_user_folder', table_name='automation')
+
+    if 'folder_id' in columns:
+        op.drop_column('automation', 'folder_id')

backend/open_webui/migrations/versions/9a1b2c3d4e5f_add_current_message_id_to_chat.py

@@ -0,0 +1,219 @@
+"""add current_message_id to chat
+
+Revision ID: 9a1b2c3d4e5f
+Revises: 856c5b02fb54
+Create Date: 2026-07-23 00:00:00.000000
+
+"""
+
+import json
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = '9a1b2c3d4e5f'
+down_revision: Union[str, None] = '856c5b02fb54'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+BATCH_SIZE = 150
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = [col['name'] for col in inspector.get_columns('chat')]
+    if 'current_message_id' not in columns:
+        op.add_column('chat', sa.Column('current_message_id', sa.Text(), nullable=True))
+
+    chat = sa.table(
+        'chat',
+        sa.column('id', sa.String()),
+        sa.column('chat', sa.Text()),
+        sa.column('current_message_id', sa.Text()),
+    )
+    chat_message = sa.table(
+        'chat_message',
+        sa.column('id', sa.Text()),
+        sa.column('chat_id', sa.Text()),
+        sa.column('parent_id', sa.Text()),
+        sa.column('created_at', sa.BigInteger()),
+    )
+
+    has_chat_message = 'chat_message' in inspector.get_table_names()
+    result = conn.execute(
+        sa.select(chat.c.id, chat.c.chat, chat.c.current_message_id).execution_options(
+            yield_per=BATCH_SIZE,
+            stream_results=True,
+        )
+    )
+
+    while True:
+        rows = result.fetchmany(BATCH_SIZE)
+        if not rows:
+            break
+
+        batch_chat_ids: list[str] = []
+        candidates_by_chat: dict[str, list[str]] = {}
+        current_by_chat: dict[str, str | None] = {}
+        json_messages_by_chat: dict[str, dict[str, dict]] = {}
+
+        for row in rows:
+            values = row._mapping
+            chat_id = values['id']
+            prefix = f'{chat_id}-'
+            batch_chat_ids.append(chat_id)
+            current_by_chat[chat_id] = values['current_message_id']
+
+            chat_data = {}
+            if isinstance(values['chat'], dict):
+                chat_data = values['chat']
+            elif isinstance(values['chat'], str):
+                try:
+                    parsed = json.loads(values['chat'])
+                    chat_data = parsed if isinstance(parsed, dict) else {}
+                except (TypeError, ValueError, json.JSONDecodeError):
+                    pass
+
+            history = chat_data.get('history') if isinstance(chat_data.get('history'), dict) else {}
+            candidates_by_chat[chat_id] = []
+            for candidate in (
+                values['current_message_id'],
+                history.get('currentId'),
+                chat_data.get('currentId'),
+                chat_data.get('branchPointMessageId'),
+            ):
+                if not isinstance(candidate, str) or not candidate:
+                    continue
+                candidate = candidate[len(prefix) :] if candidate.startswith(prefix) else candidate
+                if candidate not in candidates_by_chat[chat_id]:
+                    candidates_by_chat[chat_id].append(candidate)
+
+            messages = history.get('messages') if isinstance(history.get('messages'), dict) else {}
+            if not messages and isinstance(chat_data.get('messages'), list):
+                messages = {
+                    message['id']: message
+                    for message in chat_data['messages']
+                    if isinstance(message, dict) and message.get('id')
+                }
+            if messages:
+                json_messages_by_chat[chat_id] = {
+                    message_id: {
+                        'parent_id': message.get('parentId') if isinstance(message, dict) else None,
+                        'created_at': message.get('timestamp', 0) if isinstance(message, dict) else 0,
+                    }
+                    for message_id, message in messages.items()
+                }
+
+        resolved: dict[str, str] = {}
+
+        if has_chat_message:
+            candidate_ids = {
+                f'{chat_id}-{candidate}'
+                for chat_id, candidates in candidates_by_chat.items()
+                for candidate in candidates
+            }
+            if candidate_ids:
+                valid_by_chat: dict[str, set[str]] = {}
+                for row in conn.execute(
+                    sa.select(chat_message.c.chat_id, chat_message.c.id).where(
+                        chat_message.c.chat_id.in_(batch_chat_ids),
+                        chat_message.c.id.in_(candidate_ids),
+                    )
+                ):
+                    values = row._mapping
+                    chat_id = values['chat_id']
+                    prefix = f'{chat_id}-'
+                    message_id = values['id']
+                    if message_id and message_id.startswith(prefix):
+                        message_id = message_id[len(prefix) :]
+                    if message_id:
+                        valid_by_chat.setdefault(chat_id, set()).add(message_id)
+                for chat_id, candidates in candidates_by_chat.items():
+                    valid_ids = valid_by_chat.get(chat_id, set())
+                    for candidate in candidates:
+                        if candidate in valid_ids:
+                            resolved[chat_id] = candidate
+                            break
+
+            unresolved_chat_ids = [chat_id for chat_id in batch_chat_ids if chat_id not in resolved]
+            messages_by_chat: dict[str, dict[str, dict]] = {}
+            if unresolved_chat_ids:
+                for row in conn.execute(
+                    sa.select(
+                        chat_message.c.chat_id,
+                        chat_message.c.id,
+                        chat_message.c.parent_id,
+                        chat_message.c.created_at,
+                    ).where(chat_message.c.chat_id.in_(unresolved_chat_ids))
+                ):
+                    values = row._mapping
+                    chat_id = values['chat_id']
+                    prefix = f'{chat_id}-'
+                    message_id = values['id']
+                    if message_id and message_id.startswith(prefix):
+                        message_id = message_id[len(prefix) :]
+                    if not message_id:
+                        continue
+                    parent_id = values['parent_id']
+                    if parent_id and parent_id.startswith(prefix):
+                        parent_id = parent_id[len(prefix) :]
+                    messages_by_chat.setdefault(chat_id, {})[message_id] = {
+                        'parent_id': parent_id,
+                        'created_at': values['created_at'] or 0,
+                    }
+
+            for chat_id, messages in messages_by_chat.items():
+                parent_ids = {
+                    message['parent_id'] for message in messages.values() if message.get('parent_id') in messages
+                }
+                leaf_ids = [message_id for message_id in messages if message_id not in parent_ids]
+                resolved[chat_id] = max(
+                    leaf_ids or list(messages),
+                    key=lambda message_id: messages[message_id].get('created_at') or 0,
+                )
+
+        for chat_id in batch_chat_ids:
+            if chat_id in resolved:
+                continue
+
+            messages = json_messages_by_chat.get(chat_id, {})
+            valid_candidate = next(
+                (candidate for candidate in candidates_by_chat[chat_id] if candidate in messages),
+                None,
+            )
+            if valid_candidate:
+                resolved[chat_id] = valid_candidate
+            elif messages:
+                parent_ids = {
+                    message['parent_id'] for message in messages.values() if message.get('parent_id') in messages
+                }
+                leaf_ids = [message_id for message_id in messages if message_id not in parent_ids]
+                resolved[chat_id] = max(
+                    leaf_ids or list(messages),
+                    key=lambda message_id: messages[message_id].get('created_at') or 0,
+                )
+
+        updates = [
+            {'chat_id': chat_id, 'current_message_id': message_id}
+            for chat_id, message_id in resolved.items()
+            if message_id and message_id != current_by_chat.get(chat_id)
+        ]
+        if updates:
+            conn.execute(
+                sa.update(chat)
+                .where(chat.c.id == sa.bindparam('update_chat_id'))
+                .values(current_message_id=sa.bindparam('update_current_message_id')),
+                [
+                    {
+                        'update_chat_id': row['chat_id'],
+                        'update_current_message_id': row['current_message_id'],
+                    }
+                    for row in updates
+                ],
+            )
+
+
+def downgrade() -> None:
+    op.drop_column('chat', 'current_message_id')

backend/open_webui/migrations/versions/b0018471bbbe_add_user_variables.py

@@ -0,0 +1,32 @@
+"""add user variables
+
+Revision ID: b0018471bbbe
+Revises: c49178636c78
+Create Date: 2026-07-24 01:21:46.457057
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision: str = 'b0018471bbbe'
+down_revision: Union[str, None] = 'c49178636c78'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = [col['name'] for col in inspector.get_columns('user')]
+
+    if 'variables' not in columns:
+        op.add_column('user', sa.Column('variables', sa.JSON(), nullable=True))
+
+
+def downgrade() -> None:
+    op.drop_column('user', 'variables')

backend/open_webui/migrations/versions/b10670c03dd5_update_user_table.py

@@ -72,7 +72,6 @@ def _convert_column_to_json(table: str, column: str):
     dialect = conn.dialect.name
 
     t = sa.table(table, sa.column('id', sa.Text), sa.column(column, sa.Text))
-    t_json = sa.column(f'{column}_json', sa.JSON)
 
     # SQLite cannot ALTER COLUMN → must recreate column
     if dialect == 'sqlite':
@@ -90,9 +89,9 @@ def _convert_column_to_json(table: str, column: str):
                     parsed = None
 
             conn.execute(
-                sa.update(sa.table(table, sa.column('id'), t_json))
+                sa.update(sa.table(table, sa.column('id'), sa.column(f'{column}_json', sa.JSON)))
                 .where(sa.column('id') == uid)
-                .values({f'{column}_json': json.dumps(parsed) if parsed else None})
+                .values({f'{column}_json': parsed})
             )
 
         op.drop_column(table, column)
@@ -112,8 +111,7 @@ def _convert_column_to_text(table: str, column: str):
     conn = op.get_bind()
     dialect = conn.dialect.name
 
-    t = sa.table(table, sa.column('id', sa.Text), sa.column(column))
-    t_text = sa.column(f'{column}_text', sa.Text)
+    t = sa.table(table, sa.column('id', sa.Text), sa.column(column, sa.JSON))
 
     if dialect == 'sqlite':
         op.add_column(table, sa.Column(f'{column}_text', sa.Text(), nullable=True))
@@ -122,9 +120,9 @@ def _convert_column_to_text(table: str, column: str):
 
         for uid, raw in rows:
             conn.execute(
-                sa.update(sa.table(table, sa.column('id'), t_text))
+                sa.update(sa.table(table, sa.column('id'), sa.column(f'{column}_text', sa.Text)))
                 .where(sa.column('id') == uid)
-                .values({f'{column}_text': json.dumps(raw) if raw else None})
+                .values({f'{column}_text': json.dumps(raw) if raw is not None else None})
             )
 
         op.drop_column(table, column)

backend/open_webui/migrations/versions/c49178636c78_add_chat_variables.py

@@ -0,0 +1,32 @@
+"""add chat variables
+
+Revision ID: c49178636c78
+Revises: 9a1b2c3d4e5f
+Create Date: 2026-07-23 23:33:45.497453
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision: str = 'c49178636c78'
+down_revision: Union[str, None] = '9a1b2c3d4e5f'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    columns = [col['name'] for col in inspector.get_columns('chat')]
+
+    if 'variables' not in columns:
+        op.add_column('chat', sa.Column('variables', sa.JSON(), nullable=True))
+
+
+def downgrade() -> None:
+    op.drop_column('chat', 'variables')

backend/open_webui/migrations/versions/f0bd01a18a3d_add_unique_normalized_user_email_index.py

@@ -0,0 +1,84 @@
+"""add unique normalized user email index
+
+Revision ID: f0bd01a18a3d
+Revises: 959eaac8f909
+Create Date: 2026-07-27 04:41:12.708743
+
+"""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+from alembic import context, op
+
+
+# revision identifiers, used by Alembic.
+revision: str = 'f0bd01a18a3d'
+down_revision: str | None = '959eaac8f909'
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+INDEX_NAME = 'uq_user_email_lower'
+EMAIL_IS_NOT_NULL = sa.text('email IS NOT NULL')
+LOWER_EMAIL = sa.text('lower(email)')
+
+
+def _index_exists() -> bool:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    return INDEX_NAME in {index['name'] for index in inspector.get_indexes('user')}
+
+
+def _duplicate_emails() -> list:
+    conn = op.get_bind()
+    return conn.execute(
+        sa.text(
+            """
+            SELECT lower(email) AS email, count(*) AS duplicate_count
+            FROM "user"
+            WHERE email IS NOT NULL
+            GROUP BY lower(email)
+            HAVING count(*) > 1
+            ORDER BY lower(email)
+            """
+        )
+    ).fetchall()
+
+
+def _create_index() -> None:
+    op.create_index(
+        INDEX_NAME,
+        'user',
+        [LOWER_EMAIL],
+        unique=True,
+        postgresql_where=EMAIL_IS_NOT_NULL,
+        sqlite_where=EMAIL_IS_NOT_NULL,
+    )
+
+
+def upgrade() -> None:
+    if context.is_offline_mode():
+        _create_index()
+        return
+
+    if _index_exists():
+        return
+
+    duplicates = _duplicate_emails()
+    if duplicates:
+        details = ', '.join(f'{row.email} (x{row.duplicate_count})' for row in duplicates)
+        raise RuntimeError(
+            'Cannot add unique normalized user email index because duplicate emails exist: '
+            f'{details}. Merge or remove the duplicate users and rerun migrations.'
+        )
+
+    _create_index()
+
+
+def downgrade() -> None:
+    if context.is_offline_mode():
+        op.drop_index(INDEX_NAME, table_name='user')
+        return
+
+    if _index_exists():
+        op.drop_index(INDEX_NAME, table_name='user')

backend/open_webui/models/access_grants.py

@@ -11,6 +11,11 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 log = logging.getLogger(__name__)
 
+PRINCIPAL_TYPE_ANYONE = 'anyone'
+PRINCIPAL_TYPE_GROUP = 'group'
+PRINCIPAL_TYPE_USER = 'user'
+WILDCARD_PRINCIPAL_ID = '*'
+
 
 ####################
 # AccessGrant DB Schema
@@ -23,7 +28,7 @@ class AccessGrant(Base):
     id = Column(Text, primary_key=True)
     resource_type = Column(Text, nullable=False)  # "knowledge", "model", "prompt", "tool", "note", "channel", "file"
     resource_id = Column(Text, nullable=False)
-    principal_type = Column(Text, nullable=False)  # "user" or "group"
+    principal_type = Column(Text, nullable=False)  # "user", "group", or "anyone"
     principal_id = Column(Text, nullable=False)  # user_id, group_id, or "*" (wildcard for public)
     permission = Column(Text, nullable=False)  # "read" or "write"
     created_at = Column(BigInteger, nullable=False)
@@ -163,12 +168,14 @@ def normalize_access_grants(access_grants: Optional[list]) -> list[dict]:
         principal_id = grant.get('principal_id')
         permission = grant.get('permission')
 
-        if principal_type not in ('user', 'group'):
+        if principal_type not in (PRINCIPAL_TYPE_USER, PRINCIPAL_TYPE_GROUP, PRINCIPAL_TYPE_ANYONE):
             continue
         if permission not in ('read', 'write'):
             continue
         if not isinstance(principal_id, str) or not principal_id:
             continue
+        if principal_type == PRINCIPAL_TYPE_ANYONE and (principal_id != WILDCARD_PRINCIPAL_ID or permission != 'read'):
+            continue
 
         key = (principal_type, principal_id, permission)
         deduped[key] = {
@@ -186,7 +193,11 @@ def has_public_read_access_grant(access_grants: Optional[list]) -> bool:
     Returns True when a direct grant list includes wildcard public-read.
     """
     for grant in normalize_access_grants(access_grants):
-        if grant['principal_type'] == 'user' and grant['principal_id'] == '*' and grant['permission'] == 'read':
+        if (
+            grant['principal_type'] == PRINCIPAL_TYPE_USER
+            and grant['principal_id'] == WILDCARD_PRINCIPAL_ID
+            and grant['permission'] == 'read'
+        ):
             return True
     return False
 
@@ -196,7 +207,25 @@ def has_public_write_access_grant(access_grants: Optional[list]) -> bool:
     Returns True when a direct grant list includes wildcard public-write.
     """
     for grant in normalize_access_grants(access_grants):
-        if grant['principal_type'] == 'user' and grant['principal_id'] == '*' and grant['permission'] == 'write':
+        if (
+            grant['principal_type'] == PRINCIPAL_TYPE_USER
+            and grant['principal_id'] == WILDCARD_PRINCIPAL_ID
+            and grant['permission'] == 'write'
+        ):
+            return True
+    return False
+
+
+def has_anyone_read_access_grant(access_grants: Optional[list]) -> bool:
+    """
+    Returns True when a direct grant list includes no-auth anyone-read.
+    """
+    for grant in normalize_access_grants(access_grants):
+        if (
+            grant['principal_type'] == PRINCIPAL_TYPE_ANYONE
+            and grant['principal_id'] == WILDCARD_PRINCIPAL_ID
+            and grant['permission'] == 'read'
+        ):
             return True
     return False
 
@@ -206,7 +235,7 @@ def has_user_access_grant(access_grants: Optional[list]) -> bool:
     Returns True when a direct grant list includes any non-wildcard user grant.
     """
     for grant in normalize_access_grants(access_grants):
-        if grant['principal_type'] == 'user' and grant['principal_id'] != '*':
+        if grant['principal_type'] == PRINCIPAL_TYPE_USER and grant['principal_id'] != WILDCARD_PRINCIPAL_ID:
             return True
     return False
 
@@ -223,12 +252,27 @@ def strip_user_access_grants(access_grants: Optional[list]) -> list:
         for grant in access_grants
         if not (
             (grant.get('principal_type') if isinstance(grant, dict) else getattr(grant, 'principal_type', None))
-            == 'user'
-            and (grant.get('principal_id') if isinstance(grant, dict) else getattr(grant, 'principal_id', None)) != '*'
+            == PRINCIPAL_TYPE_USER
+            and (grant.get('principal_id') if isinstance(grant, dict) else getattr(grant, 'principal_id', None))
+            != WILDCARD_PRINCIPAL_ID
         )
     ]
 
 
+def strip_anyone_access_grants(access_grants: Optional[list]) -> list:
+    """
+    Remove no-auth anyone grants from the list.
+    """
+    if not access_grants:
+        return []
+    return [
+        grant
+        for grant in access_grants
+        if (grant.get('principal_type') if isinstance(grant, dict) else getattr(grant, 'principal_type', None))
+        != PRINCIPAL_TYPE_ANYONE
+    ]
+
+
 def grants_to_access_control(grants: list) -> Optional[dict]:
     """
     Convert a list of grant objects (AccessGrantModel or AccessGrantResponse)
@@ -316,7 +360,6 @@ class AccessGrantsTable:
             )
             db.add(grant)
             await db.commit()
-            await db.refresh(grant)
             return AccessGrantModel.model_validate(grant)
 
     async def revoke_access(
@@ -494,6 +537,28 @@ class AccessGrantsTable:
                 result_dict[g.resource_id].append(AccessGrantModel.model_validate(g))
             return result_dict
 
+    async def has_anyone_access(
+        self,
+        resource_type: str,
+        resource_id: str,
+        permission: str = 'read',
+        db: Optional[AsyncSession] = None,
+    ) -> bool:
+        """Check for a no-auth anyone:* grant. Callers must opt in explicitly."""
+        async with get_async_db_context(db) as db:
+            result = await db.execute(
+                select(AccessGrant)
+                .filter(
+                    AccessGrant.resource_type == resource_type,
+                    AccessGrant.resource_id == resource_id,
+                    AccessGrant.principal_type == PRINCIPAL_TYPE_ANYONE,
+                    AccessGrant.principal_id == WILDCARD_PRINCIPAL_ID,
+                    AccessGrant.permission == permission,
+                )
+                .limit(1)
+            )
+            return result.scalars().first() is not None
+
     async def has_access(
         self,
         user_id: str,

backend/open_webui/models/auths.py

@@ -6,15 +6,22 @@ import logging
 import uuid
 from typing import Optional
 
+import bcrypt
 from open_webui.internal.db import Base, JSONField, get_async_db_context
 from open_webui.models.users import User, UserModel, UserProfileImageResponse, Users
 from open_webui.utils.validate import validate_profile_image_url
 from pydantic import BaseModel, field_validator
 from sqlalchemy import Boolean, Column, String, Text, delete, select, update
+from sqlalchemy.exc import IntegrityError
 from sqlalchemy.ext.asyncio import AsyncSession
 
 log = logging.getLogger(__name__)
 
+# Pre-computed hash verified on signin paths that lack a real credential
+# (unknown user, inactive account) so response timing cannot reveal
+# whether an account exists (CWE-208).
+PLACEHOLDER_HASH = bcrypt.hashpw(b'placeholder', bcrypt.gensalt()).decode('utf-8')
+
 
 class Auth(Base):  # credential ↔ user linkage
     """Maps a user ID to an email/password pair with an active flag."""
@@ -118,18 +125,20 @@ class AuthsTable:
             )
             session.add(credential)
 
-            created_user = await Users.insert_new_user(
-                new_id,
-                name,
-                email,
-                profile_image_url,
-                role,
-                oauth=oauth,
-                db=session,
-            )
-            # persist both records and reload generated defaults
-            await session.commit()
-            await session.refresh(credential)
+            try:
+                created_user = await Users.insert_new_user(
+                    new_id,
+                    name,
+                    email,
+                    profile_image_url,
+                    role,
+                    oauth=oauth,
+                    db=session,
+                )
+                await session.commit()
+            except IntegrityError:
+                await session.rollback()
+                raise
             return created_user if credential and created_user else None
 
     async def authenticate_user(
@@ -142,13 +151,15 @@ class AuthsTable:
         log.info('authenticate_user: %s', email)
         resolved = await Users.get_user_by_email(email, db=db)
         if not resolved:
+            await verify_password(PLACEHOLDER_HASH)
             return
         # load the credential row and verify the password hash
         async with get_async_db_context(db) as session:
             credential = await session.get(Auth, resolved.id)
             if not credential or not credential.active:
+                await verify_password(PLACEHOLDER_HASH)
                 return
-            if not verify_password(credential.password):
+            if not await verify_password(credential.password):
                 return
             return resolved
 

backend/open_webui/models/automations.py

@@ -21,6 +21,7 @@ class Automation(Base):
 
     id = Column(Text, primary_key=True)
     user_id = Column(Text, nullable=False)
+    folder_id = Column(Text, nullable=True)
     name = Column(Text, nullable=False)
     data = Column(JSON, nullable=False)  # {prompt, model_id, rrule}
     meta = Column(JSON, nullable=True)
@@ -31,7 +32,10 @@ class Automation(Base):
     created_at = Column(BigInteger, nullable=False)
     updated_at = Column(BigInteger, nullable=False)
 
-    __table_args__ = (Index('ix_automation_next_run', 'next_run_at'),)
+    __table_args__ = (
+        Index('ix_automation_next_run', 'next_run_at'),
+        Index('ix_automation_user_folder', 'user_id', 'folder_id'),
+    )
 
 
 class AutomationRun(Base):
@@ -72,6 +76,7 @@ class AutomationModel(BaseModel):
 
     id: str
     user_id: str
+    folder_id: Optional[str] = None
     name: str
     data: dict
     meta: Optional[dict] = None
@@ -96,6 +101,7 @@ class AutomationRunModel(BaseModel):
 
 class AutomationForm(BaseModel):
     name: str
+    folder_id: Optional[str] = None
     data: AutomationData
     meta: Optional[dict] = None
     is_active: Optional[bool] = True
@@ -129,6 +135,7 @@ class AutomationTable:
             row = Automation(
                 id=str(uuid4()),
                 user_id=user_id,
+                folder_id=form.folder_id,
                 name=form.name,
                 data=form.data.model_dump(),
                 meta=form.meta,
@@ -139,7 +146,6 @@ class AutomationTable:
             )
             db.add(row)
             await db.commit()
-            await db.refresh(row)
             return AutomationModel.model_validate(row)
 
     async def count_by_user(self, user_id: str, db: Optional[AsyncSession] = None) -> int:
@@ -165,6 +171,7 @@ class AutomationTable:
         user_id: str,
         query: Optional[str] = None,
         status: Optional[str] = None,
+        folder_id: Optional[str] = None,
         skip: int = 0,
         limit: int = 30,
         db: Optional[AsyncSession] = None,
@@ -172,6 +179,9 @@ class AutomationTable:
         async with get_async_db_context(db) as db:
             stmt = select(Automation).filter_by(user_id=user_id)
 
+            if folder_id is not None:
+                stmt = stmt.filter(Automation.folder_id == (folder_id or None))
+
             if query:
                 search = f'%{query}%'
                 # Search in name and prompt inside JSON data
@@ -217,6 +227,7 @@ class AutomationTable:
             if not row:
                 return None
             row.name = form.name
+            row.folder_id = form.folder_id
             row.data = form.data.model_dump()
             row.meta = form.meta
             if form.is_active is not None:
@@ -224,9 +235,25 @@ class AutomationTable:
             row.next_run_at = next_run_at
             row.updated_at = int(time.time_ns())
             await db.commit()
-            await db.refresh(row)
             return AutomationModel.model_validate(row)
 
+    async def clear_folder_ids(
+        self,
+        user_id: str,
+        folder_ids: list[str],
+        db: Optional[AsyncSession] = None,
+    ) -> int:
+        if not folder_ids:
+            return 0
+        async with get_async_db_context(db) as db:
+            result = await db.execute(
+                update(Automation)
+                .where(Automation.user_id == user_id, Automation.folder_id.in_(folder_ids))
+                .values(folder_id=None, updated_at=int(time.time_ns()))
+            )
+            await db.commit()
+            return result.rowcount or 0
+
     async def toggle(
         self,
         id: str,
@@ -241,7 +268,6 @@ class AutomationTable:
             row.next_run_at = next_run_at if row.is_active else None
             row.updated_at = int(time.time_ns())
             await db.commit()
-            await db.refresh(row)
             return AutomationModel.model_validate(row)
 
     async def delete(self, id: str, db: Optional[AsyncSession] = None) -> bool:
@@ -324,7 +350,6 @@ class AutomationRunTable:
             )
             db.add(row)
             await db.commit()
-            await db.refresh(row)
             return AutomationRunModel.model_validate(row)
 
     async def get_latest(self, automation_id: str, db: Optional[AsyncSession] = None) -> Optional[AutomationRunModel]:

backend/open_webui/models/calendar.py

@@ -241,11 +241,11 @@ class CalendarTable:
         access_grants: Optional[list[AccessGrantModel]] = None,
         db: Optional[AsyncSession] = None,
     ) -> CalendarModel:
-        cal_data = CalendarModel.model_validate(cal).model_dump(exclude={'access_grants'})
-        cal_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(cal_data['id'], db=db)
+        calendar_model = CalendarModel.model_validate(cal)
+        calendar_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(calendar_model.id, db=db)
         )
-        return CalendarModel.model_validate(cal_data)
+        return calendar_model
 
     async def get_or_create_defaults(self, user_id: str, db: Optional[AsyncSession] = None) -> list[CalendarModel]:
         """Return user's calendars, creating 'Personal' default if none exist."""
@@ -500,9 +500,12 @@ class CalendarEventTable:
                 # Filter to requested calendars only
                 accessible_cal_ids = [c for c in accessible_cal_ids if c in calendar_ids]
 
-            # Also get event IDs where user is an attendee
+            # Also get event IDs where the user is an attendee, excluding invites they declined
             attendee_event_ids_result = await db.execute(
-                select(CalendarEventAttendee.event_id).filter(CalendarEventAttendee.user_id == user_id)
+                select(CalendarEventAttendee.event_id).filter(
+                    CalendarEventAttendee.user_id == user_id,
+                    CalendarEventAttendee.status != 'declined',
+                )
             )
             attendee_event_ids = [r[0] for r in attendee_event_ids_result.all()]
 
@@ -764,22 +767,32 @@ class CalendarEventAttendeeTable:
     async def set_attendees(
         self, event_id: str, attendees: list[dict], db: Optional[AsyncSession] = None
     ) -> list[CalendarEventAttendeeModel]:
-        """Replace all attendees for an event.
+        """Replace all attendees for an event ({user_id, meta?} per dict).
 
-        Each dict in attendees: {user_id: str, status?: str, meta?: dict}
+        RSVP status is the attendee's alone to set (via update_rsvp): an existing
+        attendee keeps their status, a newly added one starts 'pending'. A
+        caller-supplied status is ignored so an organiser cannot set it for others.
         """
         async with get_async_db_context(db) as db:
+            existing_status = {
+                row.user_id: row.status
+                for row in (
+                    await db.execute(select(CalendarEventAttendee).filter(CalendarEventAttendee.event_id == event_id))
+                ).scalars()
+            }
+
             # Remove existing
             await db.execute(delete(CalendarEventAttendee).filter(CalendarEventAttendee.event_id == event_id))
 
             now = int(time.time_ns())
             models = []
             for att in attendees:
+                user_id = att['user_id']
                 row = CalendarEventAttendee(
                     id=str(uuid4()),
                     event_id=event_id,
-                    user_id=att['user_id'],
-                    status=att.get('status', 'pending'),
+                    user_id=user_id,
+                    status=existing_status.get(user_id, 'pending'),
                     meta=att.get('meta'),
                     created_at=now,
                     updated_at=now,

backend/open_webui/models/channels.py

@@ -266,11 +266,11 @@ class ChannelTable:
         access_grants: Optional[list[AccessGrantModel]] = None,
         db: Optional[AsyncSession] = None,
     ) -> ChannelModel:
-        channel_data = ChannelModel.model_validate(channel).model_dump(exclude={'access_grants'})
-        channel_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(channel_data['id'], db=db)
+        channel_model = ChannelModel.model_validate(channel)
+        channel_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(channel_model.id, db=db)
         )
-        return ChannelModel.model_validate(channel_data)
+        return channel_model
 
     async def _collect_unique_user_ids(
         self,
@@ -869,7 +869,6 @@ class ChannelTable:
                 result = ChannelFile(**channel_file.model_dump())
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 if result:
                     return ChannelFileModel.model_validate(result)
                 else:

backend/open_webui/models/chat_messages.py

@@ -1,10 +1,15 @@
 import json
 import time
 import uuid
+from collections import Counter
+from datetime import datetime, timedelta
 from typing import Any, Optional
+from zoneinfo import ZoneInfo, ZoneInfoNotFoundError
 
+from sqlalchemy import select, delete, func, cast, Integer, distinct
+from sqlalchemy.ext.asyncio import AsyncSession
 from open_webui.internal.db import Base, get_async_db_context
-from open_webui.utils.response import normalize_usage
+from open_webui.utils.response import merge_usage, normalize_usage
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import (
     JSON,
@@ -45,6 +50,17 @@ def _normalize_timestamp(timestamp: int) -> float:
     return timestamp
 
 
+def _timezone(tz: Optional[str]) -> ZoneInfo:
+    try:
+        return ZoneInfo(tz or 'UTC')
+    except ZoneInfoNotFoundError:
+        return ZoneInfo('UTC')
+
+
+def _date_key(timestamp: int, tz: ZoneInfo) -> str:
+    return datetime.fromtimestamp(_normalize_timestamp(timestamp), tz=tz).strftime('%Y-%m-%d')
+
+
 def get_usage(data: dict) -> Optional[dict]:
     """Extract and normalize usage from message data."""
     usage = data.get('usage') or (data.get('info') or {}).get('usage')
@@ -70,6 +86,40 @@ def _token_columns(dialect: str):
     )
 
 
+def _extract_tool_names(value: Any) -> list[str]:
+    names: list[str] = []
+
+    def add(name: Any):
+        if isinstance(name, str):
+            cleaned = name.strip()
+            if cleaned and len(cleaned) <= 128:
+                names.append(cleaned)
+
+    def walk(item: Any):
+        if isinstance(item, list):
+            for child in item:
+                walk(child)
+            return
+
+        if not isinstance(item, dict):
+            return
+
+        item_type = str(item.get('type') or '')
+        looks_like_tool = 'tool' in item_type or item_type in {'function_call', 'function_call_output'}
+        if looks_like_tool:
+            add(item.get('name') or item.get('tool_name'))
+            function = item.get('function')
+            if isinstance(function, dict):
+                add(function.get('name'))
+
+        for key in ('tool_calls', 'tools', 'output', 'meta'):
+            if key in item:
+                walk(item.get(key))
+
+    walk(value)
+    return names
+
+
 ####################
 # ChatMessage DB Schema
 ####################
@@ -98,6 +148,7 @@ class ChatMessage(Base):
     files = Column(JSON, nullable=True)
     sources = Column(JSON, nullable=True)
     embeds = Column(JSON, nullable=True)
+    meta = Column(JSON, nullable=True)
 
     # Status
     done = Column(Boolean, default=True)
@@ -107,6 +158,9 @@ class ChatMessage(Base):
     # Usage (tokens, timing, etc.)
     usage = Column(JSON, nullable=True)
 
+    # Context compaction checkpoint
+    context_summary = Column(Text, nullable=True)
+
     # Timestamps
     created_at = Column(BigInteger, index=True)
     updated_at = Column(BigInteger)
@@ -137,10 +191,12 @@ class ChatMessageModel(BaseModel):
     files: Optional[list] = None
     sources: Optional[list] = None
     embeds: Optional[list] = None
+    meta: Optional[dict] = None
     done: bool = True
     status_history: Optional[list] = None
     error: Optional[dict | str] = None
     usage: Optional[dict] = None
+    context_summary: Optional[str] = None
     created_at: int
     updated_at: int
 
@@ -186,22 +242,23 @@ class ChatMessageTable:
                     existing.sources = data.get('sources')
                 if 'embeds' in data:
                     existing.embeds = data.get('embeds')
+                if 'meta' in data:
+                    existing.meta = data.get('meta')
                 if 'done' in data:
                     existing.done = data.get('done', True)
                 if 'status_history' in data or 'statusHistory' in data:
                     existing.status_history = data.get('status_history') or data.get('statusHistory')
                 if 'error' in data:
                     existing.error = data.get('error')
+                if 'context_summary' in data or 'contextSummary' in data:
+                    existing.context_summary = data.get('context_summary') or data.get('contextSummary')
                 # Extract and normalize usage
                 usage = get_usage(data)
                 if usage:
-                    # Deep-merge: preserve existing keys not present in new data
-                    # This prevents background tasks (follow-ups, title, tags)
-                    # from accidentally clearing the primary response's token counts
-                    existing.usage = {**(existing.usage or {}), **usage}
+                    existing_usage = normalize_usage(existing.usage or {}) if existing.usage else {}
+                    existing.usage = existing_usage if usage == existing_usage else merge_usage(existing_usage, usage)
                 existing.updated_at = now
                 await db.commit()
-                await db.refresh(existing)
                 return ChatMessageModel.model_validate(existing)
             else:
                 # Insert new
@@ -219,16 +276,17 @@ class ChatMessageTable:
                     files=data.get('files'),
                     sources=data.get('sources'),
                     embeds=data.get('embeds'),
+                    meta=data.get('meta'),
                     done=data.get('done', True),
                     status_history=data.get('status_history') or data.get('statusHistory'),
                     error=data.get('error'),
                     usage=usage,
+                    context_summary=data.get('context_summary') or data.get('contextSummary'),
                     created_at=timestamp,
                     updated_at=now,
                 )
                 db.add(message)
                 await db.commit()
-                await db.refresh(message)
                 return ChatMessageModel.model_validate(message)
 
     async def get_message_by_id(self, id: str, db: Optional[AsyncSession] = None) -> Optional[ChatMessageModel]:
@@ -236,6 +294,21 @@ class ChatMessageTable:
             message = await db.get(ChatMessage, id)
             return ChatMessageModel.model_validate(message) if message else None
 
+    async def has_unfinished_assistant_by_chat_id(
+        self,
+        chat_id: str,
+        db: Optional[AsyncSession] = None,
+    ) -> bool:
+        async with get_async_db_context(db) as db:
+            result = await db.execute(
+                select(ChatMessage.id)
+                .where(ChatMessage.chat_id == chat_id)
+                .where(ChatMessage.role == 'assistant')
+                .where(ChatMessage.done.is_(False))
+                .limit(1)
+            )
+            return result.scalar_one_or_none() is not None
+
     async def get_messages_by_chat_id(self, chat_id: str, db: Optional[AsyncSession] = None) -> list[ChatMessageModel]:
         async with get_async_db_context(db) as db:
             result = await db.execute(
@@ -249,6 +322,7 @@ class ChatMessageTable:
         'parent_id': 'parentId',
         'model_id': 'model',
         'status_history': 'statusHistory',
+        'context_summary': 'contextSummary',
         'created_at': 'timestamp',
     }
     # DB-internal columns excluded from the reconstructed message dict.
@@ -440,6 +514,44 @@ class ChatMessageTable:
             result = await db.execute(stmt)
             return {row.model_id: row.count for row in result.all()}
 
+    async def get_unique_counts_by_model(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        group_id: Optional[str] = None,
+        db: Optional[AsyncSession] = None,
+    ) -> dict[str, dict]:
+        """Count distinct users and chats per model."""
+        async with get_async_db_context(db) as db:
+            from open_webui.models.groups import GroupMember
+
+            stmt = select(
+                ChatMessage.model_id,
+                func.count(distinct(ChatMessage.user_id)).label('unique_users'),
+                func.count(distinct(ChatMessage.chat_id)).label('unique_chats'),
+            ).filter(
+                ChatMessage.role == 'assistant',
+                ChatMessage.model_id.isnot(None),
+            )
+
+            if start_date:
+                stmt = stmt.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                stmt = stmt.filter(ChatMessage.created_at <= end_date)
+            if group_id:
+                group_users = select(GroupMember.user_id).filter(GroupMember.group_id == group_id).scalar_subquery()
+                stmt = stmt.filter(ChatMessage.user_id.in_(group_users))
+
+            stmt = stmt.group_by(ChatMessage.model_id)
+            result = await db.execute(stmt)
+            return {
+                row.model_id: {
+                    'unique_users': row.unique_users,
+                    'unique_chats': row.unique_chats,
+                }
+                for row in result.all()
+            }
+
     async def get_token_usage_by_model(
         self,
         start_date: Optional[int] = None,
@@ -538,6 +650,233 @@ class ChatMessageTable:
                 for row in result.all()
             }
 
+    async def get_user_usage_summary(
+        self,
+        user_id: str,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        include_active_days: bool = True,
+        timezone: Optional[str] = None,
+        db: Optional[AsyncSession] = None,
+    ) -> dict:
+        async with get_async_db_context(db) as db:
+            bind = await db.connection()
+            dialect = bind.dialect.name
+            input_tokens, output_tokens = _token_columns(dialect)
+
+            messages_stmt = select(ChatMessage.role, func.count(ChatMessage.id).label('count')).filter(
+                ChatMessage.user_id == user_id,
+            )
+            token_stmt = select(
+                func.coalesce(func.sum(input_tokens), 0).label('input_tokens'),
+                func.coalesce(func.sum(output_tokens), 0).label('output_tokens'),
+            ).filter(
+                ChatMessage.user_id == user_id,
+                ChatMessage.role == 'assistant',
+                ChatMessage.usage.isnot(None),
+            )
+            models_stmt = select(func.count(distinct(ChatMessage.model_id)).label('models_used')).filter(
+                ChatMessage.user_id == user_id,
+                ChatMessage.role == 'assistant',
+                ChatMessage.model_id.isnot(None),
+            )
+            if start_date:
+                messages_stmt = messages_stmt.filter(ChatMessage.created_at >= start_date)
+                token_stmt = token_stmt.filter(ChatMessage.created_at >= start_date)
+                models_stmt = models_stmt.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                messages_stmt = messages_stmt.filter(ChatMessage.created_at <= end_date)
+                token_stmt = token_stmt.filter(ChatMessage.created_at <= end_date)
+                models_stmt = models_stmt.filter(ChatMessage.created_at <= end_date)
+
+            messages_result = await db.execute(messages_stmt.group_by(ChatMessage.role))
+            message_counts = {row.role: row.count for row in messages_result.all()}
+
+            token_result = (await db.execute(token_stmt)).one()
+            models_used = (await db.execute(models_stmt)).scalar() or 0
+
+            active_days = set()
+            if include_active_days:
+                tz = _timezone(timezone)
+                day_stmt = select(ChatMessage.created_at).filter(ChatMessage.user_id == user_id)
+                if start_date:
+                    day_stmt = day_stmt.filter(ChatMessage.created_at >= start_date)
+                if end_date:
+                    day_stmt = day_stmt.filter(ChatMessage.created_at <= end_date)
+                day_result = await db.execute(day_stmt)
+                active_days = {_date_key(row.created_at, tz) for row in day_result.all()}
+
+            input_total = int(token_result.input_tokens or 0)
+            output_total = int(token_result.output_tokens or 0)
+
+            return {
+                'messages': sum(message_counts.values()),
+                'user_messages': message_counts.get('user', 0),
+                'assistant_messages': message_counts.get('assistant', 0),
+                'input_tokens': input_total,
+                'output_tokens': output_total,
+                'total_tokens': input_total + output_total,
+                'models_used': int(models_used),
+                'active_days': len(active_days),
+            }
+
+    async def get_user_first_message_created_at(
+        self,
+        user_id: str,
+        db: Optional[AsyncSession] = None,
+    ) -> Optional[int]:
+        async with get_async_db_context(db) as db:
+            result = await db.execute(
+                select(func.min(ChatMessage.created_at)).filter(
+                    ChatMessage.user_id == user_id,
+                    ChatMessage.created_at.isnot(None),
+                )
+            )
+            value = result.scalar()
+            return int(value) if value else None
+
+    async def get_user_daily_usage(
+        self,
+        user_id: str,
+        start_date: int,
+        end_date: int,
+        timezone: Optional[str] = None,
+        db: Optional[AsyncSession] = None,
+    ) -> list[dict]:
+        async with get_async_db_context(db) as db:
+            tz = _timezone(timezone)
+            bind = await db.connection()
+            dialect = bind.dialect.name
+            input_tokens, output_tokens = _token_columns(dialect)
+
+            stmt = select(
+                ChatMessage.created_at,
+                ChatMessage.chat_id,
+                ChatMessage.role,
+                ChatMessage.model_id,
+                ChatMessage.usage,
+                input_tokens.label('input_tokens'),
+                output_tokens.label('output_tokens'),
+            ).filter(
+                ChatMessage.user_id == user_id,
+                ChatMessage.created_at >= start_date,
+                ChatMessage.created_at <= end_date,
+            )
+
+            result = await db.execute(stmt)
+            daily: dict[str, dict] = {}
+            for row in result.all():
+                date = _date_key(row.created_at, tz)
+                entry = daily.setdefault(
+                    date,
+                    {
+                        'date': date,
+                        'messages': 0,
+                        'chat_ids': set(),
+                        'tokens': 0,
+                        'models': Counter(),
+                    },
+                )
+                entry['messages'] += 1
+                entry['chat_ids'].add(row.chat_id)
+                if row.role == 'assistant' and row.model_id:
+                    entry['models'][row.model_id] += 1
+                if row.usage:
+                    entry['tokens'] += int(row.input_tokens or 0) + int(row.output_tokens or 0)
+
+            current = datetime.fromtimestamp(_normalize_timestamp(start_date), tz=tz).replace(
+                hour=0, minute=0, second=0, microsecond=0
+            )
+            end_dt = datetime.fromtimestamp(_normalize_timestamp(end_date), tz=tz).replace(
+                hour=0, minute=0, second=0, microsecond=0
+            )
+            while current <= end_dt:
+                date = current.strftime('%Y-%m-%d')
+                daily.setdefault(
+                    date,
+                    {'date': date, 'messages': 0, 'chat_ids': set(), 'tokens': 0, 'models': Counter()},
+                )
+                current += timedelta(days=1)
+
+            return [
+                {
+                    'date': item['date'],
+                    'messages': item['messages'],
+                    'chats': len(item['chat_ids']),
+                    'tokens': item['tokens'],
+                    'models': dict(item['models']),
+                }
+                for item in sorted(daily.values(), key=lambda x: x['date'])
+            ]
+
+    async def get_user_top_models(
+        self,
+        user_id: str,
+        start_date: int,
+        end_date: int,
+        limit: int = 5,
+        db: Optional[AsyncSession] = None,
+    ) -> list[dict]:
+        async with get_async_db_context(db) as db:
+            bind = await db.connection()
+            dialect = bind.dialect.name
+            input_tokens, output_tokens = _token_columns(dialect)
+
+            stmt = (
+                select(
+                    ChatMessage.model_id,
+                    func.count(ChatMessage.id).label('messages'),
+                    func.coalesce(func.sum(input_tokens), 0).label('input_tokens'),
+                    func.coalesce(func.sum(output_tokens), 0).label('output_tokens'),
+                )
+                .filter(
+                    ChatMessage.user_id == user_id,
+                    ChatMessage.role == 'assistant',
+                    ChatMessage.model_id.isnot(None),
+                    ChatMessage.created_at >= start_date,
+                    ChatMessage.created_at <= end_date,
+                )
+                .group_by(ChatMessage.model_id)
+                .order_by(func.count(ChatMessage.id).desc())
+                .limit(limit)
+            )
+            result = await db.execute(stmt)
+            return [
+                {
+                    'model_id': row.model_id,
+                    'messages': row.messages,
+                    'input_tokens': int(row.input_tokens or 0),
+                    'output_tokens': int(row.output_tokens or 0),
+                    'total_tokens': int(row.input_tokens or 0) + int(row.output_tokens or 0),
+                }
+                for row in result.all()
+            ]
+
+    async def get_user_top_tools(
+        self,
+        user_id: str,
+        start_date: int,
+        end_date: int,
+        limit: int = 5,
+        db: Optional[AsyncSession] = None,
+    ) -> list[dict]:
+        async with get_async_db_context(db) as db:
+            stmt = select(ChatMessage.output, ChatMessage.meta).filter(
+                ChatMessage.user_id == user_id,
+                ChatMessage.created_at >= start_date,
+                ChatMessage.created_at <= end_date,
+            )
+            result = await db.execute(stmt)
+
+            counts: Counter[str] = Counter()
+            for output, meta in result.all():
+                for name in _extract_tool_names(output):
+                    counts[name] += 1
+                for name in _extract_tool_names(meta):
+                    counts[name] += 1
+
+            return [{'name': name, 'count': count} for name, count in counts.most_common(limit)]
+
     async def get_message_count_by_user(
         self,
         start_date: Optional[int] = None,

backend/open_webui/models/config.py

@@ -0,0 +1,366 @@
+"""Database-backed configuration with per-key storage.
+
+Replaces the old single-row JSON blob machinery with a simple per-key model
+mirroring cptr's Config.
+
+Each config key is stored as its own row: key TEXT PK, value JSON.
+Reads are direct DB lookups. Writes are explicit awaited upserts that raise on
+failure (no more fire-and-forget create_task).
+"""
+
+from __future__ import annotations
+
+import logging
+import time
+from typing import Any, ClassVar
+
+from fastapi.encoders import jsonable_encoder
+from open_webui.internal.db import Base, get_async_db
+from sqlalchemy import JSON, BigInteger, Column, Text, delete, select
+
+log = logging.getLogger(__name__)
+
+API_CONFIG_KEYS = ('openai.api_configs', 'ollama.api_configs')
+DICT_CONFIG_KEY_ALIASES = {
+    'openai.api_configs': ('OPENAI_API_CONFIGS',),
+    'ollama.api_configs': ('OLLAMA_API_CONFIGS',),
+    'rag.mineru_params': ('MINERU_PARAMS',),
+    'rag.docling_params': ('DOCLING_PARAMS',),
+    'web.search.linkup_search_params': ('LINKUP_SEARCH_PARAMS',),
+    'image_generation.automatic1111.api_params': ('AUTOMATIC1111_PARAMS',),
+    'image_generation.openai.params': ('IMAGES_OPENAI_API_PARAMS',),
+    'audio.tts.openai.params': ('AUDIO_TTS_OPENAI_PARAMS',),
+    'models.default_metadata': ('DEFAULT_MODEL_METADATA',),
+    'models.default_params': ('DEFAULT_MODEL_PARAMS',),
+    'user.permissions': ('USER_PERMISSIONS',),
+}
+DICT_CONFIG_KEYS = tuple(DICT_CONFIG_KEY_ALIASES)
+API_CONFIG_FIELDS = (
+    'enable',
+    'key',
+    'prefix_id',
+    'tags',
+    'model_ids',
+    'connection_type',
+    'provider',
+    'auth_type',
+    'headers',
+    'azure',
+    'api_type',
+    'api_version',
+    'extra_params',
+    'passthrough_params',
+)
+
+
+def _split_api_config_fragment(fragment: str) -> tuple[str, list[str]] | None:
+    if not fragment:
+        return None
+
+    first, _, rest = fragment.partition('.')
+    if first.isdigit() and rest:
+        return first, rest.split('.')
+
+    match: tuple[int, str] | None = None
+    for field in API_CONFIG_FIELDS:
+        marker = f'.{field}'
+        marker_index = fragment.rfind(marker)
+        if marker_index != -1 and (match is None or marker_index > match[0]):
+            match = (marker_index, field)
+
+    if match:
+        marker_index, field = match
+        connection_key = fragment[:marker_index]
+        field_path = fragment[marker_index + 1 :]
+        if connection_key:
+            return connection_key, field_path.split('.')
+
+    return None
+
+
+def _assign_path(target: dict, path: list[str], value: Any) -> None:
+    current = target
+    for part in path[:-1]:
+        next_value = current.get(part)
+        if not isinstance(next_value, dict):
+            next_value = {}
+            current[part] = next_value
+        current = next_value
+    current[path[-1]] = value
+
+
+def _json_value(value: Any) -> Any:
+    return jsonable_encoder(value)
+
+
+# ── Model ────────────────────────────────────────────────────────────────────
+
+
+class Config(Base):
+    """Per-key config storage. Each row is one config key."""
+
+    __tablename__ = 'config'
+
+    key = Column(Text, primary_key=True)
+    value = Column(JSON, nullable=False)
+    updated_at = Column(BigInteger, nullable=True)
+
+    DEFAULTS: ClassVar[dict[str, Any]] = {}
+    PERSISTENT_ENABLED: ClassVar[bool] = True
+    OAUTH_PERSISTENT_ENABLED: ClassVar[bool] = False
+
+    # ── Class methods ────────────────────────────────────────
+
+    @classmethod
+    def configure(
+        cls,
+        *,
+        defaults: dict[str, Any] | None = None,
+        enable_persistent: bool = True,
+        enable_oauth_persistent: bool = False,
+    ) -> None:
+        cls.DEFAULTS = dict(defaults or {})
+        cls.PERSISTENT_ENABLED = enable_persistent
+        cls.OAUTH_PERSISTENT_ENABLED = enable_oauth_persistent
+
+    @classmethod
+    def default_value(cls, key: str, default: Any = None) -> Any:
+        return cls.DEFAULTS.get(key, default)
+
+    @classmethod
+    def persistent_enabled_for(cls, key: str) -> bool:
+        if not cls.PERSISTENT_ENABLED:
+            return False
+        if key.startswith('oauth.') and not cls.OAUTH_PERSISTENT_ENABLED:
+            return False
+        return True
+
+    @staticmethod
+    async def get(key: str, default: Any = None) -> Any:
+        """Get a config value by key. Returns default if not set."""
+        if not Config.persistent_enabled_for(key):
+            return Config.default_value(key, default)
+        async with get_async_db() as db:
+            row = await db.get(Config, key)
+            return row.value if row else Config.default_value(key, default)
+
+    @staticmethod
+    async def get_many(*keys: str) -> dict:
+        """Get multiple config values. Returns {key: value} for keys that exist."""
+        disabled_values = {
+            key: Config.default_value(key)
+            for key in keys
+            if not Config.persistent_enabled_for(key) and key in Config.DEFAULTS
+        }
+        enabled_keys = {key for key in keys if Config.persistent_enabled_for(key)}
+        if not enabled_keys:
+            return disabled_values
+        async with get_async_db() as db:
+            result = await db.execute(select(Config).where(Config.key.in_(enabled_keys)))
+            values = {row.key: row.value for row in result.scalars().all()}
+            return {
+                key: values.get(key, Config.default_value(key))
+                for key in keys
+                if key in values or key in Config.DEFAULTS or key in disabled_values
+            }
+
+    @staticmethod
+    async def get_namespace(namespace: str) -> dict:
+        """Get all config keys under a dotted namespace."""
+        default_values = {
+            key: value
+            for key, value in Config.DEFAULTS.items()
+            if key.startswith(f'{namespace}.') and not Config.persistent_enabled_for(key)
+        }
+        if not Config.PERSISTENT_ENABLED:
+            return default_values
+        async with get_async_db() as db:
+            result = await db.execute(select(Config).where(Config.key.like(f'{namespace}.%')))
+            values = {row.key: row.value for row in result.scalars().all()}
+            values.update(default_values)
+            return values
+
+    @staticmethod
+    async def get_all() -> dict:
+        """Get all config as {key: value}."""
+        if not Config.PERSISTENT_ENABLED:
+            return dict(Config.DEFAULTS)
+        async with get_async_db() as db:
+            result = await db.execute(select(Config))
+            values = {row.key: row.value for row in result.scalars().all()}
+            if not Config.OAUTH_PERSISTENT_ENABLED:
+                values.update({key: value for key, value in Config.DEFAULTS.items() if key.startswith('oauth.')})
+            return values
+
+    @staticmethod
+    async def upsert(updates: dict) -> None:
+        """Upsert multiple config key-value pairs. Raises on failure."""
+        persistent_updates = {}
+        for key, value in updates.items():
+            value = _json_value(value)
+            if Config.persistent_enabled_for(key):
+                persistent_updates[key] = value
+            else:
+                Config.DEFAULTS[key] = value
+
+        if not persistent_updates:
+            return
+
+        async with get_async_db() as db:
+            now = int(time.time())
+            for key, value in persistent_updates.items():
+                existing = await db.get(Config, key)
+                if existing:
+                    existing.value = value
+                    existing.updated_at = now
+                else:
+                    db.add(Config(key=key, value=value, updated_at=now))
+            await db.commit()
+
+    @staticmethod
+    async def delete(key: str) -> bool:
+        """Delete a config key. Returns True if it existed."""
+        async with get_async_db() as db:
+            row = await db.get(Config, key)
+            if row:
+                await db.delete(row)
+                await db.commit()
+                return True
+            return False
+
+    @staticmethod
+    async def clear() -> None:
+        """Delete all config rows."""
+        async with get_async_db() as db:
+            await db.execute(delete(Config))
+            await db.commit()
+
+    @staticmethod
+    async def seed_defaults(defaults: dict) -> None:
+        """Insert keys that don't yet exist in the DB.
+
+        Called at startup to ensure all known config keys have values.
+        Existing DB values take precedence over defaults.
+        """
+        async with get_async_db() as db:
+            result = await db.execute(select(Config.key))
+            existing_keys = {row[0] for row in result.all()}
+
+            now = int(time.time())
+            new_count = 0
+            for key, value in defaults.items():
+                # Skip keys the DB is not authoritative for (e.g. oauth.* while
+                # ENABLE_OAUTH_PERSISTENT_CONFIG is off), matching the read paths.
+                if not Config.persistent_enabled_for(key):
+                    continue
+                if key not in existing_keys:
+                    value = _json_value(value)
+                    db.add(Config(key=key, value=value, updated_at=now))
+                    existing_keys.add(key)
+                    new_count += 1
+
+            if new_count:
+                await db.commit()
+                log.info('Seeded %d new config defaults', new_count)
+
+    @staticmethod
+    async def rename_prefix(old_prefix: str, new_prefix: str) -> None:
+        """Move persisted config keys from one dotted prefix to another."""
+        if not Config.PERSISTENT_ENABLED:
+            return
+
+        async with get_async_db() as db:
+            result = await db.execute(select(Config).where(Config.key.like(f'{old_prefix}.%')))
+            rows = result.scalars().all()
+            if not rows:
+                return
+
+            now = int(time.time())
+            moved_count = 0
+            deleted_count = 0
+            for row in rows:
+                new_key = f'{new_prefix}.{row.key.removeprefix(f"{old_prefix}.")}'
+                existing = await db.get(Config, new_key)
+                if existing is None:
+                    db.add(Config(key=new_key, value=row.value, updated_at=now))
+                    moved_count += 1
+                else:
+                    deleted_count += 1
+                await db.delete(row)
+
+            await db.commit()
+            log.info(
+                'Renamed %d config keys from %s.* to %s.*; deleted %d old duplicates',
+                moved_count,
+                old_prefix,
+                new_prefix,
+                deleted_count,
+            )
+
+    @staticmethod
+    async def repair_flattened_dict_configs() -> None:
+        """Reassemble dict config values flattened by the per-key migration."""
+        if not Config.PERSISTENT_ENABLED:
+            return
+
+        async with get_async_db() as db:
+            repaired_keys: list[str] = []
+            orphan_keys: list[str] = []
+
+            for config_key, aliases in DICT_CONFIG_KEY_ALIASES.items():
+                prefixes = (config_key, *aliases)
+                rows = []
+                for key_prefix in prefixes:
+                    result = await db.execute(select(Config).where(Config.key.like(f'{key_prefix}.%')))
+                    rows.extend(result.scalars().all())
+                if not rows:
+                    continue
+
+                existing = await db.get(Config, config_key)
+                repaired = existing.value if existing and isinstance(existing.value, dict) else {}
+
+                repaired_any = False
+                for row in rows:
+                    fragment = None
+                    for key_prefix in prefixes:
+                        prefix = f'{key_prefix}.'
+                        if row.key.startswith(prefix):
+                            fragment = row.key.removeprefix(prefix)
+                            break
+                    if fragment is None:
+                        continue
+
+                    if config_key in API_CONFIG_KEYS:
+                        split = _split_api_config_fragment(fragment)
+                        if not split:
+                            continue
+                        object_key, field_path = split
+                    else:
+                        object_key, field_path = None, fragment.split('.')
+
+                    target = repaired
+                    if object_key is not None:
+                        target = repaired.setdefault(object_key, {})
+                        if not isinstance(target, dict):
+                            continue
+
+                    _assign_path(target, field_path, row.value)
+                    orphan_keys.append(row.key)
+                    repaired_any = True
+
+                if not repaired_any:
+                    continue
+
+                if existing:
+                    existing.value = repaired
+                    existing.updated_at = int(time.time())
+                else:
+                    db.add(Config(key=config_key, value=repaired, updated_at=int(time.time())))
+                repaired_keys.append(config_key)
+
+            if orphan_keys:
+                await db.execute(delete(Config).where(Config.key.in_(orphan_keys)))
+
+            if repaired_keys or orphan_keys:
+                await db.commit()
+                log.info('Repaired flattened dict config rows for %s', ', '.join(repaired_keys))

backend/open_webui/models/feedbacks.py

@@ -133,6 +133,12 @@ class ModelHistoryEntry(BaseModel):
     lost: int
 
 
+class ModelHistoryCounts(BaseModel):
+    date: str
+    won: int = 0
+    lost: int = 0
+
+
 class ModelHistoryResponse(BaseModel):
     model_id: str
     history: list[ModelHistoryEntry]
@@ -159,7 +165,6 @@ class FeedbackTable:
                 result = Feedback(**feedback.model_dump())
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 if result:
                     return FeedbackModel.model_validate(result)
                 else:
@@ -216,12 +221,15 @@ class FeedbackTable:
     ) -> FeedbackListResponse:
         async with get_async_db_context(db) as db:
             stmt = select(Feedback, User).join(User, Feedback.user_id == User.id)
+            count_stmt = select(func.count(Feedback.id)).select_from(Feedback).join(User, Feedback.user_id == User.id)
 
             if filter:
                 # Apply model_id filter (exact match)
                 model_id = filter.get('model_id')
                 if model_id:
-                    stmt = stmt.filter(Feedback.data['model_id'].as_string() == model_id)
+                    model_id_filter = Feedback.data['model_id'].as_string() == model_id
+                    stmt = stmt.filter(model_id_filter)
+                    count_stmt = count_stmt.filter(model_id_filter)
 
                 order_by = filter.get('order_by')
                 direction = filter.get('direction')
@@ -250,9 +258,9 @@ class FeedbackTable:
             else:
                 stmt = stmt.order_by(Feedback.created_at.desc())
 
-            # Count BEFORE pagination
-            count_result = await db.execute(select(func.count()).select_from(stmt.subquery()))
-            total = count_result.scalar()
+            # Count before pagination without wrapping the ordered item query.
+            count_result = await db.execute(count_stmt)
+            total = count_result.scalar() or 0
 
             if skip:
                 stmt = stmt.offset(skip)
@@ -375,6 +383,45 @@ class FeedbackTable:
 
         return result
 
+    async def get_model_feedback_counts_by_day(
+        self,
+        model_id: str,
+        start_date: Optional[int] = None,
+        db: Optional[AsyncSession] = None,
+    ) -> list[ModelHistoryCounts]:
+        """Get aggregated feedback counts per day for a model, preserving all matching days."""
+        from collections import defaultdict
+        from datetime import datetime
+
+        async with get_async_db_context(db) as db:
+            stmt = select(Feedback.created_at, Feedback.data).filter(Feedback.data['model_id'].as_string() == model_id)
+            if start_date is not None:
+                stmt = stmt.filter(Feedback.created_at >= start_date)
+
+            result = await db.execute(stmt.order_by(Feedback.created_at.asc()))
+            rows = result.all()
+
+        daily_counts = defaultdict(lambda: {'won': 0, 'lost': 0})
+
+        for created_at, data in rows:
+            if not data:
+                continue
+
+            rating_str = str(data.get('rating', ''))
+            if rating_str not in ('1', '-1'):
+                continue
+
+            date_str = datetime.fromtimestamp(created_at).strftime('%Y-%m-%d')
+            if rating_str == '1':
+                daily_counts[date_str]['won'] += 1
+            else:
+                daily_counts[date_str]['lost'] += 1
+
+        return [
+            ModelHistoryCounts(date=date_str, won=counts['won'], lost=counts['lost'])
+            for date_str, counts in sorted(daily_counts.items())
+        ]
+
     async def get_feedbacks_by_type(self, type: str, db: Optional[AsyncSession] = None) -> list[FeedbackModel]:
         async with get_async_db_context(db) as db:
             result = await db.execute(select(Feedback).filter_by(type=type).order_by(Feedback.updated_at.desc()))

backend/open_webui/models/files.py

@@ -142,7 +142,6 @@ class FilesTable:
                 result = File(**file.model_dump())
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 if result:
                     return FileModel.model_validate(result)
                 else:
@@ -201,6 +200,18 @@ class FilesTable:
             result = await db.execute(select(File))
             return [FileModel.model_validate(file) for file in result.scalars().all()]
 
+    async def count_files_by_user_id(
+        self,
+        user_id: str | None = None,
+        db: AsyncSession | None = None,
+    ) -> int:
+        async with get_async_db_context(db) as db:
+            stmt = select(func.count(File.id))
+            if user_id:
+                stmt = stmt.filter_by(user_id=user_id)
+            result = await db.execute(stmt)
+            return result.scalar() or 0
+
     async def check_access_by_user_id(self, id, user_id, permission='write', db: AsyncSession | None = None) -> bool:
         file = await self.get_file_by_id(id, db=db)
         if not file:

backend/open_webui/models/folders.py

@@ -6,7 +6,7 @@ from typing import Optional
 
 from open_webui.internal.db import Base, JSONField, get_async_db_context
 from pydantic import BaseModel, ConfigDict
-from sqlalchemy import JSON, BigInteger, Boolean, Column, Text, delete, func, select
+from sqlalchemy import JSON, BigInteger, Boolean, Column, Text, delete, func, select, or_, and_
 from sqlalchemy.ext.asyncio import AsyncSession
 
 log = logging.getLogger(__name__)
@@ -58,6 +58,21 @@ class FolderNameIdResponse(BaseModel):
     meta: Optional[FolderMetadataResponse] = None
     parent_id: Optional[str] = None
     is_expanded: bool = False
+    unread_count: int = 0
+    created_at: int
+    updated_at: int
+
+
+class SharedFolderResponse(BaseModel):
+    id: str
+    name: str
+    parent_id: Optional[str] = None
+    user_id: str
+    owner_name: Optional[str] = None
+    permission: str = 'read'
+    access_grants: list = []
+    is_expanded: bool = False
+    meta: Optional[dict] = None
     created_at: int
     updated_at: int
 
@@ -130,6 +145,52 @@ class FolderTable:
         except Exception:
             return None
 
+    async def get_folder_by_id(self, id: str, db: Optional[AsyncSession] = None) -> Optional[FolderModel]:
+        """Fetch folder by ID only (no user_id filter). Used for shared access."""
+        try:
+            async with get_async_db_context(db) as db:
+                result = await db.execute(select(Folder).filter_by(id=id))
+                folder = result.scalars().first()
+                if not folder:
+                    return None
+                return FolderModel.model_validate(folder)
+        except Exception:
+            return None
+
+    async def get_shared_folder_ids_for_user(
+        self, user_id: str, user_group_ids: set[str], db: Optional[AsyncSession] = None
+    ) -> dict[str, str]:
+        """
+        Returns {folder_id: highest_permission} for all folders shared with user.
+        Checks direct user grants, group grants, and public (user:*) grants.
+        """
+        from open_webui.models.access_grants import AccessGrant
+
+        async with get_async_db_context(db) as db:
+            conditions = [
+                and_(AccessGrant.principal_type == 'user', AccessGrant.principal_id == '*'),
+                and_(AccessGrant.principal_type == 'user', AccessGrant.principal_id == user_id),
+            ]
+            if user_group_ids:
+                conditions.append(
+                    and_(AccessGrant.principal_type == 'group', AccessGrant.principal_id.in_(user_group_ids))
+                )
+            result = await db.execute(
+                select(AccessGrant).filter(
+                    AccessGrant.resource_type == 'folder',
+                    or_(*conditions),
+                )
+            )
+            grants = result.scalars().all()
+
+            # Build {folder_id: highest_permission} ('write' > 'read')
+            folder_perms = {}
+            for g in grants:
+                existing = folder_perms.get(g.resource_id)
+                if existing != 'write':
+                    folder_perms[g.resource_id] = g.permission
+            return folder_perms
+
     async def get_children_folders_by_id_and_user_id(
         self, id: str, user_id: str, db: Optional[AsyncSession] = None
     ) -> Optional[list[FolderModel]]:
@@ -188,6 +249,25 @@ class FolderTable:
             result = await db.execute(select(Folder).filter_by(parent_id=parent_id, user_id=user_id))
             return [FolderModel.model_validate(folder) for folder in result.scalars().all()]
 
+    async def get_folder_ids_by_id_and_user_id_in_subtree(
+        self, id: str, user_id: str, db: Optional[AsyncSession] = None
+    ) -> list[str]:
+        async with get_async_db_context(db) as db:
+            result = await db.execute(select(Folder).filter_by(id=id, user_id=user_id))
+            folder = result.scalars().first()
+            if not folder:
+                return []
+
+            folder_ids = [folder.id]
+            folders = [FolderModel.model_validate(folder)]
+            while folders:
+                current_folder = folders.pop()
+                children = await self.get_folders_by_parent_id_and_user_id(current_folder.id, user_id, db=db)
+                folder_ids.extend(child.id for child in children)
+                folders.extend(children)
+
+            return folder_ids
+
     async def update_folder_parent_id_by_id_and_user_id(
         self,
         id: str,

backend/open_webui/models/functions.py

@@ -7,7 +7,8 @@ import time
 
 # local imports
 from open_webui.internal.db import Base, JSONField, get_async_db_context
-from open_webui.models.users import UserModel, UserResponse, Users
+from open_webui.models.users import User, UserResponse, Users, UserSettings
+from open_webui.utils.valves import decrypt_valves, encrypt_valves
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Boolean, Column, Index, String, Text, delete, select, update
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -41,7 +42,7 @@ class FunctionMeta(BaseModel):
 
 class FunctionModel(BaseModel):
     id: str
-    user_id: str
+    user_id: str | None = None  # may be null for legacy/malformed records
     name: str
     type: str
     content: str
@@ -57,7 +58,7 @@ class FunctionModel(BaseModel):
 # --- form / schema definitions ---
 class FunctionWithValvesModel(BaseModel):
     id: str
-    user_id: str
+    user_id: str | None = None  # may be null for legacy/malformed records
     name: str
     type: str
     content: str
@@ -78,7 +79,7 @@ class FunctionWithValvesModel(BaseModel):
 
 class FunctionResponse(BaseModel):
     id: str
-    user_id: str
+    user_id: str | None = None  # may be null for legacy/malformed records
     type: str
     name: str
     meta: FunctionMeta
@@ -128,7 +129,6 @@ class FunctionsTable:
                 result = Function(**function.model_dump())
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 if result:
                     return FunctionModel.model_validate(result)
                 else:
@@ -143,7 +143,8 @@ class FunctionsTable:
         functions: list[FunctionWithValvesModel],
         db: AsyncSession | None = None,
     ) -> list[FunctionWithValvesModel]:
-        # Synchronize functions for a user by updating existing ones, inserting new ones, and removing those that are no longer present.
+        # Synchronize functions by updating existing ones, inserting new ones,
+        # and removing those that are no longer present.
         try:
             async with get_async_db_context(db) as db:
                 # Get existing functions
@@ -156,24 +157,15 @@ class FunctionsTable:
 
                 # Update or insert functions
                 for func in functions:
+                    func_data = func.model_dump()
+                    func_data['valves'] = encrypt_valves(func_data['valves']) if func_data.get('valves') else None
+                    func_data['user_id'] = user_id
+                    func_data['updated_at'] = int(time.time())
+
                     if func.id in existing_ids:
-                        await db.execute(
-                            update(Function)
-                            .filter_by(id=func.id)
-                            .values(
-                                **func.model_dump(),
-                                user_id=user_id,
-                                updated_at=int(time.time()),
-                            )
-                        )
+                        await db.execute(update(Function).filter_by(id=func.id).values(**func_data))
                     else:
-                        new_func = Function(
-                            **{
-                                **func.model_dump(),
-                                'user_id': user_id,
-                                'updated_at': int(time.time()),
-                            }
-                        )
+                        new_func = Function(**func_data)
                         db.add(new_func)
 
                 # Remove functions that are no longer present
@@ -227,7 +219,15 @@ class FunctionsTable:
             functions = result.scalars().all()
 
             if include_valves:
-                return [FunctionWithValvesModel.model_validate(function) for function in functions]
+                return [
+                    FunctionWithValvesModel.model_validate(
+                        {
+                            **FunctionModel.model_validate(function).model_dump(),
+                            'valves': decrypt_valves(function.valves),
+                        }
+                    )
+                    for function in functions
+                ]
             else:
                 return [FunctionModel.model_validate(function) for function in functions]
 
@@ -274,6 +274,18 @@ class FunctionsTable:
             result = await db.execute(select(Function).filter_by(type='filter', is_active=True, is_global=True))
             return [FunctionModel.model_validate(function) for function in result.scalars().all()]
 
+    async def get_active_function_ids_by_type(
+        self, type: str, db: AsyncSession | None = None
+    ) -> list[tuple[str, bool]]:
+        """Return (id, is_global) for active functions without fetching plugin source."""
+        async with get_async_db_context(db) as db:
+            result = await db.execute(select(Function.id, Function.is_global).filter_by(type=type, is_active=True))
+            return [(id, bool(is_global)) for id, is_global in result.all()]
+
+    async def get_active_filter_ids(self, db: AsyncSession | None = None) -> list[tuple[str, bool]]:
+        """Return (id, is_global) for active filters without fetching plugin source."""
+        return await self.get_active_function_ids_by_type('filter', db=db)
+
     async def get_global_action_functions(self, db: AsyncSession | None = None) -> list[FunctionModel]:
         async with get_async_db_context(db) as db:
             result = await db.execute(select(Function).filter_by(type='action', is_active=True, is_global=True))
@@ -282,8 +294,8 @@ class FunctionsTable:
     async def get_function_valves_by_id(self, id: str, db: AsyncSession | None = None) -> dict | None:
         async with get_async_db_context(db) as db:
             try:
-                function = await db.get(Function, id)
-                return function.valves if function.valves else {}
+                result = await db.execute(select(Function.valves).filter_by(id=id))
+                return decrypt_valves(result.scalar_one_or_none())
             except Exception as e:
                 log.exception(f'Error getting function valves by id {id}: {e}')
                 return None
@@ -299,8 +311,7 @@ class FunctionsTable:
         try:
             async with get_async_db_context(db) as db:
                 result = await db.execute(select(Function.id, Function.valves).filter(Function.id.in_(ids)))
-                functions = result.all()
-                return {f.id: (f.valves if f.valves else {}) for f in functions}
+                return {id: decrypt_valves(valves) for id, valves in result.all()}
         except Exception as e:
             log.exception(f'Error batch-fetching function valves: {e}')
             return {}
@@ -311,10 +322,9 @@ class FunctionsTable:
         async with get_async_db_context(db) as db:
             try:
                 function = await db.get(Function, id)
-                function.valves = valves
+                function.valves = encrypt_valves(valves)
                 function.updated_at = int(time.time())
                 await db.commit()
-                await db.refresh(function)
                 return FunctionModel.model_validate(function)
             except Exception:
                 return None
@@ -334,7 +344,6 @@ class FunctionsTable:
 
                     function.updated_at = int(time.time())
                     await db.commit()
-                    await db.refresh(function)
                     return FunctionModel.model_validate(function)
                 else:
                     return None
@@ -346,8 +355,11 @@ class FunctionsTable:
         self, id: str, user_id: str, db: AsyncSession | None = None
     ) -> dict | None:
         try:
-            user = await Users.get_user_by_id(user_id, db=db)
-            user_settings = user.settings.model_dump() if user.settings else {}
+            async with get_async_db_context(db) as db:
+                result = await db.execute(select(User.settings).filter_by(id=user_id))
+                settings = result.scalar_one_or_none()
+
+            user_settings = UserSettings(**settings).model_dump() if settings else {}
 
             # Check if user has "functions" and "valves" settings
             if 'functions' not in user_settings:
@@ -355,8 +367,8 @@ class FunctionsTable:
             if 'valves' not in user_settings['functions']:
                 user_settings['functions']['valves'] = {}
 
-            return user_settings['functions']['valves'].get(id, {})
-        except Exception as e:
+            return decrypt_valves(user_settings['functions']['valves'].get(id))
+        except Exception:
             log.exception(f'Error getting user values by id {id} and user id {user_id}')
             return None
 
@@ -373,12 +385,12 @@ class FunctionsTable:
             if 'valves' not in user_settings['functions']:
                 user_settings['functions']['valves'] = {}
 
-            user_settings['functions']['valves'][id] = valves
+            user_settings['functions']['valves'][id] = encrypt_valves(valves)
 
             # Update the user settings in the database
             await Users.update_user_by_id(user_id, {'settings': user_settings}, db=db)
 
-            return user_settings['functions']['valves'][id]
+            return valves
         except Exception as e:
             log.exception(f'Error updating user valves by id {id} and user_id {user_id}: {e}')
             return None

backend/open_webui/models/knowledge.py

@@ -4,6 +4,7 @@ import time
 import uuid
 from typing import Optional
 
+from open_webui.config import RAG_FILE_CONTENT_SEARCH_MAX_CHARS
 from open_webui.internal.db import Base, JSONField, get_async_db_context
 from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 from open_webui.models.files import (
@@ -31,9 +32,13 @@ from sqlalchemy import (
     update,
 )
 from sqlalchemy.ext.asyncio import AsyncSession
+from sqlalchemy.orm import defer
 
 log = logging.getLogger(__name__)
 
+# Columns the knowledge base list may be ordered by; anything else falls back to the default.
+KNOWLEDGE_SORTABLE_FIELDS = {'name', 'created_at', 'updated_at'}
+
 ####################
 # Knowledge DB Schema
 # Let what was gathered here outlast the one who gathered it,
@@ -147,6 +152,7 @@ class KnowledgeDirectoryForm(BaseModel):
 ####################
 class KnowledgeUserModel(KnowledgeModel):
     user: Optional[UserResponse] = None
+    file_count: int | None = None
 
 
 class KnowledgeResponse(KnowledgeModel):
@@ -189,11 +195,11 @@ class KnowledgeTable:
         access_grants: Optional[list[AccessGrantModel]] = None,
         db: Optional[AsyncSession] = None,
     ) -> KnowledgeModel:
-        knowledge_data = KnowledgeModel.model_validate(knowledge).model_dump(exclude={'access_grants'})
-        knowledge_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(knowledge_data['id'], db=db)
+        knowledge_model = KnowledgeModel.model_validate(knowledge)
+        knowledge_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(knowledge_model.id, db=db)
         )
-        return KnowledgeModel.model_validate(knowledge_data)
+        return knowledge_model
 
     async def insert_new_knowledge(
         self, user_id: str, form_data: KnowledgeForm, db: Optional[AsyncSession] = None
@@ -286,6 +292,17 @@ class KnowledgeTable:
                     elif view_option == 'shared':
                         stmt = stmt.filter(Knowledge.user_id != user_id)
 
+                    source = filter.get('source')
+                    if source == 'external':
+                        stmt = stmt.filter(Knowledge.meta['source'].as_string() == 'external')
+                    elif source == 'local':
+                        stmt = stmt.filter(
+                            or_(
+                                Knowledge.meta.is_(None),
+                                Knowledge.meta['source'].as_string() != 'external',
+                            )
+                        )
+
                     stmt = AccessGrants.has_permission_filter(
                         db=db,
                         query=stmt,
@@ -295,7 +312,17 @@ class KnowledgeTable:
                         permission='read',
                     )
 
-                stmt = stmt.order_by(Knowledge.updated_at.desc(), Knowledge.id.asc())
+                order_by = (filter or {}).get('order_by')
+                direction = (filter or {}).get('direction')
+
+                if order_by in KNOWLEDGE_SORTABLE_FIELDS:
+                    column = getattr(Knowledge, order_by)
+                    if (direction or 'desc').lower() == 'asc':
+                        stmt = stmt.order_by(column.asc(), Knowledge.id.asc())
+                    else:
+                        stmt = stmt.order_by(column.desc(), Knowledge.id.asc())
+                else:
+                    stmt = stmt.order_by(Knowledge.updated_at.desc(), Knowledge.id.asc())
 
                 count_result = await db.execute(select(func.count()).select_from(stmt.subquery()))
                 total = count_result.scalar()
@@ -309,6 +336,14 @@ class KnowledgeTable:
 
                 knowledge_ids = [kb.id for kb, _ in items]
                 grants_map = await AccessGrants.get_grants_by_resources('knowledge', knowledge_ids, db=db)
+                file_counts = {}
+                if knowledge_ids:
+                    file_count_result = await db.execute(
+                        select(KnowledgeFile.knowledge_id, func.count(KnowledgeFile.id))
+                        .where(KnowledgeFile.knowledge_id.in_(knowledge_ids))
+                        .group_by(KnowledgeFile.knowledge_id)
+                    )
+                    file_counts = dict(file_count_result.all())
 
                 knowledge_bases = []
                 for knowledge_base, user in items:
@@ -323,6 +358,7 @@ class KnowledgeTable:
                                     )
                                 ).model_dump(),
                                 'user': (UserModel.model_validate(user).model_dump() if user else None),
+                                'file_count': file_counts.get(knowledge_base.id, 0),
                             }
                         )
                     )
@@ -369,6 +405,7 @@ class KnowledgeTable:
                             # to avoid PostgreSQL "invalid memory alloc request
                             # size" on large extracted-content rows (#24670).
                             content_text = File.data['content'].as_string()
+                            content_text = func.substr(content_text, 1, RAG_FILE_CONTENT_SEARCH_MAX_CHARS)
                             search_filter = or_(
                                 File.filename.ilike(f'%{q}%'),
                                 content_text.ilike(f'%{q}%'),
@@ -405,6 +442,7 @@ class KnowledgeTable:
                 if limit:
                     stmt = stmt.limit(limit)
 
+                stmt = stmt.options(defer(File.data))
                 result = await db.execute(stmt)
                 rows = result.all()
 
@@ -412,7 +450,13 @@ class KnowledgeTable:
                 for file, user, knowledge in rows:
                     items.append(
                         FileUserResponse(
-                            **FileModel.model_validate(file).model_dump(),
+                            id=file.id,
+                            user_id=file.user_id,
+                            hash=file.hash,
+                            filename=file.filename,
+                            meta=file.meta,
+                            created_at=file.created_at,
+                            updated_at=file.updated_at,
                             user=(UserResponse(**UserModel.model_validate(user).model_dump()) if user else None),
                             collection=(await self._to_knowledge_model(knowledge, db=db)).model_dump(),
                         )
@@ -448,20 +492,16 @@ class KnowledgeTable:
         user_groups = await Groups.get_groups_by_member_id(user_id, db=db)
         user_group_ids = {group.id for group in user_groups}
 
-        result = []
-        for knowledge_base in knowledge_bases:
-            if knowledge_base.user_id == user_id:
-                result.append(knowledge_base)
-            elif await AccessGrants.has_access(
-                user_id=user_id,
-                resource_type='knowledge',
-                resource_id=knowledge_base.id,
-                permission=permission,
-                user_group_ids=user_group_ids,
-                db=db,
-            ):
-                result.append(knowledge_base)
-        return result
+        # One grants query for all non-owned knowledge bases instead of one each
+        accessible_ids = await AccessGrants.get_accessible_resource_ids(
+            user_id=user_id,
+            resource_type='knowledge',
+            resource_ids=[kb.id for kb in knowledge_bases if kb.user_id != user_id],
+            permission=permission,
+            user_group_ids=user_group_ids,
+            db=db,
+        )
+        return [kb for kb in knowledge_bases if kb.user_id == user_id or kb.id in accessible_ids]
 
     async def get_knowledge_by_id(self, id: str, db: Optional[AsyncSession] = None) -> Optional[KnowledgeModel]:
         try:
@@ -554,6 +594,7 @@ class KnowledgeTable:
                             # to avoid PostgreSQL memory allocation failures on
                             # large content (#24670).
                             content_text = File.data['content'].as_string()
+                            content_text = func.substr(content_text, 1, RAG_FILE_CONTENT_SEARCH_MAX_CHARS)
                             stmt = stmt.filter(
                                 or_(
                                     File.filename.ilike(f'%{query_key}%'),
@@ -592,17 +633,23 @@ class KnowledgeTable:
                 if limit:
                     stmt = stmt.limit(limit)
 
+                stmt = stmt.options(defer(File.data))
                 result = await db.execute(stmt)
                 items = result.all()
 
-                files = []
-                for file, user in items:
-                    files.append(
-                        FileUserResponse(
-                            **FileModel.model_validate(file).model_dump(),
-                            user=(UserResponse(**UserModel.model_validate(user).model_dump()) if user else None),
-                        )
+                files = [
+                    FileUserResponse(
+                        id=file.id,
+                        user_id=file.user_id,
+                        hash=file.hash,
+                        filename=file.filename,
+                        meta=file.meta,
+                        created_at=file.created_at,
+                        updated_at=file.updated_at,
+                        user=(UserResponse(**UserModel.model_validate(user).model_dump()) if user else None),
                     )
+                    for file, user in items
+                ]
 
                 return KnowledgeFileListResponse(
                     items=files,
@@ -637,9 +684,25 @@ class KnowledgeTable:
     async def get_file_metadatas_by_id(
         self, knowledge_id: str, db: Optional[AsyncSession] = None
     ) -> list[FileMetadataResponse]:
+        """Column-only listing: File.data holds each file's full extracted
+        text, which metadata views must never load."""
         try:
-            files = await self.get_files_by_id(knowledge_id, db=db)
-            return [FileMetadataResponse(**file.model_dump()) for file in files]
+            async with get_async_db_context(db) as db:
+                result = await db.execute(
+                    select(File.id, File.hash, File.meta, File.created_at, File.updated_at)
+                    .join(KnowledgeFile, File.id == KnowledgeFile.file_id)
+                    .filter(KnowledgeFile.knowledge_id == knowledge_id)
+                )
+                return [
+                    FileMetadataResponse(
+                        id=row.id,
+                        hash=row.hash,
+                        meta=row.meta,
+                        created_at=row.created_at,
+                        updated_at=row.updated_at,
+                    )
+                    for row in result.all()
+                ]
         except Exception:
             return []
 
@@ -765,6 +828,25 @@ class KnowledgeTable:
             log.exception(e)
             return None
 
+    async def update_knowledge_meta_by_id(
+        self, id: str, meta: dict, db: Optional[AsyncSession] = None
+    ) -> Optional[KnowledgeModel]:
+        try:
+            async with get_async_db_context(db) as db:
+                await db.execute(
+                    update(Knowledge)
+                    .filter_by(id=id)
+                    .values(
+                        meta=meta,
+                        updated_at=int(time.time()),
+                    )
+                )
+                await db.commit()
+                return await self.get_knowledge_by_id(id=id, db=db)
+        except Exception as e:
+            log.exception(e)
+            return None
+
     async def delete_knowledge_by_id(self, id: str, db: Optional[AsyncSession] = None) -> bool:
         try:
             async with get_async_db_context(db) as db:

backend/open_webui/models/memories.py

@@ -4,11 +4,11 @@ from __future__ import annotations
 
 import time
 import uuid
-from typing import Optional
+from typing import Literal
 
 from open_webui.internal.db import Base, get_async_db_context
 from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Column, String, Text, delete, select
+from sqlalchemy import JSON, BigInteger, Column, Index, String, Text, delete, select
 from sqlalchemy.ext.asyncio import AsyncSession
 
 
@@ -16,10 +16,14 @@ class Memory(Base):  # user memory store
     """Stores user-created memory entries linked to a vector collection."""
 
     __tablename__ = 'memory'
+    __table_args__ = (Index('ix_memory_id_user_id', 'id', 'user_id'),)
 
     id = Column(String, primary_key=True, unique=True)
     user_id = Column(String, index=True)
+    type = Column(String, default='context', server_default='context', index=True)
+    path = Column(Text, nullable=True)
     content = Column(Text)  # free-form text learned from conversation
+    meta = Column(JSON, nullable=True)
     updated_at = Column(BigInteger)  # epoch seconds
     created_at = Column(BigInteger)  # epoch seconds
 
@@ -29,17 +33,27 @@ class MemoryModel(BaseModel):
 
     id: str
     user_id: str
+    type: Literal['user', 'context'] = 'context'
+    path: str | None = None
     content: str
+    meta: dict | None = None
     updated_at: int  # timestamp in epoch
     created_at: int  # timestamp in epoch
     model_config = ConfigDict(from_attributes=True)  # allows ORM mapping
 
 
 class MemoriesTable:
+    @staticmethod
+    def normalize_memory_type(memory_type: str | None = None) -> str:
+        return 'user' if memory_type == 'user' else 'context'
+
     async def insert_new_memory(
         self,
         user_id: str,
         content: str,
+        memory_type: str | None = None,
+        path: str | None = None,
+        meta: dict | None = None,
         db: AsyncSession | None = None,
     ) -> MemoryModel | None:
         """Persist a new memory entry and return the created model."""
@@ -48,20 +62,26 @@ class MemoriesTable:
             record = Memory(
                 id=str(uuid.uuid4()),
                 user_id=user_id,
+                type=self.normalize_memory_type(memory_type),
+                path=path,
                 content=content,
+                meta=meta,
                 created_at=now,
                 updated_at=now,
             )
             db.add(record)
             await db.commit()
-            await db.refresh(record)
             return MemoryModel.model_validate(record) if record else None
 
     async def update_memory_by_id_and_user_id(
         self,
         id: str,
         user_id: str,
-        content: str,
+        content: str | None,
+        memory_type: str | None = None,
+        path: str | None = None,
+        update_path: bool = False,
+        meta: dict | None = None,
         db: AsyncSession | None = None,
     ) -> MemoryModel | None:
         async with get_async_db_context(db) as db:
@@ -70,11 +90,17 @@ class MemoriesTable:
                 if not memory or memory.user_id != user_id:
                     return None
 
-                memory.content = content
+                if content is not None:
+                    memory.content = content
+                if memory_type is not None:
+                    memory.type = self.normalize_memory_type(memory_type)
+                if update_path:
+                    memory.path = path
+                if meta is not None:
+                    memory.meta = {**(memory.meta or {}), **meta}
                 memory.updated_at = int(time.time())
 
                 await db.commit()
-                await db.refresh(memory)
                 return MemoryModel.model_validate(memory)
             except Exception:
                 return None
@@ -139,5 +165,104 @@ class MemoriesTable:
             except Exception:
                 return False
 
+    async def apply_memory_operations(
+        self,
+        user_id: str,
+        operations: list[dict],
+        db: AsyncSession | None = None,
+    ) -> list[dict]:
+        now = int(time.time())
+        results: list[dict] = []
+
+        async with get_async_db_context(db) as db:
+            for operation in operations:
+                action = operation.get('action')
+
+                if action == 'add':
+                    content = operation.get('content', '').strip()
+                    memory_type = self.normalize_memory_type(operation.get('type'))
+                    path = operation.get('path')
+                    result = await db.execute(
+                        select(Memory).filter_by(user_id=user_id, content=content, type=memory_type, path=path)
+                    )
+                    existing = result.scalars().first()
+                    if existing:
+                        results.append(
+                            {
+                                'action': action,
+                                'status': 'skipped',
+                                'memory': MemoryModel.model_validate(existing),
+                                'reason': 'duplicate',
+                            }
+                        )
+                        continue
+
+                    memory = Memory(
+                        id=str(uuid.uuid4()),
+                        user_id=user_id,
+                        type=memory_type,
+                        path=path,
+                        content=content,
+                        meta=operation.get('meta'),
+                        created_at=now,
+                        updated_at=now,
+                    )
+                    db.add(memory)
+                    await db.flush()
+                    results.append(
+                        {'action': action, 'status': 'created', 'memory': MemoryModel.model_validate(memory)}
+                    )
+
+                elif action == 'replace':
+                    memory_id = operation.get('id')
+                    content = operation.get('content', '').strip()
+                    memory = await db.get(Memory, memory_id)
+                    if not memory or memory.user_id != user_id:
+                        raise ValueError(f'Memory not found: {memory_id}')
+
+                    memory.content = content
+                    if operation.get('type') is not None:
+                        memory.type = self.normalize_memory_type(operation.get('type'))
+                    if 'path' in operation:
+                        memory.path = operation.get('path')
+                    if operation.get('meta') is not None:
+                        memory.meta = {**(memory.meta or {}), **operation.get('meta')}
+                    memory.updated_at = now
+                    await db.flush()
+                    results.append(
+                        {'action': action, 'status': 'updated', 'memory': MemoryModel.model_validate(memory)}
+                    )
+
+                elif action == 'move':
+                    memory_id = operation.get('id')
+                    memory = await db.get(Memory, memory_id)
+                    if not memory or memory.user_id != user_id:
+                        raise ValueError(f'Memory not found: {memory_id}')
+
+                    memory.path = operation.get('path')
+                    if operation.get('meta') is not None:
+                        memory.meta = {**(memory.meta or {}), **operation.get('meta')}
+                    memory.updated_at = now
+                    await db.flush()
+                    results.append(
+                        {'action': action, 'status': 'updated', 'memory': MemoryModel.model_validate(memory)}
+                    )
+
+                elif action == 'remove':
+                    memory_id = operation.get('id')
+                    memory = await db.get(Memory, memory_id)
+                    if not memory or memory.user_id != user_id:
+                        raise ValueError(f'Memory not found: {memory_id}')
+
+                    await db.delete(memory)
+                    results.append({'action': action, 'status': 'deleted', 'id': memory_id})
+
+                else:
+                    raise ValueError(f'Unsupported memory operation: {action}')
+
+            await db.commit()
+
+        return results
+
 
 Memories = MemoriesTable()  # user memory registry

backend/open_webui/models/messages.py

@@ -328,7 +328,8 @@ class MessageTable:
         async with get_async_db_context(db) as db:
             message = await db.get(Message, parent_id)
 
-            if not message:
+            # Thread parent must belong to the requested channel; never disclose a foreign-channel message.
+            if not message or message.channel_id != channel_id:
                 return []
 
             result = await db.execute(
@@ -500,6 +501,71 @@ class MessageTable:
 
             return [Reactions(**reaction) for reaction in reactions.values()]
 
+    async def get_reactions_by_message_ids(
+        self, ids: list[str], db: Optional[AsyncSession] = None
+    ) -> dict[str, list[Reactions]]:
+        """Batch-fetch reactions for multiple messages in a single query.
+
+        Returns a dict mapping each message_id to its list of Reactions.
+        Messages with no reactions map to an empty list.
+        """
+        if not ids:
+            return {}
+
+        async with get_async_db_context(db) as db:
+            result = await db.execute(
+                select(MessageReaction, User)
+                .join(User, MessageReaction.user_id == User.id)
+                .filter(MessageReaction.message_id.in_(ids))
+            )
+            rows = result.all()
+
+            # Group by (message_id, reaction_name)
+            grouped: dict[str, dict[str, dict]] = {mid: {} for mid in ids}
+            for reaction, user in rows:
+                mid = reaction.message_id
+                if mid not in grouped:
+                    grouped[mid] = {}
+                if reaction.name not in grouped[mid]:
+                    grouped[mid][reaction.name] = {
+                        'name': reaction.name,
+                        'users': [],
+                        'count': 0,
+                    }
+                grouped[mid][reaction.name]['users'].append(
+                    {
+                        'id': user.id,
+                        'name': user.name,
+                    }
+                )
+                grouped[mid][reaction.name]['count'] += 1
+
+            return {mid: [Reactions(**r) for r in reactions.values()] for mid, reactions in grouped.items()}
+
+    async def get_thread_reply_counts_by_message_ids(
+        self, ids: list[str], db: Optional[AsyncSession] = None
+    ) -> dict[str, tuple[int, int | None]]:
+        """Batch-fetch reply counts and latest reply timestamps for multiple parent messages.
+
+        Returns a dict mapping each parent message_id to a
+        (reply_count, latest_reply_created_at) tuple.
+        Messages with no replies are omitted from the result.
+        """
+        if not ids:
+            return {}
+
+        async with get_async_db_context(db) as db:
+            result = await db.execute(
+                select(
+                    Message.parent_id,
+                    func.count(Message.id),
+                    func.max(Message.created_at),
+                )
+                .filter(Message.parent_id.in_(ids))
+                .group_by(Message.parent_id)
+            )
+            return {row[0]: (row[1], row[2]) for row in result.all()}
+
     async def remove_reaction_by_id_and_user_id_and_name(
         self, id: str, user_id: str, name: str, db: Optional[AsyncSession] = None
     ) -> bool:

backend/open_webui/models/models.py

@@ -3,7 +3,8 @@ from __future__ import annotations
 import json
 import logging
 import time
-from typing import Optional
+from copy import deepcopy
+from typing import Any, Optional
 
 from open_webui.internal.db import Base, JSONField, get_async_db_context
 from open_webui.models.access_grants import AccessGrantModel, AccessGrants
@@ -22,6 +23,38 @@ log = logging.getLogger(__name__)
 _warned_profile_urls: set[str] = set()
 
 
+def strip_extracted_content_from_model_knowledge(knowledge: Any) -> Any:
+    """Drop duplicated extracted text from ModelMeta.knowledge."""
+    if not isinstance(knowledge, list):
+        return knowledge
+
+    sanitized = []
+
+    for item in knowledge:
+        if not isinstance(item, dict):
+            sanitized.append(item)
+            continue
+
+        next_item = item
+        data = item.get('data')
+        if isinstance(data, dict) and 'content' in data:
+            next_item = deepcopy(item)
+            next_item.get('data', {}).pop('content', None)
+
+        file = next_item.get('file')
+        file_data = file.get('data') if isinstance(file, dict) else None
+        if isinstance(file_data, dict) and 'content' in file_data:
+            if next_item is item:
+                next_item = deepcopy(item)
+                file = next_item.get('file')
+                file_data = file.get('data') if isinstance(file, dict) else None
+            file_data.pop('content', None)
+
+        sanitized.append(next_item)
+
+    return sanitized
+
+
 # --- Models DB Schema ---
 
 
@@ -37,6 +70,7 @@ class ModelMeta(BaseModel):
     profile_image_url: str | None = None
     description: str | None = Field(default=None, description='User-facing description of the model.')
     capabilities: dict | None = None
+    knowledge: list[Any] | None = None
 
     model_config = ConfigDict(extra='allow')
 
@@ -56,6 +90,11 @@ class ModelMeta(BaseModel):
                 )
             return None
 
+    @field_validator('knowledge', mode='before')
+    @classmethod
+    def strip_knowledge_content(cls, v):
+        return strip_extracted_content_from_model_knowledge(v)
+
     @model_validator(mode='before')
     @classmethod
     def normalize_tags(cls, data):
@@ -152,11 +191,19 @@ class ModelsTable:
         access_grants: list[AccessGrantModel | None] = None,
         db: AsyncSession | None = None,
     ) -> ModelModel:
-        model_data = ModelModel.model_validate(model).model_dump(exclude={'access_grants'})
-        model_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(model_data['id'], db=db)
+        if isinstance(model.meta, dict):
+            knowledge = model.meta.get('knowledge')
+            stripped_knowledge = strip_extracted_content_from_model_knowledge(knowledge)
+            if stripped_knowledge != knowledge:
+                model.meta = {**model.meta, 'knowledge': stripped_knowledge}
+                if db is not None:
+                    await db.commit()
+
+        model_model = ModelModel.model_validate(model)
+        model_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(model_model.id, db=db)
         )
-        return ModelModel.model_validate(model_data)
+        return model_model
 
     async def insert_new_model(
         self, form_data: ModelForm, user_id: str, db: AsyncSession | None = None
@@ -173,7 +220,6 @@ class ModelsTable:
                 )
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 await AccessGrants.set_access_grants('model', result.id, form_data.access_grants, db=db)
 
                 if result:
@@ -229,10 +275,25 @@ class ModelsTable:
                 )
             return models
 
-    async def get_base_models(self, db: AsyncSession | None = None) -> list[ModelModel]:
+    @staticmethod
+    def _meta_has_tag(meta: dict | None, tag: str) -> bool:
+        if not meta:
+            return False
+
+        for raw_tag in meta.get('tags', []):
+            name = raw_tag.get('name') if isinstance(raw_tag, dict) else str(raw_tag)
+            if name == tag:
+                return True
+
+        return False
+
+    async def get_base_models(self, tag: str | None = None, db: AsyncSession | None = None) -> list[ModelModel]:
         async with get_async_db_context(db) as db:
-            result = await db.execute(select(Model).filter(Model.base_model_id == None))
+            result = await db.execute(select(Model).filter(Model.base_model_id.is_(None)))
             all_models = result.scalars().all()
+            if tag:
+                all_models = [model for model in all_models if self._meta_has_tag(model.meta, tag)]
+
             model_ids = [model.id for model in all_models]
             grants_map = await AccessGrants.get_grants_by_resources('model', model_ids, db=db)
             return [
@@ -241,26 +302,26 @@ class ModelsTable:
             ]
 
     async def get_models_by_user_id(
-        self, user_id: str, permission: str = 'write', db: AsyncSession | None = None
+        self,
+        user_id: str,
+        permission: str = 'write',
+        db: AsyncSession | None = None,
+        user_group_ids: set[str] | None = None,
     ) -> list[ModelUserResponse]:
         models = await self.get_models(db=db)
-        user_groups = await Groups.get_groups_by_member_id(user_id, db=db)
-        user_group_ids = {group.id for group in user_groups}
+        if user_group_ids is None:
+            user_group_ids = {group.id for group in await Groups.get_groups_by_member_id(user_id, db=db)}
 
-        result = []
-        for model in models:
-            if model.user_id == user_id:
-                result.append(model)
-            elif await AccessGrants.has_access(
-                user_id=user_id,
-                resource_type='model',
-                resource_id=model.id,
-                permission=permission,
-                user_group_ids=user_group_ids,
-                db=db,
-            ):
-                result.append(model)
-        return result
+        # One grants query for all non-owned models instead of one per model
+        accessible_ids = await AccessGrants.get_accessible_resource_ids(
+            user_id=user_id,
+            resource_type='model',
+            resource_ids=[model.id for model in models if model.user_id != user_id],
+            permission=permission,
+            user_group_ids=user_group_ids,
+            db=db,
+        )
+        return [model for model in models if model.user_id == user_id or model.id in accessible_ids]
 
     def _has_permission(self, db, query, filter: dict, permission: str = 'read'):
         return AccessGrants.has_permission_filter(
@@ -395,11 +456,14 @@ class ModelsTable:
         self,
         user_id: str,
         is_admin: bool = False,
+        is_base_model: bool = False,
         db: AsyncSession | None = None,
     ) -> set[str]:
         """Extract unique tag names from model meta, querying only the meta column."""
         async with get_async_db_context(db) as db:
-            stmt = select(Model.meta).filter(Model.base_model_id != None)
+            stmt = select(Model.meta).filter(
+                Model.base_model_id.is_(None) if is_base_model else Model.base_model_id.is_not(None)
+            )
 
             if not is_admin:
                 user_groups = await Groups.get_groups_by_member_id(user_id, db=db)
@@ -465,7 +529,6 @@ class ModelsTable:
                 model.is_active = not model.is_active
                 model.updated_at = int(time.time())
                 await db.commit()
-                await db.refresh(model)
 
                 return await self._to_model_model(model, db=db)
             except Exception:
@@ -492,13 +555,12 @@ class ModelsTable:
         try:
             async with get_async_db_context(db) as db:
                 result = await db.execute(select(Model).filter_by(id=id))
-                model_obj = result.scalars().first()
-                if not model_obj:
+                model = result.scalars().first()
+                if not model:
                     return None
-                model_obj.updated_at = int(time.time())
+                model.updated_at = int(time.time())
                 await db.commit()
-                await db.refresh(model_obj)
-                return await self._to_model_model(model_obj, db=db)
+                return await self._to_model_model(model, db=db)
         except Exception as e:
             log.exception(f'Failed to update the model updated_at by id {id}: {e}')
             return None

backend/open_webui/models/notes.py

@@ -106,11 +106,12 @@ class NoteTable:
         db: Optional[AsyncSession] = None,
     ) -> NoteModel:
         # We exclude access_grants to inject them
-        note_data = NoteModel.model_validate(note).model_dump(exclude={'access_grants'})
-        note_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(note_data['id'], db=db)
+        note_model = NoteModel.model_validate(note)
+        note_model.data = note_model.data or {}
+        note_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(note_model.id, db=db)
         )
-        return NoteModel.model_validate(note_data)
+        return note_model
 
     def _has_permission(self, db, query, filter: dict, permission: str = 'read'):
         return AccessGrants.has_permission_filter(
@@ -308,9 +309,12 @@ class NoteTable:
             if 'title' in form_data:
                 note.title = form_data['title']
             if 'data' in form_data:
-                note.data = {**note.data, **form_data['data']}
+                note.data = {**(note.data or {}), **(form_data['data'] or {})}
             if 'meta' in form_data:
-                note.meta = {**note.meta, **form_data['meta']}
+                note.meta = {**(note.meta or {}), **(form_data['meta'] or {})}
+
+            if not db.is_modified(note) and 'access_grants' not in form_data:
+                return await self._to_note_model(note, db=db)
 
             if 'access_grants' in form_data:
                 await AccessGrants.set_access_grants('note', id, form_data['access_grants'], db=db)

backend/open_webui/models/oauth_sessions.py

@@ -128,7 +128,6 @@ class OAuthSessionTable:
 
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
 
                 if result:
                     # Make a copy of the model data before closing session

backend/open_webui/models/prompt_history.py

@@ -70,7 +70,6 @@ class PromptHistoryTable:
             )
             db.add(history)
             await db.commit()
-            await db.refresh(history)
             return PromptHistoryModel.model_validate(history)
 
     async def get_history_by_prompt_id(

backend/open_webui/models/prompts.py

@@ -103,11 +103,11 @@ class PromptsTable:
         access_grants: list[AccessGrantModel | None] = None,
         db: AsyncSession | None = None,
     ) -> PromptModel:
-        prompt_data = PromptModel.model_validate(prompt).model_dump(exclude={'access_grants'})
-        prompt_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(prompt_data['id'], db=db)
+        prompt_model = PromptModel.model_validate(prompt)
+        prompt_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(prompt_model.id, db=db)
         )
-        return PromptModel.model_validate(prompt_data)
+        return prompt_model
 
     async def insert_new_prompt(
         self, user_id: str, form_data: PromptForm, db: AsyncSession | None = None
@@ -132,7 +132,6 @@ class PromptsTable:
                 )
                 session.add(record)
                 await session.commit()
-                await session.refresh(record)  # populate generated defaults
 
                 await AccessGrants.set_access_grants(
                     'prompt',
@@ -169,7 +168,6 @@ class PromptsTable:
                 if history_entry:
                     record.version_id = history_entry.id
                     await session.commit()
-                    await session.refresh(record)  # re-read version_id
 
                 return await self._to_prompt_model(record, db=session)
             except Exception as e:
@@ -637,7 +635,6 @@ class PromptsTable:
                     prompt.is_active = not prompt.is_active
                     prompt.updated_at = int(time.time())
                     await session.commit()
-                    await session.refresh(prompt)
                     return await self._to_prompt_model(prompt, db=session)
                 return None
         except Exception:

backend/open_webui/models/shared_chats.py

@@ -201,5 +201,15 @@ class SharedChatsTable:
         except Exception:
             return False
 
+    async def delete_all_by_user_id(self, user_id: str, db: Optional[AsyncSession] = None) -> bool:
+        """Delete all shared chats created by a user."""
+        try:
+            async with get_async_db_context(db) as db:
+                await db.execute(delete(SharedChat).filter_by(user_id=user_id))
+                await db.commit()
+                return True
+        except Exception:
+            return False
+
 
 SharedChats = SharedChatsTable()

backend/open_webui/models/skills.py

@@ -113,11 +113,11 @@ class SkillsTable:
         access_grants: Optional[list[AccessGrantModel]] = None,
         db: Optional[AsyncSession] = None,
     ) -> SkillModel:
-        skill_data = SkillModel.model_validate(skill).model_dump(exclude={'access_grants'})
-        skill_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(skill_data['id'], db=db)
+        skill_model = SkillModel.model_validate(skill)
+        skill_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(skill_model.id, db=db)
         )
-        return SkillModel.model_validate(skill_data)
+        return skill_model
 
     async def insert_new_skill(
         self,
@@ -137,7 +137,6 @@ class SkillsTable:
                 )
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 await AccessGrants.set_access_grants('skill', result.id, form_data.access_grants, db=db)
                 if result:
                     return await self._to_skill_model(result, db=db)
@@ -259,7 +258,26 @@ class SkillsTable:
                         permission='read',
                     )
 
-                stmt = stmt.order_by(Skill.updated_at.desc())
+                order_by = filter.get('order_by')
+                direction = filter.get('direction')
+
+                if order_by == 'name':
+                    if direction == 'asc':
+                        stmt = stmt.order_by(Skill.name.asc())
+                    else:
+                        stmt = stmt.order_by(Skill.name.desc())
+                elif order_by == 'created_at':
+                    if direction == 'asc':
+                        stmt = stmt.order_by(Skill.created_at.asc())
+                    else:
+                        stmt = stmt.order_by(Skill.created_at.desc())
+                elif order_by == 'updated_at':
+                    if direction == 'asc':
+                        stmt = stmt.order_by(Skill.updated_at.asc())
+                    else:
+                        stmt = stmt.order_by(Skill.updated_at.desc())
+                else:
+                    stmt = stmt.order_by(Skill.updated_at.desc())
 
                 # Count BEFORE pagination
                 count_result = await db.execute(select(func.count()).select_from(stmt.subquery()))
@@ -307,8 +325,8 @@ class SkillsTable:
                 if access_grants is not None:
                     await AccessGrants.set_access_grants('skill', id, access_grants, db=db)
 
-                skill = await db.get(Skill, id)
-                await db.refresh(skill)
+                # populate_existing: the Core update above bypasses any identity-map copy
+                skill = await db.get(Skill, id, populate_existing=True)
                 return await self._to_skill_model(skill, db=db)
         except Exception:
             return None
@@ -324,7 +342,6 @@ class SkillsTable:
                 skill.is_active = not skill.is_active
                 skill.updated_at = int(time.time())
                 await db.commit()
-                await db.refresh(skill)
 
                 return await self._to_skill_model(skill, db=db)
             except Exception:

backend/open_webui/models/tags.py

@@ -63,7 +63,6 @@ class TagTable:
                 record = Tag(id=tag_id, user_id=user_id, name=name)
                 db.add(record)
                 await db.commit()
-                await db.refresh(record)
                 return TagModel.model_validate(record) if record else None
             except Exception as e:
                 log.exception('Error inserting tag %r: %s', name, e)

backend/open_webui/models/tools.py

@@ -10,6 +10,7 @@ from open_webui.internal.db import Base, JSONField, get_async_db_context
 from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 from open_webui.models.groups import Groups
 from open_webui.models.users import UserResponse, Users
+from open_webui.utils.valves import decrypt_valves, encrypt_valves
 from pydantic import BaseModel, ConfigDict, Field
 from sqlalchemy import BigInteger, Column, String, Text, delete, select, update
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -35,13 +36,15 @@ class Tool(Base):  # database table definition
 class ToolMeta(BaseModel):
     description: str | None = None
     manifest: dict | None = {}
+    has_user_valves: bool = False
 
 
 class ToolModel(BaseModel):
     id: str
-    user_id: str
+    user_id: str | None = None  # may be null for legacy/malformed records
     name: str
-    content: str
+    # None when listed with defer_content=True (source skipped for listings)
+    content: str | None = None
     specs: list[dict]
     meta: ToolMeta
     access_grants: list[AccessGrantModel] = Field(default_factory=list)
@@ -63,7 +66,7 @@ class ToolUserModel(ToolModel):
 
 class ToolResponse(BaseModel):
     id: str
-    user_id: str
+    user_id: str | None = None  # may be null for legacy/malformed records
     name: str
     meta: ToolMeta
     access_grants: list[AccessGrantModel] = Field(default_factory=list)
@@ -103,11 +106,11 @@ class ToolsTable:
         access_grants: list[AccessGrantModel | None] = None,
         db: AsyncSession | None = None,
     ) -> ToolModel:
-        tool_data = ToolModel.model_validate(tool).model_dump(exclude={'access_grants'})
-        tool_data['access_grants'] = (
-            access_grants if access_grants is not None else await self._get_access_grants(tool_data['id'], db=db)
+        tool_model = ToolModel.model_validate(tool)
+        tool_model.access_grants = (
+            access_grants if access_grants is not None else await self._get_access_grants(tool_model.id, db=db)
         )
-        return ToolModel.model_validate(tool_data)
+        return tool_model
 
     async def insert_new_tool(
         self,
@@ -129,7 +132,6 @@ class ToolsTable:
                 )
                 db.add(result)
                 await db.commit()
-                await db.refresh(result)
                 await AccessGrants.set_access_grants('tool', result.id, form_data.access_grants, db=db)
                 if result:
                     return await self._to_tool_model(result, db=db)
@@ -169,11 +171,18 @@ class ToolsTable:
 
     async def get_tools(self, defer_content: bool = False, db: AsyncSession | None = None) -> list[ToolUserModel]:
         async with get_async_db_context(db) as db:
-            stmt = select(Tool).order_by(Tool.updated_at.desc())
             if defer_content:
-                stmt = stmt
-            result = await db.execute(stmt)
-            all_tools = result.scalars().all()
+                # Skip Tool.content (plugin source, potentially large) via a
+                # column select; Row attributes satisfy from_attributes.
+                result = await db.execute(
+                    select(
+                        Tool.id, Tool.user_id, Tool.name, Tool.specs, Tool.meta, Tool.updated_at, Tool.created_at
+                    ).order_by(Tool.updated_at.desc())
+                )
+                all_tools = result.all()
+            else:
+                result = await db.execute(select(Tool).order_by(Tool.updated_at.desc()))
+                all_tools = result.scalars().all()
 
             user_ids = list(set(tool.user_id for tool in all_tools))
             tool_ids = [tool.id for tool in all_tools]
@@ -212,27 +221,23 @@ class ToolsTable:
         user_groups = await Groups.get_groups_by_member_id(user_id, db=db)
         user_group_ids = {group.id for group in user_groups}
 
-        result = []
-        for tool in tools:
-            if tool.user_id == user_id:
-                result.append(tool)
-            elif await AccessGrants.has_access(
-                user_id=user_id,
-                resource_type='tool',
-                resource_id=tool.id,
-                permission=permission,
-                user_group_ids=user_group_ids,
-                db=db,
-            ):
-                result.append(tool)
-        return result
+        # One grants query for all non-owned tools instead of one per tool
+        accessible_ids = await AccessGrants.get_accessible_resource_ids(
+            user_id=user_id,
+            resource_type='tool',
+            resource_ids=[tool.id for tool in tools if tool.user_id != user_id],
+            permission=permission,
+            user_group_ids=user_group_ids,
+            db=db,
+        )
+        return [tool for tool in tools if tool.user_id == user_id or tool.id in accessible_ids]
 
     async def get_tool_valves_by_id(self, id: str, db: AsyncSession | None = None) -> dict | None:
         try:
             async with get_async_db_context(db) as db:
                 tool = await db.get(Tool, id)
-                return tool.valves if tool.valves else {}
-        except Exception as e:
+                return decrypt_valves(tool.valves if tool else None)
+        except Exception:
             log.exception(f'Error getting tool valves by id {id}')
             return None
 
@@ -241,7 +246,9 @@ class ToolsTable:
     ) -> ToolValves | None:
         try:
             async with get_async_db_context(db) as db:
-                await db.execute(update(Tool).filter_by(id=id).values(valves=valves, updated_at=int(time.time())))
+                await db.execute(
+                    update(Tool).filter_by(id=id).values(valves=encrypt_valves(valves), updated_at=int(time.time()))
+                )
                 await db.commit()
                 return await self.get_tool_by_id(id, db=db)
         except Exception:
@@ -260,7 +267,7 @@ class ToolsTable:
             if 'valves' not in user_settings['tools']:
                 user_settings['tools']['valves'] = {}
 
-            return user_settings['tools']['valves'].get(id, {})
+            return decrypt_valves(user_settings['tools']['valves'].get(id))
         except Exception as e:
             log.exception(f'Error getting user values by id {id} and user_id {user_id}: {e}')
             return None
@@ -278,12 +285,12 @@ class ToolsTable:
             if 'valves' not in user_settings['tools']:
                 user_settings['tools']['valves'] = {}
 
-            user_settings['tools']['valves'][id] = valves
+            user_settings['tools']['valves'][id] = encrypt_valves(valves)
 
             # Update the user settings in the database
             await Users.update_user_by_id(user_id, {'settings': user_settings}, db=db)
 
-            return user_settings['tools']['valves'][id]
+            return valves
         except Exception as e:
             log.exception(f'Error updating user valves by id {id} and user_id {user_id}: {e}')
             return None
@@ -297,8 +304,8 @@ class ToolsTable:
                 if access_grants is not None:
                     await AccessGrants.set_access_grants('tool', id, access_grants, db=db)
 
-                tool = await db.get(Tool, id)
-                await db.refresh(tool)
+                # populate_existing: the Core update above bypasses any identity-map copy
+                tool = await db.get(Tool, id, populate_existing=True)
                 return await self._to_tool_model(tool, db=db)
         except Exception:
             return None

backend/open_webui/models/users.py

@@ -9,7 +9,7 @@ from open_webui.env import DATABASE_USER_ACTIVE_STATUS_UPDATE_INTERVAL
 from open_webui.internal.db import Base, JSONField, get_async_db_context
 from open_webui.utils.misc import throttle
 from open_webui.utils.validate import validate_profile_image_url
-from pydantic import BaseModel, ConfigDict, field_validator, model_validator
+from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator
 from sqlalchemy import (
     JSON,
     BigInteger,
@@ -69,6 +69,7 @@ class User(Base):  # identity & profile
 
     # Metadata
     info = Column(JSON, nullable=True)
+    variables = Column(JSON, nullable=True)
     settings = Column(JSON, nullable=True)
     oauth = Column(JSON, nullable=True)
     scim = Column(JSON, nullable=True)
@@ -105,6 +106,7 @@ class UserModel(BaseModel):
     status_expires_at: int | None = None
 
     info: dict | None = None
+    variables: dict = Field(default_factory=dict, exclude=True)
     settings: UserSettings | None = None
 
     oauth: dict | None = None
@@ -126,6 +128,11 @@ class UserModel(BaseModel):
         self.profile_image_url = self.profile_image_url or _DEFAULT_PROFILE_IMAGE_URL.format(user_id=self.id)
         return self
 
+    @field_validator('variables', mode='before')
+    @classmethod
+    def normalize_variables(cls, value):
+        return value if isinstance(value, dict) else {}
+
 
 class UserStatusModel(UserModel):
     is_active: bool = False
@@ -279,6 +286,11 @@ class UsersTable:
         oauth: dict | None = None,
         db: AsyncSession | None = None,
     ) -> UserModel | None:
+        try:
+            profile_image_url = validate_profile_image_url(profile_image_url)
+        except ValueError:
+            profile_image_url = '/user.png'
+
         async with get_async_db_context(db) as session:
             user = UserModel(
                 **{
@@ -297,7 +309,6 @@ class UsersTable:
             result = User(**user.model_dump())
             session.add(result)
             await session.commit()
-            await session.refresh(result)
             return user if result else None
 
     # database read methods
@@ -561,13 +572,6 @@ class UsersTable:
             row = (await session.execute(stmt)).scalars().first()
             return UserModel.model_validate(row) if row else None
 
-    async def get_user_webhook_url_by_id(self, id: str, db: AsyncSession | None = None) -> str | None:
-        async with get_async_db_context(db) as session:
-            user = await session.get(User, id)
-            if user and user.settings:
-                return user.settings.get('ui', {}).get('notifications', {}).get('webhook_url', None)
-            return None
-
     async def get_num_users_active_today(self, db: AsyncSession | None = None) -> int | None:
         async with get_async_db_context(db) as session:
             current_timestamp = int(time.time())
@@ -584,7 +588,6 @@ class UsersTable:
                 return None
             user.role = role
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     async def update_user_status_by_id(
@@ -597,7 +600,6 @@ class UsersTable:
             for key, value in form_data.model_dump(exclude_none=True).items():
                 setattr(user, key, value)
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     async def update_user_profile_image_url_by_id(
@@ -606,13 +608,17 @@ class UsersTable:
         profile_image_url: str,
         db: AsyncSession | None = None,
     ) -> UserModel | None:
+        try:
+            profile_image_url = validate_profile_image_url(profile_image_url)
+        except ValueError:
+            profile_image_url = '/user.png'
+
         async with get_async_db_context(db) as session:
             user = await session.get(User, id)
             if user is None:
                 return None
             user.profile_image_url = profile_image_url
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     @throttle(DATABASE_USER_ACTIVE_STATUS_UPDATE_INTERVAL)
@@ -633,7 +639,6 @@ class UsersTable:
             oauth[provider] = {'sub': sub}
             user.oauth = oauth
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     async def update_user_scim_by_id(
@@ -652,7 +657,6 @@ class UsersTable:
             scim[provider] = {'external_id': external_id}
             user.scim = scim
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     async def update_user_by_id(self, id: str, updated: dict, db: AsyncSession | None = None) -> UserModel | None:
@@ -663,7 +667,6 @@ class UsersTable:
             for key, value in updated.items():
                 setattr(user, key, value)
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     # settings update helper
@@ -678,7 +681,6 @@ class UsersTable:
             user_settings.update(updated)
             user.settings = user_settings
             await session.commit()
-            await session.refresh(user)
             return UserModel.model_validate(user)
 
     async def delete_user_by_id(self, id: str, db: AsyncSession | None = None) -> bool:
@@ -725,8 +727,8 @@ class UsersTable:
 
     async def get_valid_user_ids(self, user_ids: list[str], db: AsyncSession | None = None) -> list[str]:
         async with get_async_db_context(db) as session:
-            result = await session.execute(select(User).where(User.id.in_(user_ids)))
-            return [u.id for u in result.scalars().all()]
+            result = await session.execute(select(User.id).where(User.id.in_(user_ids)))
+            return list(result.scalars().all())
 
     async def get_super_admin_user(self, db: AsyncSession | None = None) -> UserModel | None:
         async with get_async_db_context(db) as session:
@@ -752,11 +754,11 @@ class UsersTable:
 
     async def is_user_active(self, user_id: str, db: AsyncSession | None = None) -> bool:
         async with get_async_db_context(db) as session:
-            user = await session.get(User, user_id)
-            if user and user.last_active_at:
+            last_active_at = await session.scalar(select(User.last_active_at).where(User.id == user_id))
+            if last_active_at:
                 # Consider user active if last_active_at within the last 3 minutes
                 three_minutes_ago = int(time.time()) - 180
-                return user.last_active_at >= three_minutes_ago
+                return last_active_at >= three_minutes_ago
             return False
 
 

backend/open_webui/retrieval/external.py

@@ -0,0 +1,379 @@
+import asyncio
+import logging
+import re
+import time
+from typing import Any, Optional
+
+from open_webui.config import RAG_EMBEDDING_QUERY_PREFIX
+from open_webui.models.config import Config
+from open_webui.models.knowledge import KnowledgeModel
+
+log = logging.getLogger(__name__)
+
+EXTERNAL_KNOWLEDGE_CONNECTIONS_CONFIG_KEY = 'external_knowledge.connections'
+IDENTIFIER_RE = re.compile(r'^[A-Za-z_][A-Za-z0-9_]*$')
+
+
+async def _get_external_connection(connection_id: str) -> Optional[dict]:
+    connections = await Config.get(EXTERNAL_KNOWLEDGE_CONNECTIONS_CONFIG_KEY, []) or []
+    return next((connection for connection in connections if connection.get('id') == connection_id), None)
+
+
+def _get_path(data: Any, path: Optional[str], default=None):
+    if not path:
+        return default
+    value = data
+    for part in path.split('.'):
+        if isinstance(value, dict):
+            value = value.get(part, default)
+        else:
+            return default
+    return value
+
+
+def _normalize_result(result: dict, mapping: dict, knowledge: KnowledgeModel, distance: Optional[float] = None) -> dict:
+    content = _get_path(result, mapping.get('content_field', 'content'), '')
+    title = _get_path(result, mapping.get('title_field', 'title'), None)
+    source = _get_path(result, mapping.get('source_field', 'source'), None)
+    url = _get_path(result, mapping.get('url_field', 'url'), None)
+    document_id = _get_path(result, mapping.get('document_id_field', 'document_id'), None)
+    page = _get_path(result, mapping.get('page_field', 'page'), None)
+    metadata = _get_path(result, mapping.get('metadata_field', 'metadata'), {}) or {}
+    score = _get_path(result, mapping.get('score_field', 'score'), distance)
+
+    if not isinstance(metadata, dict):
+        metadata = {'external_metadata': metadata}
+
+    source_name = source or title or metadata.get('source') or metadata.get('name') or knowledge.name
+    metadata.update(
+        {
+            'name': title or source_name,
+            'source': source_name,
+            'url': url,
+            'file_id': document_id or f'external-{knowledge.id}',
+            'knowledge_id': knowledge.id,
+            'knowledge_name': knowledge.name,
+            'external': True,
+        }
+    )
+    if page is not None:
+        metadata['page'] = page
+    if document_id is not None:
+        metadata['document_id'] = document_id
+
+    return {
+        'content': content,
+        'metadata': metadata,
+        'distance': score,
+    }
+
+
+def _source_config(knowledge: KnowledgeModel) -> dict:
+    external = (knowledge.meta or {}).get('external', {})
+    source = external.get('source') or {}
+    return source.get('config') or {}
+
+
+def _root_field(path: Optional[str]) -> Optional[str]:
+    if not path:
+        return None
+    return path.split('.')[0]
+
+
+def _safe_identifier(value: str, label: str) -> str:
+    if not value or not IDENTIFIER_RE.match(value):
+        raise RuntimeError(f'Invalid {label}')
+    return value
+
+
+async def _retrieve_qdrant(connection, auth_config, knowledge, query, count, embedding_function) -> list[dict]:
+    try:
+        from qdrant_client import QdrantClient
+    except ImportError as exc:
+        raise RuntimeError('qdrant-client is not installed') from exc
+
+    if not embedding_function:
+        raise RuntimeError('Embedding function is not configured')
+
+    config = connection.get('config') or {}
+    external = (knowledge.meta or {}).get('external', {})
+    source = external.get('source') or {}
+    collection_name = source.get('name')
+    if not collection_name:
+        raise RuntimeError('External source collection is not configured')
+    source_config = _source_config(knowledge)
+    vector_field = source_config.get('vector_field') or None
+
+    vector = await embedding_function(query, prefix=RAG_EMBEDDING_QUERY_PREFIX)
+
+    def _search():
+        client = QdrantClient(
+            url=connection.get('endpoint'),
+            api_key=(auth_config or {}).get('api_key'),
+            timeout=config.get('timeout') or 30,
+        )
+        return client.query_points(
+            collection_name=collection_name,
+            query=vector,
+            using=vector_field,
+            limit=count,
+        )
+
+    response = await asyncio.to_thread(_search)
+    mapping = {
+        'content_field': source_config.get('content_field') or 'payload.text',
+        'metadata_field': source_config.get('metadata_field') or 'payload.metadata',
+        'document_id_field': source_config.get('document_id_field') or 'id',
+        'score_field': 'score',
+    }
+
+    normalized = []
+    for point in response.points:
+        normalized.append(_normalize_result(point.model_dump(), mapping, knowledge, distance=point.score))
+    return normalized
+
+
+async def _retrieve_milvus(connection, auth_config, knowledge, query, count, embedding_function) -> list[dict]:
+    try:
+        from pymilvus import MilvusClient
+    except ImportError as exc:
+        raise RuntimeError('pymilvus is not installed') from exc
+
+    if not embedding_function:
+        raise RuntimeError('Embedding function is not configured')
+
+    config = connection.get('config') or {}
+    external = (knowledge.meta or {}).get('external', {})
+    source = external.get('source') or {}
+    collection_name = source.get('name')
+    if not collection_name:
+        raise RuntimeError('Milvus collection is not configured')
+    source_config = _source_config(knowledge)
+    vector_field = source_config.get('vector_field') or 'vector'
+    content_field = source_config.get('content_field') or 'data.text'
+    metadata_field = source_config.get('metadata_field') or 'metadata'
+
+    vector = await embedding_function(query, prefix=RAG_EMBEDDING_QUERY_PREFIX)
+
+    def _search():
+        client_kwargs = {
+            'uri': connection.get('endpoint'),
+        }
+        token = (auth_config or {}).get('api_key') or (auth_config or {}).get('token')
+        if token:
+            client_kwargs['token'] = token
+        if config.get('db_name'):
+            client_kwargs['db_name'] = config.get('db_name')
+
+        client = MilvusClient(**client_kwargs)
+        output_fields = {
+            field
+            for field in (
+                _root_field(content_field),
+                _root_field(metadata_field),
+                _root_field(source_config.get('document_id_field')),
+            )
+            if field and field != vector_field
+        }
+        kwargs = {
+            'collection_name': collection_name,
+            'data': [vector],
+            'anns_field': vector_field,
+            'limit': count,
+            'output_fields': list(output_fields),
+        }
+        return client.search(**kwargs)
+
+    response = await asyncio.to_thread(_search)
+    mapping = {
+        'content_field': content_field,
+        'metadata_field': metadata_field,
+        'document_id_field': source_config.get('document_id_field') or 'id',
+        'score_field': 'distance',
+    }
+
+    normalized = []
+    for hit in response[0] if response else []:
+        item = dict(hit)
+        entity = item.get('entity') or {}
+        result = {
+            **entity,
+            'id': item.get('id') or entity.get('id'),
+            'distance': item.get('distance'),
+        }
+        normalized.append(_normalize_result(result, mapping, knowledge, distance=item.get('distance')))
+    return normalized
+
+
+async def _retrieve_pgvector(connection, auth_config, knowledge, query, count, embedding_function) -> list[dict]:
+    try:
+        import psycopg
+        from pgvector.psycopg import register_vector
+        from psycopg.rows import dict_row
+    except ImportError as exc:
+        raise RuntimeError('psycopg and pgvector are required for pgvector retrieval') from exc
+
+    if not embedding_function:
+        raise RuntimeError('Embedding function is not configured')
+
+    config = connection.get('config') or {}
+    external = (knowledge.meta or {}).get('external', {})
+    source = external.get('source') or {}
+    collection_name = source.get('name')
+    if not collection_name:
+        raise RuntimeError('pgvector collection is not configured')
+    source_config = _source_config(knowledge)
+    table_name = source_config.get('table_name') or 'document_chunk'
+    collection_field = source_config.get('collection_field') or 'collection_name'
+    content_field = source_config.get('content_field') or 'text'
+    vector_field = source_config.get('vector_field') or 'vector'
+    metadata_field = source_config.get('metadata_field') or 'vmetadata'
+    document_id_field = source_config.get('document_id_field') or 'id'
+
+    vector = await embedding_function(query, prefix=RAG_EMBEDDING_QUERY_PREFIX)
+
+    def _search():
+        from psycopg import sql
+
+        table_identifier = sql.SQL('.').join(
+            sql.Identifier(_safe_identifier(part, 'table name')) for part in table_name.split('.')
+        )
+        collection_identifier = sql.Identifier(_safe_identifier(collection_field, 'collection field'))
+        content_identifier = sql.Identifier(_safe_identifier(content_field, 'content field'))
+        vector_identifier = sql.Identifier(_safe_identifier(vector_field, 'vector field'))
+        document_id_identifier = sql.Identifier(_safe_identifier(document_id_field, 'document id field'))
+        metadata_sql = (
+            sql.Identifier(_safe_identifier(metadata_field, 'metadata field'))
+            if metadata_field
+            else sql.SQL("'{}'::jsonb")
+        )
+
+        with psycopg.connect(
+            connection.get('endpoint'),
+            row_factory=dict_row,
+            connect_timeout=config.get('timeout') or 30,
+        ) as conn:
+            register_vector(conn)
+            with conn.cursor() as cur:
+                cur.execute(
+                    sql.SQL(
+                        """
+                    SELECT {document_id} AS id,
+                           {content} AS content,
+                           {metadata} AS metadata,
+                           {vector_column} <=> %s AS distance
+                    FROM {table_name}
+                    WHERE {collection} = %s
+                    ORDER BY distance ASC
+                    LIMIT %s
+                    """
+                    ).format(
+                        document_id=document_id_identifier,
+                        content=content_identifier,
+                        metadata=metadata_sql,
+                        vector_column=vector_identifier,
+                        table_name=table_identifier,
+                        collection=collection_identifier,
+                    ),
+                    (vector, collection_name, count),
+                )
+                return cur.fetchall()
+
+    rows = await asyncio.to_thread(_search)
+    mapping = {
+        'content_field': 'content',
+        'metadata_field': 'metadata',
+        'document_id_field': 'id',
+        'score_field': 'distance',
+    }
+    return [_normalize_result(row, mapping, knowledge, distance=row.get('distance')) for row in rows]
+
+
+async def retrieve_external_knowledge(
+    request,
+    knowledge: KnowledgeModel,
+    queries: list[str],
+    count: int,
+    user=None,
+) -> dict:
+    external = (knowledge.meta or {}).get('external', {})
+    connection_id = external.get('connection_id')
+    if not connection_id:
+        raise RuntimeError('External knowledge connection is not configured')
+
+    connection = await _get_external_connection(connection_id)
+    if not connection:
+        raise RuntimeError('External knowledge connection not found')
+
+    return await retrieve_external_knowledge_for_connection(request, knowledge, connection, queries, count, user=user)
+
+
+async def retrieve_external_knowledge_for_connection(
+    request,
+    knowledge: KnowledgeModel,
+    connection: dict,
+    queries: list[str],
+    count: int,
+    user=None,
+) -> dict:
+    auth_config = connection.get('auth_config') or {}
+    if not connection.get('enabled', True):
+        raise RuntimeError('External knowledge connection is disabled')
+
+    started_at = time.monotonic()
+    chunks = []
+    provider = (connection.get('provider') or '').lower()
+
+    for query in queries:
+        if provider == 'qdrant':
+            chunks.extend(
+                await _retrieve_qdrant(
+                    connection,
+                    auth_config,
+                    knowledge,
+                    query,
+                    count,
+                    getattr(request.app.state, 'EMBEDDING_FUNCTION', None),
+                )
+            )
+        elif provider == 'milvus':
+            chunks.extend(
+                await _retrieve_milvus(
+                    connection,
+                    auth_config,
+                    knowledge,
+                    query,
+                    count,
+                    getattr(request.app.state, 'EMBEDDING_FUNCTION', None),
+                )
+            )
+        elif provider == 'pgvector':
+            chunks.extend(
+                await _retrieve_pgvector(
+                    connection,
+                    auth_config,
+                    knowledge,
+                    query,
+                    count,
+                    getattr(request.app.state, 'EMBEDDING_FUNCTION', None),
+                )
+            )
+        else:
+            raise RuntimeError(f'Unsupported external knowledge provider: {connection.get("provider")}')
+
+    chunks = chunks[:count]
+    log.info(
+        'external_knowledge_retrieval knowledge_id=%s connection_id=%s provider=%s user_id=%s latency_ms=%s result_count=%s',
+        knowledge.id,
+        connection.get('id'),
+        connection.get('provider'),
+        getattr(user, 'id', None),
+        round((time.monotonic() - started_at) * 1000),
+        len(chunks),
+    )
+
+    return {
+        'documents': [[chunk['content'] for chunk in chunks]],
+        'metadatas': [[chunk['metadata'] for chunk in chunks]],
+        'distances': [[chunk['distance'] for chunk in chunks]],
+    }

backend/open_webui/retrieval/loaders/external_document.py

@@ -6,7 +6,7 @@ from urllib.parse import quote
 import requests
 from langchain_core.document_loaders import BaseLoader
 from langchain_core.documents import Document
-from open_webui.utils.headers import include_user_info_headers
+from open_webui.utils.headers import include_user_info_headers, parse_custom_headers
 
 log = logging.getLogger(__name__)
 
@@ -19,6 +19,9 @@ class ExternalDocumentLoader(BaseLoader):
         api_key: str,
         mime_type=None,
         user=None,
+        user_groups=None,
+        headers=None,
+        metadata=None,
         **kwargs,
     ) -> None:
         self.url = url
@@ -28,6 +31,9 @@ class ExternalDocumentLoader(BaseLoader):
         self.mime_type = mime_type
 
         self.user = user
+        self.user_groups = user_groups
+        self.headers = headers
+        self.metadata = metadata
 
     def load(self) -> List[Document]:
         with open(self.file_path, 'rb') as f:
@@ -45,6 +51,8 @@ class ExternalDocumentLoader(BaseLoader):
         except Exception:
             pass
 
+        headers.update(parse_custom_headers(self.headers, self.user, self.metadata, user_groups=self.user_groups))
+
         if self.user is not None:
             headers = include_user_info_headers(headers, self.user)
 

backend/open_webui/retrieval/loaders/main.py

@@ -11,18 +11,23 @@ from langchain_community.document_loaders import (
     BSHTMLLoader,
     CSVLoader,
     Docx2txtLoader,
-    OutlookMessageLoader,
     PyPDFLoader,
     TextLoader,
     YoutubeLoader,
 )
 from langchain_core.documents import Document
-from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, GLOBAL_LOG_LEVEL, REQUESTS_VERIFY
+from open_webui.env import (
+    AIOHTTP_CLIENT_SESSION_SSL,
+    GLOBAL_LOG_LEVEL,
+    MINERU_MAX_MARKDOWN_BYTES,
+    REQUESTS_VERIFY,
+)
 from open_webui.retrieval.loaders.datalab_marker import DatalabMarkerLoader
 from open_webui.retrieval.loaders.external_document import ExternalDocumentLoader
 from open_webui.retrieval.loaders.mineru import MinerULoader
 from open_webui.retrieval.loaders.mistral import MistralLoader
-from open_webui.retrieval.loaders.paddleocr_vl import PaddleOCRVLLoader
+from open_webui.retrieval.loaders.paddleocr_vl import PADDLEOCR_VL_SUPPORTED_EXTENSIONS, PaddleOCRVLLoader
+from open_webui.utils.headers import get_user_groups_for_custom_headers
 
 logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
 log = logging.getLogger(__name__)
@@ -183,6 +188,7 @@ class DoclingLoader:
         self.params = params or {}
 
     def load(self) -> list[Document]:
+        page_break_marker = '\f'
         with open(self.file_path, 'rb') as f:
             headers = {}
             if self.api_key:
@@ -199,6 +205,10 @@ class DoclingLoader:
                 },
                 data={
                     'image_export_mode': 'placeholder',
+                    'md_page_break_placeholder': page_break_marker,
+                    # Keep Docling params as user-provided form values. Encoding nested
+                    # values here would make Open WebUI responsible for Docling's API
+                    # quirks and could break when Docling changes its form contract.
                     **self.params,
                 },
                 headers=headers,
@@ -207,9 +217,19 @@ class DoclingLoader:
         if r.ok:
             result = r.json()
             document_data = result.get('document', {})
-            text = document_data.get('md_content', '<No text content found>')
+            md_content = document_data.get('md_content', '')
+            text = md_content or '<No text content found>'
 
             metadata = {'Content-Type': self.mime_type} if self.mime_type else {}
+            if page_break_marker in md_content:
+                documents = [
+                    Document(page_content=page.strip(), metadata={**metadata, 'page': page_idx})
+                    for page_idx, page in enumerate(md_content.split(page_break_marker))
+                    if page.strip()
+                ]
+                if documents:
+                    log.debug('Docling extracted text: %s', text)
+                    return documents
 
             log.debug('Docling extracted text: %s', text)
             return [Document(page_content=text, metadata=metadata)]
@@ -229,6 +249,8 @@ class Loader:
     def __init__(self, engine: str = '', **kwargs):
         self.engine = engine
         self.user = kwargs.get('user', None)
+        self.user_groups = kwargs.get('user_groups', None)
+        self.metadata = kwargs.get('metadata', {})
         self.kwargs = kwargs
 
     def load(self, filename: str, file_content_type: str, file_path: str) -> list[Document]:
@@ -246,6 +268,13 @@ class Loader:
         loop for the entire parse — minutes for large PDFs. This offloads
         the work to a worker thread so the loop stays responsive.
         """
+        # Group lookup is async-only, so it must happen before `load`
+        # is offloaded to a thread without a running event loop.
+        if self.engine == 'external' and self.user_groups is None:
+            self.user_groups = await get_user_groups_for_custom_headers(
+                self.kwargs.get('EXTERNAL_DOCUMENT_LOADER_HEADERS'), self.user
+            )
+
         return await asyncio.to_thread(self.load, filename, file_content_type, file_path)
 
     def _is_text_file(self, file_ext: str, file_content_type: str) -> bool:
@@ -285,13 +314,20 @@ class Loader:
         try:
             raw.decode('utf-8')
             return 'utf-8'
-        except UnicodeDecodeError:
-            pass
+        except UnicodeDecodeError as e:
+            first_non_utf8 = e.start
 
         # Use chardet as a hint, not as ground truth
         import chardet
 
-        detected = chardet.detect(raw)
+        # chardet is pure Python (~1.3s/MB), so sample around the first bad byte
+        window = 256 * 1024
+        sample_start = max(0, first_non_utf8 - window // 2)
+        sample = raw[sample_start : sample_start + window]
+        detected = chardet.detect(sample)
+        # A stray byte can sit far from the real payload, leaving the sample with nothing to read
+        if len(sample.translate(None, delete=bytes(range(128)))) < 64 and len(sample) < len(raw):
+            detected = chardet.detect(raw)
         detected_enc = (detected.get('encoding') or '').lower().replace('-', '').replace('_', '')
 
         # Map chardet's detected encoding to the correct superset codec.
@@ -404,6 +440,13 @@ class Loader:
                 api_key=self.kwargs.get('EXTERNAL_DOCUMENT_LOADER_API_KEY'),
                 mime_type=file_content_type,
                 user=self.user,
+                user_groups=self.user_groups,
+                headers=self.kwargs.get('EXTERNAL_DOCUMENT_LOADER_HEADERS'),
+                metadata={
+                    **self.metadata,
+                    'file_name': filename,
+                    'file_content_type': file_content_type,
+                },
             )
         elif self.engine == 'tika' and self.kwargs.get('TIKA_SERVER_URL'):
             if self._is_text_file(file_ext, file_content_type):
@@ -511,7 +554,6 @@ class Loader:
                     mineru_timeout = int(mineru_timeout)
                 except ValueError:
                     mineru_timeout = 300
-
             loader = MinerULoader(
                 file_path=file_path,
                 api_mode=self.kwargs.get('MINERU_API_MODE', 'local'),
@@ -519,6 +561,7 @@ class Loader:
                 api_key=self.kwargs.get('MINERU_API_KEY', ''),
                 params=self.kwargs.get('MINERU_PARAMS', {}),
                 timeout=mineru_timeout,
+                max_markdown_bytes=MINERU_MAX_MARKDOWN_BYTES,
             )
         elif (
             self.engine == 'mistral_ocr'
@@ -529,8 +572,15 @@ class Loader:
                 base_url=self.kwargs.get('MISTRAL_OCR_API_BASE_URL'),
                 api_key=self.kwargs.get('MISTRAL_OCR_API_KEY'),
                 file_path=file_path,
+                use_base64=self.kwargs.get('MISTRAL_OCR_USE_BASE64', False),
+                user=self.user,
             )
-        elif self.engine == 'paddleocr_vl' and self.kwargs.get('PADDLEOCR_VL_TOKEN') != '':
+        elif (
+            self.engine == 'paddleocr_vl'
+            and self.kwargs.get('PADDLEOCR_VL_BASE_URL')
+            and self.kwargs.get('PADDLEOCR_VL_TOKEN')
+            and file_ext in PADDLEOCR_VL_SUPPORTED_EXTENSIONS
+        ):
             loader = PaddleOCRVLLoader(
                 api_url=self.kwargs.get('PADDLEOCR_VL_BASE_URL'),
                 token=self.kwargs.get('PADDLEOCR_VL_TOKEN'),
@@ -629,7 +679,18 @@ class Loader:
                     )
                     loader = PptxLoader(file_path)
             elif file_ext == 'msg':
-                loader = OutlookMessageLoader(file_path)
+                try:
+                    from langchain_community.document_loaders import (
+                        UnstructuredEmailLoader,
+                    )
+
+                    # unstructured parses .msg via python-oxmsg; avoids extract_msg's beautifulsoup4<4.14 conflict
+                    loader = UnstructuredEmailLoader(file_path, process_attachments=False)
+                except ImportError:
+                    raise ValueError(
+                        "Processing .msg files requires the 'unstructured' package. "
+                        'Install it with: pip install unstructured'
+                    )
             elif file_ext == 'odt':
                 try:
                     from langchain_community.document_loaders import UnstructuredODTLoader

backend/open_webui/retrieval/loaders/microsoft_web_iq.py

@@ -0,0 +1,110 @@
+import logging
+import time
+from collections.abc import Iterator
+from typing import Any
+from urllib.parse import urlparse
+
+import requests
+from langchain_core.document_loaders import BaseLoader
+from langchain_core.documents import Document
+
+log = logging.getLogger(__name__)
+
+DEFAULT_MICROSOFT_WEB_IQ_API_BASE_URL = 'https://api.microsoft.ai/v3'
+MICROSOFT_BROWSE_RETRY_STATUS_CODES = {202, 429, 500, 502, 503, 504}
+MICROSOFT_BROWSE_MAX_RETRIES = 2
+
+
+class MicrosoftWebIQLoader(BaseLoader):
+    def __init__(
+        self,
+        urls: str | list[str],
+        api_base_url: str,
+        api_key: str,
+        language: str = 'en',
+        verify_ssl: bool = True,
+        timeout: Any = None,
+        continue_on_failure: bool = True,
+    ) -> None:
+        self.urls = urls if isinstance(urls, list) else [urls]
+        self.api_base_url = (api_base_url or DEFAULT_MICROSOFT_WEB_IQ_API_BASE_URL).rstrip('/')
+        self.api_key = api_key
+        self.language = language
+        self.verify_ssl = verify_ssl
+        self.timeout = timeout
+        self.continue_on_failure = continue_on_failure
+
+    def lazy_load(self) -> Iterator[Document]:
+        for url in self.urls:
+            try:
+                doc = self._browse_url(url)
+                if doc is not None:
+                    yield doc
+            except Exception as e:
+                if self.continue_on_failure:
+                    log.warning(f'Error browsing {url} with Microsoft Web IQ: {e}')
+                else:
+                    raise e
+
+    def _browse_url(self, url: str) -> Document | None:
+        headers = {
+            'host': urlparse(self.api_base_url).netloc or 'api.microsoft.ai',
+            'x-apikey': self.api_key,
+            'content-type': 'application/json',
+        }
+        payload = {
+            'url': url,
+            'contentFormat': 'markdown',
+            'liveCrawl': 'fallback',
+            'renderDynamicPages': True,
+            'language': self.language,
+        }
+        try:
+            request_timeout = float(self.timeout)
+        except (TypeError, ValueError):
+            request_timeout = 60
+        request_timeout = request_timeout if request_timeout > 0 else 60
+
+        data: dict[str, Any] = {}
+        for attempt in range(MICROSOFT_BROWSE_MAX_RETRIES + 1):
+            response = requests.post(
+                f'{self.api_base_url}/browse',
+                json=payload,
+                headers=headers,
+                timeout=request_timeout,
+                verify=self.verify_ssl,
+            )
+
+            if response.status_code in MICROSOFT_BROWSE_RETRY_STATUS_CODES and attempt < MICROSOFT_BROWSE_MAX_RETRIES:
+                try:
+                    body = response.json()
+                except Exception:
+                    body = {}
+                retry_after = body.get('retryAfter') if isinstance(body, dict) else None
+                retry_after = retry_after or response.headers.get('Retry-After')
+                try:
+                    delay = min(10.0, max(0.0, float(str(retry_after).rstrip('s'))))
+                except (TypeError, ValueError):
+                    delay = min(8.0, float(2**attempt))
+                log.warning(
+                    'Microsoft Browse %s returned HTTP %s; retrying in %.1fs',
+                    url,
+                    response.status_code,
+                    delay,
+                )
+                time.sleep(delay)
+                continue
+
+            response.raise_for_status()
+            data = response.json()
+            break
+
+        content = data.get('content') or ''
+        if not isinstance(content, str) or not content.strip():
+            return None
+
+        metadata = {'source': data.get('url') or url}
+        if data.get('title'):
+            metadata['title'] = data['title']
+
+        return Document(page_content=content, metadata=metadata)

backend/open_webui/retrieval/loaders/mineru.py

@@ -28,20 +28,22 @@ class MinerULoader:
         api_key: str = '',
         params: dict = None,
         timeout: Optional[int] = 300,
+        max_markdown_bytes: Optional[int] = None,
     ):
         self.file_path = file_path
         self.api_mode = api_mode.lower()
         self.api_url = api_url.rstrip('/')
         self.api_key = api_key
         self.timeout = timeout
+        self.max_markdown_bytes = max_markdown_bytes
 
         # Parse params dict with defaults
         self.params = params or {}
-        self.enable_ocr = params.get('enable_ocr', False)
-        self.enable_formula = params.get('enable_formula', True)
-        self.enable_table = params.get('enable_table', True)
-        self.language = params.get('language', 'en')
-        self.model_version = params.get('model_version', 'pipeline')
+        self.enable_ocr = self.params.get('enable_ocr', False)
+        self.enable_formula = self.params.get('enable_formula', True)
+        self.enable_table = self.params.get('enable_table', True)
+        self.language = self.params.get('language', 'en')
+        self.model_version = self.params.get('model_version', 'pipeline')
 
         self.page_ranges = self.params.pop('page_ranges', '')
 
@@ -435,67 +437,77 @@ class MinerULoader:
                 detail=f'Error downloading results: {str(e)}',
             )
 
-        # Save ZIP to temporary file and extract
+        # Save ZIP to temporary file before reading.
+        tmp_zip_path = None
+        markdown_content = None
         try:
             with tempfile.NamedTemporaryFile(delete=False, suffix='.zip') as tmp_zip:
                 tmp_zip.write(response.content)
                 tmp_zip_path = tmp_zip.name
 
-            with tempfile.TemporaryDirectory() as tmp_dir:
-                # Extract ZIP
-                with zipfile.ZipFile(tmp_zip_path, 'r') as zip_ref:
-                    zip_ref.extractall(tmp_dir)
+            with zipfile.ZipFile(tmp_zip_path, 'r') as zip_ref:
+                members = zip_ref.infolist()
+                all_files = [member.filename for member in members]
+                md_members = [member for member in members if member.filename.endswith('.md')]
+                read_errors = []
 
-                # Find markdown file - search recursively for any .md file
-                markdown_content = None
-                found_md_path = None
-
-                # First, list all files in the ZIP for debugging
-                all_files = []
-                for root, dirs, files in os.walk(tmp_dir):
-                    for file in files:
-                        full_path = os.path.join(root, file)
-                        all_files.append(full_path)
-                        # Look for any .md file
-                        if file.endswith('.md'):
-                            found_md_path = full_path
-                            log.info(f'Found markdown file at: {full_path}')
-                            try:
-                                with open(full_path, 'r', encoding='utf-8') as f:
-                                    markdown_content = f.read()
-                                if markdown_content:  # Use the first non-empty markdown file
-                                    break
-                            except Exception as e:
-                                log.warning(f'Failed to read {full_path}: {e}')
+                for member in md_members:
+                    log.info(f'Found markdown file in ZIP: {member.filename}')
+                    try:
+                        with zip_ref.open(member, 'r') as f:
+                            if self.max_markdown_bytes is None:
+                                content = f.read()
+                            else:
+                                content = f.read(self.max_markdown_bytes + 1)
+                                if len(content) > self.max_markdown_bytes:
+                                    raise HTTPException(
+                                        status.HTTP_502_BAD_GATEWAY,
+                                        detail=f'Markdown file in results ZIP is too large: {member.filename}',
+                                    )
+                            markdown_content = content.decode('utf-8')
+                    except UnicodeDecodeError as e:
+                        read_errors.append(f'{member.filename}: {e}')
+                        log.warning(f'Failed to decode {member.filename}: {e}')
+                        continue
+                    except HTTPException:
+                        raise
+                    except Exception as e:
+                        read_errors.append(f'{member.filename}: {e}')
+                        log.warning(f'Failed to read {member.filename}: {e}')
+                        continue
                     if markdown_content:
                         break
 
                 if markdown_content is None:
                     log.error(f'Available files in ZIP: {all_files}')
-                    # Try to provide more helpful error message
-                    md_files = [f for f in all_files if f.endswith('.md')]
-                    if md_files:
-                        error_msg = f"Found .md files but couldn't read them: {md_files}"
+                    if read_errors:
+                        error_msg = f"Found .md files but couldn't read them: {read_errors}"
                     else:
                         error_msg = f'No .md files found in ZIP. Available files: {all_files}'
                     raise HTTPException(
                         status.HTTP_502_BAD_GATEWAY,
                         detail=error_msg,
                     )
-
-            # Clean up temporary ZIP file
-            os.unlink(tmp_zip_path)
-
         except zipfile.BadZipFile as e:
             raise HTTPException(
                 status.HTTP_502_BAD_GATEWAY,
                 detail=f'Invalid ZIP file received: {e}',
             )
+        except HTTPException:
+            raise
         except Exception as e:
             raise HTTPException(
                 status.HTTP_500_INTERNAL_SERVER_ERROR,
                 detail=f'Error extracting ZIP: {str(e)}',
             )
+        finally:
+            if tmp_zip_path:
+                try:
+                    os.unlink(tmp_zip_path)
+                except FileNotFoundError:
+                    pass
+                except Exception as e:
+                    log.warning(f'Failed to remove temporary ZIP file {tmp_zip_path}: {e}')
 
         if not markdown_content:
             raise HTTPException(

backend/open_webui/retrieval/loaders/mistral.py

@@ -1,15 +1,17 @@
 import asyncio
+import base64
 import logging
 import os
 import sys
 import time
 from contextlib import asynccontextmanager
-from typing import Any, Dict, List
+from typing import Any, Dict, List, Optional
 
 import aiohttp
 import requests
 from langchain_core.documents import Document
-from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, GLOBAL_LOG_LEVEL
+from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, ENABLE_FORWARD_USER_INFO_HEADERS, GLOBAL_LOG_LEVEL
+from open_webui.utils.headers import include_user_info_headers
 
 logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
 log = logging.getLogger(__name__)
@@ -37,6 +39,8 @@ class MistralLoader:
         timeout: int = 300,  # 5 minutes default
         max_retries: int = 3,
         enable_debug_logging: bool = False,
+        use_base64: bool = False,
+        user: Optional[Any] = None,
     ):
         """
         Initializes the loader with enhanced features.
@@ -47,6 +51,9 @@ class MistralLoader:
             timeout: Request timeout in seconds.
             max_retries: Maximum number of retry attempts.
             enable_debug_logging: Enable detailed debug logs.
+            use_base64: Send the document as a data URL instead of uploading it first.
+            user: The requesting user, forwarded to Mistral via user-info headers
+                when ENABLE_FORWARD_USER_INFO_HEADERS is enabled.
         """
         if not api_key:
             raise ValueError('API key cannot be empty.')
@@ -59,6 +66,8 @@ class MistralLoader:
         self.timeout = timeout
         self.max_retries = max_retries
         self.debug = enable_debug_logging
+        self.use_base64 = use_base64
+        self.user = user
 
         # PERFORMANCE OPTIMIZATION: Differentiated timeouts for different operations
         # This prevents long-running OCR operations from affecting quick operations
@@ -78,6 +87,8 @@ class MistralLoader:
             'Authorization': f'Bearer {self.api_key}',
             'User-Agent': 'OpenWebUI-MistralLoader/2.0',  # Helps API provider track usage
         }
+        if self.user is not None and ENABLE_FORWARD_USER_INFO_HEADERS:
+            self.headers = include_user_info_headers(self.headers, self.user)
 
     def _debug_log(self, message: str, *args) -> None:
         """
@@ -261,33 +272,32 @@ class MistralLoader:
         url = f'{self.base_url}/files'
 
         async def upload_request():
-            # Create multipart writer for streaming upload
-            writer = aiohttp.MultipartWriter('form-data')
+            # Open inside the request so the handle stays valid for the whole
+            # streamed POST and is closed right after.
+            with open(self.file_path, 'rb') as f:
+                writer = aiohttp.MultipartWriter('form-data')
 
-            # Add purpose field
-            purpose_part = writer.append('ocr')
-            purpose_part.set_content_disposition('form-data', name='purpose')
+                # Add purpose field
+                purpose_part = writer.append('ocr')
+                purpose_part.set_content_disposition('form-data', name='purpose')
 
-            # Add file part with streaming
-            file_part = writer.append_payload(
-                aiohttp.streams.FilePayload(
-                    self.file_path,
-                    filename=self.file_name,
-                    content_type='application/pdf',
-                )
-            )
-            file_part.set_content_disposition('form-data', name='file', filename=self.file_name)
+                # Stream the file. aiohttp builds a payload from the file object;
+                # the previous aiohttp.streams.FilePayload was removed upstream
+                # (payloads live in aiohttp.payload and there is no FilePayload),
+                # so this path raised AttributeError on every async OCR upload.
+                file_part = writer.append(f, {'Content-Type': 'application/pdf'})
+                file_part.set_content_disposition('form-data', name='file', filename=self.file_name)
 
-            self._debug_log(f'Uploading file: {self.file_name} ({self.file_size:,} bytes)')
+                self._debug_log(f'Uploading file: {self.file_name} ({self.file_size:,} bytes)')
 
-            async with session.post(
-                url,
-                data=writer,
-                headers=self.headers,
-                timeout=aiohttp.ClientTimeout(total=self.upload_timeout),
-                ssl=AIOHTTP_CLIENT_SESSION_SSL,
-            ) as response:
-                return await self._handle_response_async(response)
+                async with session.post(
+                    url,
+                    data=writer,
+                    headers=self.headers,
+                    timeout=aiohttp.ClientTimeout(total=self.upload_timeout),
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
+                ) as response:
+                    return await self._handle_response_async(response)
 
         response_data = await self._retry_request_async(upload_request)
 
@@ -417,6 +427,11 @@ class MistralLoader:
 
         return await self._retry_request_async(ocr_request)
 
+    def _get_file_data_url(self) -> str:
+        with open(self.file_path, 'rb') as f:
+            encoded_file = base64.b64encode(f.read()).decode('utf-8')
+        return f'data:application/pdf;base64,{encoded_file}'
+
     def _delete_file(self, file_id: str) -> None:
         """Deletes the file from Mistral storage (sync version)."""
         log.info(f'Deleting uploaded file ID: {file_id}')
@@ -566,6 +581,12 @@ class MistralLoader:
         start_time = time.time()
 
         try:
+            if self.use_base64:
+                documents = self._process_results(self._process_ocr(self._get_file_data_url()))
+                total_time = time.time() - start_time
+                log.info(f'Sync OCR workflow completed in {total_time:.2f}s, produced {len(documents)} documents')
+                return documents
+
             # 1. Upload file
             file_id = self._upload_file()
 
@@ -617,6 +638,13 @@ class MistralLoader:
 
         try:
             async with self._get_session() as session:
+                if self.use_base64:
+                    ocr_response = await self._process_ocr_async(session, self._get_file_data_url())
+                    documents = self._process_results(ocr_response)
+                    total_time = time.time() - start_time
+                    log.info(f'Async OCR workflow completed in {total_time:.2f}s, produced {len(documents)} documents')
+                    return documents
+
                 # 1. Upload file with streaming
                 file_id = await self._upload_file_async(session)
 

backend/open_webui/retrieval/loaders/paddleocr_vl.py

@@ -11,6 +11,9 @@ from open_webui.env import GLOBAL_LOG_LEVEL
 logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
 log = logging.getLogger(__name__)
 
+PADDLEOCR_VL_IMAGE_EXTENSIONS = ['png', 'jpg', 'jpeg', 'bmp', 'tiff', 'webp']
+PADDLEOCR_VL_SUPPORTED_EXTENSIONS = ['pdf'] + PADDLEOCR_VL_IMAGE_EXTENSIONS
+
 
 class PaddleOCRVLLoader:
     """Loader that uses PaddleOCR-vl API to extract text from PDF/images."""
@@ -46,8 +49,7 @@ class PaddleOCRVLLoader:
 
         # Detect fileType based on file extension
         ext = self.file_path.lower().split('.')[-1]
-        image_extensions = ['png', 'jpg', 'jpeg', 'bmp', 'tiff', 'webp']
-        file_type = 1 if ext in image_extensions else 0
+        file_type = 1 if ext in PADDLEOCR_VL_IMAGE_EXTENSIONS else 0
 
         payload = {
             'file': file_data,

backend/open_webui/retrieval/utils.py

@@ -37,17 +37,21 @@ from open_webui.env import (
 from open_webui.models.access_grants import AccessGrants
 from open_webui.models.chats import Chats
 from open_webui.models.files import Files
+from open_webui.models.folders import Folders
 from open_webui.models.knowledge import Knowledges
 from open_webui.models.notes import Notes
+from open_webui.models.config import Config
 from open_webui.models.users import UserModel
 from open_webui.retrieval.loaders.youtube import YoutubeLoader
 from open_webui.retrieval.vector.async_client import ASYNC_VECTOR_DB_CLIENT
+from open_webui.retrieval.external import retrieve_external_knowledge
 from open_webui.retrieval.vector.factory import VECTOR_DB_CLIENT
-from open_webui.retrieval.vector.main import GetResult
+from open_webui.retrieval.vector.main import GetResult, SearchResult
 from open_webui.retrieval.web.utils import get_web_loader
-from open_webui.utils.access_control.files import has_access_to_file
+from open_webui.utils.access_control.files import get_owner_accessible_folder_files, has_access_to_file
+from open_webui.utils.access_control.folders import has_folder_access
 from open_webui.utils.headers import include_user_info_headers
-from open_webui.utils.misc import get_message_list
+from open_webui.utils.misc import get_content_from_message, get_message_list
 
 log = logging.getLogger(__name__)
 
@@ -63,65 +67,99 @@ def is_youtube_url(url: str) -> bool:
     return re.match(youtube_regex, url) is not None
 
 
-def get_loader(request, url: str):
+LOADER_CONFIG_KEYS = {
+    'youtube_language': 'rag.youtube_loader_language',
+    'youtube_proxy_url': 'rag.youtube_loader_proxy_url',
+    'web_loader_ssl_verification': 'web.loader.ssl_verification',
+    'web_loader_concurrent_requests': 'web.loader.concurrent_requests',
+    'web_search_trust_env': 'web.search.trust_env',
+    'web_loader_engine': 'web.loader.engine',
+    'web_loader_timeout': 'web.loader.timeout',
+    'playwright_ws_url': 'web.loader.playwright_ws_url',
+    'playwright_timeout': 'web.loader.playwright_timeout',
+    'firecrawl_api_key': 'web.loader.firecrawl_api_key',
+    'firecrawl_api_url': 'web.loader.firecrawl_api_url',
+    'firecrawl_timeout': 'web.loader.firecrawl_timeout',
+    'tavily_api_key': 'web.search.tavily_api_key',
+    'tavily_extract_depth': 'web.search.tavily_extract_depth',
+    'microsoft_web_iq_api_base_url': 'web.search.microsoft_web_iq_api_base_url',
+    'microsoft_web_iq_api_key': 'web.search.microsoft_web_iq_api_key',
+    'microsoft_web_iq_language': 'web.search.microsoft_web_iq_language',
+    'external_web_loader_url': 'web.loader.external_web_loader_url',
+    'external_web_loader_api_key': 'web.loader.external_web_loader_api_key',
+    'CONTENT_EXTRACTION_ENGINE': 'rag.content_extraction_engine',
+    'DATALAB_MARKER_API_KEY': 'rag.datalab_marker_api_key',
+    'DATALAB_MARKER_API_BASE_URL': 'rag.datalab_marker_api_base_url',
+    'DATALAB_MARKER_ADDITIONAL_CONFIG': 'rag.datalab_marker_additional_config',
+    'DATALAB_MARKER_SKIP_CACHE': 'rag.datalab_marker_skip_cache',
+    'DATALAB_MARKER_FORCE_OCR': 'rag.datalab_marker_force_ocr',
+    'DATALAB_MARKER_PAGINATE': 'rag.datalab_marker_paginate',
+    'DATALAB_MARKER_STRIP_EXISTING_OCR': 'rag.datalab_marker_strip_existing_ocr',
+    'DATALAB_MARKER_DISABLE_IMAGE_EXTRACTION': 'rag.datalab_marker_disable_image_extraction',
+    'DATALAB_MARKER_FORMAT_LINES': 'rag.datalab_marker_format_lines',
+    'DATALAB_MARKER_USE_LLM': 'rag.datalab_marker_use_llm',
+    'DATALAB_MARKER_OUTPUT_FORMAT': 'rag.datalab_marker_output_format',
+    'EXTERNAL_DOCUMENT_LOADER_URL': 'rag.external_document_loader_url',
+    'EXTERNAL_DOCUMENT_LOADER_API_KEY': 'rag.external_document_loader_api_key',
+    'EXTERNAL_DOCUMENT_LOADER_HEADERS': 'rag.external_document_loader_headers',
+    'TIKA_SERVER_URL': 'rag.tika_server_url',
+    'DOCLING_SERVER_URL': 'rag.docling_server_url',
+    'DOCLING_API_KEY': 'rag.docling_api_key',
+    'DOCLING_PARAMS': 'rag.docling_params',
+    'PDF_EXTRACT_IMAGES': 'rag.pdf_extract_images',
+    'PDF_LOADER_MODE': 'rag.pdf_loader_mode',
+    'DOCUMENT_INTELLIGENCE_ENDPOINT': 'rag.document_intelligence_endpoint',
+    'DOCUMENT_INTELLIGENCE_KEY': 'rag.document_intelligence_key',
+    'DOCUMENT_INTELLIGENCE_MODEL': 'rag.document_intelligence_model',
+    'MISTRAL_OCR_API_BASE_URL': 'rag.mistral_ocr_api_base_url',
+    'MISTRAL_OCR_API_KEY': 'rag.mistral_ocr_api_key',
+    'MISTRAL_OCR_USE_BASE64': 'rag.mistral_ocr_use_base64',
+    'PADDLEOCR_VL_BASE_URL': 'rag.paddleocr_vl_base_url',
+    'PADDLEOCR_VL_TOKEN': 'rag.paddleocr_vl_token',
+    'MINERU_API_MODE': 'rag.mineru_api_mode',
+    'MINERU_API_URL': 'rag.mineru_api_url',
+    'MINERU_API_KEY': 'rag.mineru_api_key',
+    'MINERU_API_TIMEOUT': 'rag.mineru_api_timeout',
+    'MINERU_PARAMS': 'rag.mineru_params',
+    'MINERU_FILE_EXTENSIONS': 'rag.mineru_file_extensions',
+}
+
+
+async def get_loader_config():
+    values = await Config.get_many(*LOADER_CONFIG_KEYS.values())
+    return {name: values.get(key) for name, key in LOADER_CONFIG_KEYS.items()}
+
+
+def get_loader(request, url: str, config: dict):
     if is_youtube_url(url):
         return YoutubeLoader(
             url,
-            language=request.app.state.config.YOUTUBE_LOADER_LANGUAGE,
-            proxy_url=request.app.state.config.YOUTUBE_LOADER_PROXY_URL,
+            language=config.get('youtube_language'),
+            proxy_url=config.get('youtube_proxy_url'),
         )
-    else:
-        return get_web_loader(
-            url,
-            verify_ssl=request.app.state.config.ENABLE_WEB_LOADER_SSL_VERIFICATION,
-            requests_per_second=request.app.state.config.WEB_LOADER_CONCURRENT_REQUESTS,
-            trust_env=request.app.state.config.WEB_SEARCH_TRUST_ENV,
-        )
-
-
-def build_loader_from_config(request):
-    """Build a Loader instance with the admin's configured extraction engine settings."""
-    from open_webui.retrieval.loaders.main import Loader
-
-    config = request.app.state.config
-    return Loader(
-        engine=config.CONTENT_EXTRACTION_ENGINE,
-        DATALAB_MARKER_API_KEY=config.DATALAB_MARKER_API_KEY,
-        DATALAB_MARKER_API_BASE_URL=config.DATALAB_MARKER_API_BASE_URL,
-        DATALAB_MARKER_ADDITIONAL_CONFIG=config.DATALAB_MARKER_ADDITIONAL_CONFIG,
-        DATALAB_MARKER_SKIP_CACHE=config.DATALAB_MARKER_SKIP_CACHE,
-        DATALAB_MARKER_FORCE_OCR=config.DATALAB_MARKER_FORCE_OCR,
-        DATALAB_MARKER_PAGINATE=config.DATALAB_MARKER_PAGINATE,
-        DATALAB_MARKER_STRIP_EXISTING_OCR=config.DATALAB_MARKER_STRIP_EXISTING_OCR,
-        DATALAB_MARKER_DISABLE_IMAGE_EXTRACTION=config.DATALAB_MARKER_DISABLE_IMAGE_EXTRACTION,
-        DATALAB_MARKER_FORMAT_LINES=config.DATALAB_MARKER_FORMAT_LINES,
-        DATALAB_MARKER_USE_LLM=config.DATALAB_MARKER_USE_LLM,
-        DATALAB_MARKER_OUTPUT_FORMAT=config.DATALAB_MARKER_OUTPUT_FORMAT,
-        EXTERNAL_DOCUMENT_LOADER_URL=config.EXTERNAL_DOCUMENT_LOADER_URL,
-        EXTERNAL_DOCUMENT_LOADER_API_KEY=config.EXTERNAL_DOCUMENT_LOADER_API_KEY,
-        TIKA_SERVER_URL=config.TIKA_SERVER_URL,
-        DOCLING_SERVER_URL=config.DOCLING_SERVER_URL,
-        DOCLING_API_KEY=config.DOCLING_API_KEY,
-        DOCLING_PARAMS=config.DOCLING_PARAMS,
-        PDF_EXTRACT_IMAGES=config.PDF_EXTRACT_IMAGES,
-        PDF_LOADER_MODE=config.PDF_LOADER_MODE,
-        DOCUMENT_INTELLIGENCE_ENDPOINT=config.DOCUMENT_INTELLIGENCE_ENDPOINT,
-        DOCUMENT_INTELLIGENCE_KEY=config.DOCUMENT_INTELLIGENCE_KEY,
-        DOCUMENT_INTELLIGENCE_MODEL=config.DOCUMENT_INTELLIGENCE_MODEL,
-        MISTRAL_OCR_API_BASE_URL=config.MISTRAL_OCR_API_BASE_URL,
-        MISTRAL_OCR_API_KEY=config.MISTRAL_OCR_API_KEY,
-        PADDLEOCR_VL_BASE_URL=config.PADDLEOCR_VL_BASE_URL,
-        PADDLEOCR_VL_TOKEN=config.PADDLEOCR_VL_TOKEN,
-        MINERU_API_MODE=config.MINERU_API_MODE,
-        MINERU_API_URL=config.MINERU_API_URL,
-        MINERU_API_KEY=config.MINERU_API_KEY,
-        MINERU_API_TIMEOUT=config.MINERU_API_TIMEOUT,
-        MINERU_PARAMS=config.MINERU_PARAMS,
-        MINERU_FILE_EXTENSIONS=config.MINERU_FILE_EXTENSIONS,
+    return get_web_loader(
+        url,
+        verify_ssl=config.get('web_loader_ssl_verification'),
+        requests_per_second=config.get('web_loader_concurrent_requests'),
+        trust_env=config.get('web_search_trust_env'),
+        loader_config=config,
     )
 
 
-def _extract_text_from_binary_response(request, response: requests.Response, url: str) -> tuple[str, list]:
+def build_loader_from_config(request, config: dict):
+    """Build a Loader instance with the admin's configured extraction engine settings."""
+    from open_webui.retrieval.loaders.main import Loader
+
+    loader_config = {key: config.get(key) for key in LOADER_CONFIG_KEYS if key.isupper()}
+    return Loader(
+        engine=loader_config['CONTENT_EXTRACTION_ENGINE'],
+        **{key: value for key, value in loader_config.items() if key != 'CONTENT_EXTRACTION_ENGINE'},
+    )
+
+
+def _extract_text_from_binary_response(
+    request, response: requests.Response, url: str, loader_config: dict
+) -> tuple[str, list]:
     """Download response body to a temp file and extract text using the Loader pipeline."""
     import mimetypes
     import tempfile
@@ -150,7 +188,7 @@ def _extract_text_from_binary_response(request, response: requests.Response, url
         tmp_path = tmp.name
 
     try:
-        loader = build_loader_from_config(request)
+        loader = build_loader_from_config(request, loader_config)
         docs = loader.load(filename, content_type, tmp_path)
         for doc in docs:
             doc.metadata['source'] = url
@@ -170,8 +208,17 @@ def _is_text_content_type(content_type: str) -> bool:
     return not ct  # empty / missing → assume HTML
 
 
-def get_content_from_url(request, url: str) -> str:
-    from open_webui.retrieval.web.utils import validate_url
+async def get_content_from_url(request, url: str) -> str:
+    loader_config = await get_loader_config()
+
+    # The rest of this function performs synchronous, blocking work: an SSRF-guarded
+    # `requests` probe and a synchronous document loader (`loader.load()`). Run it in a
+    # worker thread so the event loop stays free while waiting on network/parsing.
+    return await asyncio.to_thread(_get_content_from_url_sync, request, url, loader_config)
+
+
+def _get_content_from_url_sync(request, url: str, loader_config):
+    from open_webui.retrieval.web.utils import validate_url, _SSRFSafeAdapter
 
     # Validate URL before making any request (blocks private IPs, non-HTTP, filter list)
     validate_url(url)
@@ -183,7 +230,7 @@ def get_content_from_url(request, url: str) -> str:
     # when allow_redirects=False, causing the binary-content path to run
     # and produce empty docs → HTTP 400.
     if is_youtube_url(url):
-        loader = get_loader(request, url)
+        loader = get_loader(request, url, loader_config)
         docs = loader.load()
         content = ' '.join([doc.page_content for doc in docs])
         return content, docs
@@ -194,7 +241,11 @@ def get_content_from_url(request, url: str) -> str:
     # re-validation would let an attacker reach private IPs (RFC1918, loopback,
     # cloud-metadata 169.254.169.254) via a public host that redirects internally.
     try:
-        response = requests.get(url, stream=True, timeout=30, allow_redirects=AIOHTTP_CLIENT_ALLOW_REDIRECTS)
+        # Probe through the connect-time SSRF guard; bare requests.get re-resolves (DNS-rebinding gap).
+        session = requests.Session()
+        session.mount('http://', _SSRFSafeAdapter())
+        session.mount('https://', _SSRFSafeAdapter())
+        response = session.get(url, stream=True, timeout=30, allow_redirects=AIOHTTP_CLIENT_ALLOW_REDIRECTS)
         response.raise_for_status()
         content_type = response.headers.get('Content-Type', '')
     except Exception:
@@ -205,14 +256,14 @@ def get_content_from_url(request, url: str) -> str:
     if response is None or _is_text_content_type(content_type):
         if response is not None:
             response.close()
-        loader = get_loader(request, url)
+        loader = get_loader(request, url, loader_config)
         docs = loader.load()
         content = ' '.join([doc.page_content for doc in docs])
         return content, docs
 
     # Binary content (PDF, DOCX, XLSX, PPTX, etc.) — download and extract
     try:
-        return _extract_text_from_binary_response(request, response, url)
+        return _extract_text_from_binary_response(request, response, url, loader_config)
     finally:
         response.close()
 
@@ -255,21 +306,7 @@ class VectorSearchRetriever(BaseRetriever):
             limit=self.top_k,
         )
 
-        ids = result.ids[0]
-        metadatas = result.metadatas[0]
-        documents = result.documents[0]
-
-        results = []
-        for idx in range(len(ids)):
-            metadata = metadatas[idx]
-            metadata[CHUNK_HASH_KEY] = _content_hash(documents[idx])
-            results.append(
-                Document(
-                    metadata=metadata,
-                    page_content=documents[idx],
-                )
-            )
-        return results
+        return _search_result_to_documents(result)
 
 
 def query_doc(collection_name: str, query_embedding: list[float], k: int, user: UserModel = None):
@@ -338,9 +375,96 @@ def get_enriched_texts(collection_result: GetResult) -> list[str]:
     return enriched_texts
 
 
+def _search_result_to_documents(result: SearchResult | None) -> list[Document]:
+    ids = result.ids[0] if result and result.ids else []
+    metadatas = result.metadatas[0] if result and result.metadatas else []
+    documents = result.documents[0] if result and result.documents else []
+    distances = result.distances[0] if result and result.distances else []
+
+    docs = []
+    for idx in range(len(ids)):
+        document = documents[idx]
+        metadata = dict(metadatas[idx] or {})
+        metadata[CHUNK_HASH_KEY] = _content_hash(document)
+        if idx < len(distances):
+            metadata.setdefault('score', distances[idx])
+        docs.append(Document(metadata=metadata, page_content=document))
+    return docs
+
+
+def _supports_native_hybrid_search() -> bool:
+    supports_hybrid_search = getattr(ASYNC_VECTOR_DB_CLIENT, 'supports_hybrid_search', None)
+    if supports_hybrid_search is not None:
+        return bool(supports_hybrid_search)
+    return callable(getattr(ASYNC_VECTOR_DB_CLIENT, 'hybrid_search', None))
+
+
+async def query_doc_with_native_hybrid_search(
+    collection_name: str,
+    query: str,
+    embedding_function,
+    k: int,
+    reranking_function,
+    k_reranker: int,
+    r: float,
+    hybrid_bm25_weight: float,
+) -> Optional[dict]:
+    try:
+        if not _supports_native_hybrid_search():
+            return None
+
+        query_vectors = []
+        if hybrid_bm25_weight < 1:
+            query_vectors = [await embedding_function(query, RAG_EMBEDDING_QUERY_PREFIX)]
+
+        result = await ASYNC_VECTOR_DB_CLIENT.hybrid_search(
+            collection_name=collection_name,
+            query=query,
+            vectors=query_vectors,
+            limit=k,
+            hybrid_bm25_weight=hybrid_bm25_weight,
+        )
+        if result is None:
+            return None
+
+        documents = _search_result_to_documents(result)
+        if not documents:
+            return {'distances': [[]], 'documents': [[]], 'metadatas': [[]]}
+
+        compressor = RerankCompressor(
+            embedding_function=embedding_function,
+            top_n=k_reranker,
+            reranking_function=reranking_function,
+            r_score=r,
+        )
+        compressed = await compressor.acompress_documents(documents, query)
+
+        distances = [d.metadata.get('score') for d in compressed]
+        documents = [d.page_content for d in compressed]
+        metadatas = [d.metadata for d in compressed]
+
+        if k < k_reranker:
+            sorted_items = sorted(zip(distances, documents, metadatas), key=lambda x: x[0], reverse=True)
+            sorted_items = sorted_items[:k]
+
+            if sorted_items:
+                distances, documents, metadatas = map(list, zip(*sorted_items))
+            else:
+                distances, documents, metadatas = [], [], []
+
+        return {
+            'distances': [distances],
+            'documents': [documents],
+            'metadatas': [metadatas],
+        }
+    except Exception as e:
+        log.debug(f'Native hybrid search failed for {collection_name}, falling back to legacy hybrid search: {e}')
+        return None
+
+
 async def query_doc_with_hybrid_search(
     collection_name: str,
-    collection_result: GetResult,
+    collection_result: Optional[GetResult],
     query: str,
     embedding_function,
     k: int,
@@ -349,8 +473,26 @@ async def query_doc_with_hybrid_search(
     r: float,
     hybrid_bm25_weight: float,
     enable_enriched_texts: bool = False,
+    native_hybrid_search: bool = True,
 ) -> dict:
     try:
+        if native_hybrid_search and not enable_enriched_texts:
+            native_result = await query_doc_with_native_hybrid_search(
+                collection_name=collection_name,
+                query=query,
+                embedding_function=embedding_function,
+                k=k,
+                reranking_function=reranking_function,
+                k_reranker=k_reranker,
+                r=r,
+                hybrid_bm25_weight=hybrid_bm25_weight,
+            )
+            if native_result is not None:
+                return native_result
+
+        if collection_result is None:
+            collection_result = await ASYNC_VECTOR_DB_CLIENT.get(collection_name=collection_name)
+
         # First check if collection_result has the required attributes
         if (
             not collection_result
@@ -490,7 +632,7 @@ def merge_and_sort_query_results(query_results: list[dict], k: int) -> dict:
 
         for distance, document, metadata in zip(distances, documents, metadatas):
             if isinstance(document, str):
-                doc_hash = hashlib.sha256(document.encode()).hexdigest()  # Compute a hash for uniqueness
+                doc_hash = (metadata or {}).get(CHUNK_HASH_KEY) or _content_hash(document)
 
                 if doc_hash not in combined.keys():
                     combined[doc_hash] = (distance, document, metadata)
@@ -539,8 +681,15 @@ async def query_collection(
     embedding_function,
     k: int,
 ) -> dict:
+    config = await Config.get_many(
+        'rag.enable_hybrid_search',
+        'rag.top_k_reranker',
+        'rag.relevance_threshold',
+        'rag.hybrid_bm25_weight',
+        'rag.enable_hybrid_search_enriched_texts',
+    )
     # When request is provided, try hybrid search + reranking if enabled
-    if request and request.app.state.config.ENABLE_RAG_HYBRID_SEARCH:
+    if request and config.get('rag.enable_hybrid_search'):
         try:
             reranking_function = (
                 (lambda query, documents: request.app.state.RERANKING_FUNCTION(query, documents))
@@ -553,10 +702,10 @@ async def query_collection(
                 embedding_function=embedding_function,
                 k=k,
                 reranking_function=reranking_function,
-                k_reranker=request.app.state.config.TOP_K_RERANKER,
-                r=request.app.state.config.RELEVANCE_THRESHOLD,
-                hybrid_bm25_weight=request.app.state.config.HYBRID_BM25_WEIGHT,
-                enable_enriched_texts=request.app.state.config.ENABLE_RAG_HYBRID_SEARCH_ENRICHED_TEXTS,
+                k_reranker=config.get('rag.top_k_reranker'),
+                r=config.get('rag.relevance_threshold'),
+                hybrid_bm25_weight=config.get('rag.hybrid_bm25_weight'),
+                enable_enriched_texts=config.get('rag.enable_hybrid_search_enriched_texts'),
             )
         except Exception as e:
             log.debug(f'Hybrid search failed, falling back to vector search: {e}')
@@ -623,6 +772,28 @@ async def query_collection_with_hybrid_search(
 ) -> dict:
     results = []
     error = False
+
+    if not enable_enriched_texts:
+
+        async def process_native_query(collection_name, query):
+            result = await query_doc_with_native_hybrid_search(
+                collection_name=collection_name,
+                query=query,
+                embedding_function=embedding_function,
+                k=k,
+                reranking_function=reranking_function,
+                k_reranker=k_reranker,
+                r=r,
+                hybrid_bm25_weight=hybrid_bm25_weight,
+            )
+            return result
+
+        native_task_results = await asyncio.gather(
+            *[process_native_query(collection_name, query) for collection_name in collection_names for query in queries]
+        )
+        if native_task_results and all(result is not None for result in native_task_results):
+            return merge_and_sort_query_results(native_task_results, k=k)
+
     # Fetch every collection's contents once up front so the
     # per-query/per-document loop below can reuse them. Each fetch
     # offloads to a worker thread, so run them concurrently with
@@ -657,6 +828,7 @@ async def query_collection_with_hybrid_search(
                 r=r,
                 hybrid_bm25_weight=hybrid_bm25_weight,
                 enable_enriched_texts=enable_enriched_texts,
+                native_hybrid_search=False,
             )
             return result, None
         except Exception as e:
@@ -927,15 +1099,15 @@ def get_embedding_function(
     concurrent_requests=0,
 ) -> Awaitable:
     if embedding_engine == '':
-        if embedding_function is None:
-            raise ValueError(
-                'No embedding model is loaded. Set RAG_EMBEDDING_MODEL to a valid '
-                'SentenceTransformer model name, or configure an external '
-                'RAG_EMBEDDING_ENGINE (ollama, openai, azure_openai).'
-            )
-
         # Sentence transformers: CPU-bound sync operation
         async def async_embedding_function(query, prefix=None, user=None):
+            # Deferred so a missing local model degrades RAG instead of crashing boot.
+            if embedding_function is None:
+                raise ValueError(
+                    'No embedding model is loaded. Set RAG_EMBEDDING_MODEL to a valid '
+                    'SentenceTransformer model name, or configure an external '
+                    'RAG_EMBEDDING_ENGINE (ollama, openai, azure_openai).'
+                )
             return await asyncio.to_thread(
                 (
                     lambda query, prefix=None: embedding_function.encode(
@@ -1095,7 +1267,7 @@ async def filter_accessible_collections(
       - any name with characters outside [A-Za-z0-9_-] → rejected
       - file-*          → validated via has_access_to_file
       - user-memory-*   → must match user's own memory collection
-      - web-search-*    → ephemeral per-query collections, always allowed
+      - web-search-*    → ephemeral per-query collections, owner-bound to web-search-{user.id}-*
       - knowledge-bases → always denied (system meta-collection)
       - everything else → if the name matches a knowledge base, validated
                           via Knowledges.check_access_by_user_id; if no
@@ -1130,10 +1302,10 @@ async def filter_accessible_collections(
             if name == f'user-memory-{user.id}':
                 validated.add(name)
         elif name.startswith('web-search-'):
-            # Ephemeral collections created by process_web_search — safe
-            # to allow because they contain only transient web-search
-            # results scoped to the requesting user's session.
-            validated.add(name)
+            # Ephemeral per-query collections, owner-bound: process_web_search mints
+            # them as web-search-{user.id}-<hash>, so only the creator may read/write.
+            if name.startswith(f'web-search-{user.id}-'):
+                validated.add(name)
         else:
             # May be a knowledge-base ID or a legacy/ephemeral collection.
             # If it IS a KB, enforce access control.  If no such KB
@@ -1163,10 +1335,28 @@ async def get_sources_from_items(
     full_context=False,
     user: UserModel | None = None,
 ):
-    log.debug(f'items: {items} {queries} {embedding_function} {reranking_function} {full_context}')
+    log.debug('items: %s %s %s %s %s', items, queries, embedding_function, reranking_function, full_context)
 
+    bypass_embedding_and_retrieval = await Config.get('rag.bypass_embedding_and_retrieval')
     extracted_collections = []
     query_results = []
+    folder_items = set()
+    expanded_folders = set()
+
+    items = list(items)
+    for item in items:
+        if item.get('type') != 'folder' or not user:
+            continue
+        folder_id = item.get('id')
+        if not folder_id or folder_id in expanded_folders:
+            continue
+        expanded_folders.add(folder_id)
+
+        folder = await Folders.get_folder_by_id(folder_id)
+        if folder and (user.role == 'admin' or await has_folder_access(user.id, folder, 'read', db=None)):
+            files = await get_owner_accessible_folder_files(folder)
+            folder_items.update((entry.get('type'), entry.get('id')) for entry in files if isinstance(entry, dict))
+            items.extend(files)
 
     for item in items:
         query_result = None
@@ -1234,7 +1424,10 @@ async def get_sources_from_items(
                     # Reconstruct the message list in order
                     message_list = get_message_list(messages_map, message_id)
                     message_history = '\n'.join(
-                        [f'#### {m.get("role", "user").capitalize()}\n{m.get("content")}\n' for m in message_list]
+                        [
+                            f'#### {m.get("role", "user").capitalize()}\n{get_content_from_message(m) or ""}\n'
+                            for m in message_list
+                        ]
                     )
 
                     # User has access to the chat
@@ -1244,14 +1437,14 @@ async def get_sources_from_items(
                     }
 
         elif item.get('type') == 'url':
-            content, docs = get_content_from_url(request, item.get('url'))
+            content, docs = await get_content_from_url(request, item.get('url'))
             if docs:
                 query_result = {
                     'documents': [[content]],
                     'metadatas': [[{'url': item.get('url'), 'name': item.get('url')}]],
                 }
         elif item.get('type') == 'file':
-            if item.get('context') == 'full' or request.app.state.config.BYPASS_EMBEDDING_AND_RETRIEVAL:
+            if item.get('context') == 'full' or bypass_embedding_and_retrieval:
                 if item.get('file', {}).get('data', {}).get('content', ''):
                     # Manual Full Mode Toggle
                     # Used from chat file modal, we can assume that the file content will be available from item.get("file").get("data", {}).get("content")
@@ -1273,6 +1466,7 @@ async def get_sources_from_items(
                         user.role == 'admin'
                         or file_object.user_id == user.id
                         or await has_access_to_file(item.get('id'), 'read', user)
+                        or ('file', item.get('id')) in folder_items
                     ):
                         query_result = {
                             'documents': [[file_object.data.get('content', '')]],
@@ -1303,6 +1497,7 @@ async def get_sources_from_items(
                             user.role == 'admin'
                             or file_object.user_id == user.id
                             or await has_access_to_file(file_id, 'read', user)
+                            or ('file', file_id) in folder_items
                         ):
                             if item.get('legacy'):
                                 collection_names.append(f'{file_id}')
@@ -1322,51 +1517,64 @@ async def get_sources_from_items(
                     resource_id=knowledge_base.id,
                     permission='read',
                 )
+                or ('collection', item.get('id')) in folder_items
             ):
-                if item.get('context') == 'full' or request.app.state.config.BYPASS_EMBEDDING_AND_RETRIEVAL:
-                    if knowledge_base and (
-                        user.role == 'admin'
-                        or knowledge_base.user_id == user.id
-                        or await AccessGrants.has_access(
-                            user_id=user.id,
-                            resource_type='knowledge',
-                            resource_id=knowledge_base.id,
-                            permission='read',
-                        )
-                    ):
-                        files = await Knowledges.get_files_by_id(knowledge_base.id)
+                if (knowledge_base.meta or {}).get('source') == 'external':
+                    query_result = await retrieve_external_knowledge(
+                        request,
+                        knowledge_base,
+                        queries=queries,
+                        count=k,
+                        user=user,
+                    )
+                    extracted_collections.append(knowledge_base.id)
 
-                        documents = []
-                        metadatas = []
-                        for file in files:
-                            documents.append(file.data.get('content', ''))
-                            metadatas.append(
-                                {
-                                    'file_id': file.id,
-                                    'name': file.filename,
-                                    'source': file.filename,
-                                }
-                            )
-
-                        query_result = {
-                            'documents': [documents],
-                            'metadatas': [metadatas],
-                        }
                 else:
-                    if item.get('legacy'):
-                        if BYPASS_RETRIEVAL_ACCESS_CONTROL:
-                            collection_names = item.get('collection_names', [])
-                        else:
-                            # Legacy KB: item.collection_names is client-supplied.
-                            # Validate against the KB's actual files to prevent
-                            # cross-tenant collection name substitution.
+                    if item.get('context') == 'full' or bypass_embedding_and_retrieval:
+                        if knowledge_base and (
+                            user.role == 'admin'
+                            or knowledge_base.user_id == user.id
+                            or await AccessGrants.has_access(
+                                user_id=user.id,
+                                resource_type='knowledge',
+                                resource_id=knowledge_base.id,
+                                permission='read',
+                            )
+                            or ('collection', item.get('id')) in folder_items
+                        ):
                             files = await Knowledges.get_files_by_id(knowledge_base.id)
-                            owned_names = {f'file-{f.id}' for f in files}
-                            owned_names.add(knowledge_base.id)
-                            valid_names = [n for n in (item.get('collection_names') or []) if n in owned_names]
-                            collection_names = valid_names if valid_names else [knowledge_base.id]
+
+                            documents = []
+                            metadatas = []
+                            for file in files:
+                                documents.append(file.data.get('content', ''))
+                                metadatas.append(
+                                    {
+                                        'file_id': file.id,
+                                        'name': file.filename,
+                                        'source': file.filename,
+                                    }
+                                )
+
+                            query_result = {
+                                'documents': [documents],
+                                'metadatas': [metadatas],
+                            }
                     else:
-                        collection_names.append(item['id'])
+                        if item.get('legacy'):
+                            if BYPASS_RETRIEVAL_ACCESS_CONTROL:
+                                collection_names = item.get('collection_names', [])
+                            else:
+                                # Legacy KB: item.collection_names is client-supplied.
+                                # Validate against the KB's actual files to prevent
+                                # cross-tenant collection name substitution.
+                                files = await Knowledges.get_files_by_id(knowledge_base.id)
+                                owned_names = {f'file-{f.id}' for f in files}
+                                owned_names.add(knowledge_base.id)
+                                valid_names = [n for n in (item.get('collection_names') or []) if n in owned_names]
+                                collection_names = valid_names if valid_names else [knowledge_base.id]
+                        else:
+                            collection_names.append(item['id'])
 
         elif item.get('docs'):
             # BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL
@@ -1374,6 +1582,10 @@ async def get_sources_from_items(
                 'documents': [[doc.get('content') for doc in item.get('docs')]],
                 'metadatas': [[doc.get('metadata') for doc in item.get('docs')]],
             }
+        elif item.get('type') == 'web_search' and item.get('collection_name'):
+            # Trusted server-generated collection; authorized by
+            # filter_accessible_collections below (allowlists web-search-*).
+            collection_names.append(item['collection_name'])
         elif item.get('collection_name'):
             if BYPASS_RETRIEVAL_ACCESS_CONTROL:
                 collection_names.append(item['collection_name'])
@@ -1399,7 +1611,7 @@ async def get_sources_from_items(
                 continue
 
             # Filter out collections the user cannot read
-            if user:
+            if user and (item.get('type'), item.get('id')) not in folder_items:
                 collection_names = await filter_accessible_collections(collection_names, user)
                 if not collection_names:
                     log.debug(f'access denied for all collections in item {item}')

backend/open_webui/retrieval/vector/async_client.py

@@ -82,6 +82,10 @@ class AsyncVectorDBClient:
         (e.g. already inside a worker thread)."""
         return self._sync
 
+    @property
+    def supports_hybrid_search(self) -> bool:
+        return type(self._sync).hybrid_search is not VectorDBBase.hybrid_search
+
     async def has_collection(self, collection_name: str) -> bool:
         return await asyncio.to_thread(self._sync.has_collection, collection_name)
 
@@ -103,6 +107,25 @@ class AsyncVectorDBClient:
     ) -> Optional[SearchResult]:
         return await asyncio.to_thread(self._sync.search, collection_name, vectors, filter, limit)
 
+    async def hybrid_search(
+        self,
+        collection_name: str,
+        query: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[Dict] = None,
+        limit: int = 10,
+        hybrid_bm25_weight: float = 0.5,
+    ) -> Optional[SearchResult]:
+        return await asyncio.to_thread(
+            self._sync.hybrid_search,
+            collection_name,
+            query,
+            vectors,
+            filter,
+            limit,
+            hybrid_bm25_weight,
+        )
+
     async def query(
         self,
         collection_name: str,

backend/open_webui/retrieval/vector/dbs/chroma.py

@@ -3,6 +3,7 @@ from typing import Optional
 
 import chromadb
 from chromadb import Settings
+from chromadb.errors import NotFoundError
 from chromadb.utils.batch_utils import create_batches
 from open_webui.config import (
     CHROMA_CLIENT_AUTH_CREDENTIALS,
@@ -56,9 +57,11 @@ class ChromaClient(VectorDBBase):
             )
 
     def has_collection(self, collection_name: str) -> bool:
-        # Check if the collection exists based on the collection name.
-        collection_names = self.client.list_collections()
-        return collection_name in collection_names
+        try:
+            self.client.get_collection(name=collection_name)
+            return True
+        except NotFoundError:
+            return False
 
     def delete_collection(self, collection_name: str):
         # Delete the collection based on the collection name.

backend/open_webui/retrieval/vector/dbs/milvus.py

@@ -25,11 +25,17 @@ from open_webui.retrieval.vector.main import (
     VectorItem,
 )
 from open_webui.retrieval.vector.utils import process_metadata
-from pymilvus import Collection, DataType, FieldSchema, connections
+from pymilvus import DataType
 from pymilvus import MilvusClient as Client
+from pymilvus.exceptions import MilvusException
 
 log = logging.getLogger(__name__)
 
+# Milvus caps stored text length (here the chunk lives under the JSON `data`
+# field). Clamp long chunks before insert so one oversized chunk can't fail the
+# whole batch and leave the file with zero embeddings.
+MILVUS_TEXT_MAX_LENGTH = 65535
+
 
 class MilvusClient(VectorDBBase):
     def __init__(self):
@@ -196,8 +202,6 @@ class MilvusClient(VectorDBBase):
         return self._result_to_search_result(result)
 
     def query(self, collection_name: str, filter: dict, limit: int = -1):
-        connections.connect(uri=MILVUS_URI, token=MILVUS_TOKEN, db_name=MILVUS_DB)
-
         collection_name = collection_name.replace('-', '_')
         if not self.has_collection(collection_name):
             log.warning(f'Query attempted on non-existent collection: {self.collection_prefix}_{collection_name}')
@@ -212,16 +216,16 @@ class MilvusClient(VectorDBBase):
 
         filter_string = ' && '.join(filter_expressions)
 
-        collection = Collection(f'{self.collection_prefix}_{collection_name}')
-        collection.load()
+        self.client.load_collection(collection_name=f'{self.collection_prefix}_{collection_name}')
 
         try:
             log.info(
                 f"Querying collection {self.collection_prefix}_{collection_name} with filter: '{filter_string}', limit: {limit}"
             )
 
-            iterator = collection.query_iterator(
-                expr=filter_string,
+            iterator = self.client.query_iterator(
+                collection_name=f'{self.collection_prefix}_{collection_name}',
+                filter=filter_string,
                 output_fields=[
                     'id',
                     'data',
@@ -270,18 +274,28 @@ class MilvusClient(VectorDBBase):
             self._create_collection(collection_name=collection_name, dimension=len(items[0]['vector']))
 
         log.info(f'Inserting {len(items)} items into collection {self.collection_prefix}_{collection_name}.')
-        return self.client.insert(
-            collection_name=f'{self.collection_prefix}_{collection_name}',
-            data=[
+        data = []
+        for item in items:
+            text = item['text'] or ''
+            if len(text) > MILVUS_TEXT_MAX_LENGTH:
+                log.warning(f'Milvus: truncating text id={item["id"]} {len(text)}->{MILVUS_TEXT_MAX_LENGTH} chars')
+                text = text[:MILVUS_TEXT_MAX_LENGTH]
+            data.append(
                 {
                     'id': item['id'],
                     'vector': item['vector'],
-                    'data': {'text': item['text']},
+                    'data': {'text': text},
                     'metadata': process_metadata(item['metadata']),
                 }
-                for item in items
-            ],
-        )
+            )
+        try:
+            return self.client.insert(
+                collection_name=f'{self.collection_prefix}_{collection_name}',
+                data=data,
+            )
+        except MilvusException as e:
+            log.error(f'Milvus insert failed for {self.collection_prefix}_{collection_name} ({len(items)} items): {e}')
+            raise
 
     def upsert(self, collection_name: str, items: list[VectorItem]):
         # Update the items in the collection, if the items are not present, insert them. If the collection does not exist, it will be created.
@@ -298,18 +312,28 @@ class MilvusClient(VectorDBBase):
             self._create_collection(collection_name=collection_name, dimension=len(items[0]['vector']))
 
         log.info(f'Upserting {len(items)} items into collection {self.collection_prefix}_{collection_name}.')
-        return self.client.upsert(
-            collection_name=f'{self.collection_prefix}_{collection_name}',
-            data=[
+        data = []
+        for item in items:
+            text = item['text'] or ''
+            if len(text) > MILVUS_TEXT_MAX_LENGTH:
+                log.warning(f'Milvus: truncating text id={item["id"]} {len(text)}->{MILVUS_TEXT_MAX_LENGTH} chars')
+                text = text[:MILVUS_TEXT_MAX_LENGTH]
+            data.append(
                 {
                     'id': item['id'],
                     'vector': item['vector'],
-                    'data': {'text': item['text']},
+                    'data': {'text': text},
                     'metadata': process_metadata(item['metadata']),
                 }
-                for item in items
-            ],
-        )
+            )
+        try:
+            return self.client.upsert(
+                collection_name=f'{self.collection_prefix}_{collection_name}',
+                data=data,
+            )
+        except MilvusException as e:
+            log.error(f'Milvus upsert failed for {self.collection_prefix}_{collection_name} ({len(items)} items): {e}')
+            raise
 
     def delete(
         self,

backend/open_webui/retrieval/vector/dbs/milvus_multitenancy.py

@@ -23,18 +23,17 @@ from open_webui.retrieval.vector.main import (
     VectorDBBase,
     VectorItem,
 )
-from pymilvus import (
-    Collection,
-    CollectionSchema,
-    DataType,
-    FieldSchema,
-    connections,
-    utility,
-)
+from pymilvus import DataType
+from pymilvus import MilvusClient as Client
+from pymilvus.exceptions import MilvusException
 
 log = logging.getLogger(__name__)
 
 RESOURCE_ID_FIELD = 'resource_id'
+# Milvus VARCHAR hard cap for the `text` field (see _create_shared_collection).
+# Chunks longer than this are truncated before insert so one oversized chunk
+# can't fail the whole batch (and leave the file with zero embeddings).
+MILVUS_TEXT_MAX_LENGTH = 65535
 
 # Milvus expressions are SQL-like strings with no parameterized-query API;
 # values get interpolated into single-quoted literals. Reject anything that
@@ -65,12 +64,7 @@ class MilvusClient(VectorDBBase):
     def __init__(self):
         # Milvus collection names can only contain numbers, letters, and underscores.
         self.collection_prefix = MILVUS_COLLECTION_PREFIX.replace('-', '_')
-        connections.connect(
-            alias='default',
-            uri=MILVUS_URI,
-            token=MILVUS_TOKEN,
-            db_name=MILVUS_DB,
-        )
+        self.client = Client(uri=MILVUS_URI, token=MILVUS_TOKEN, db_name=MILVUS_DB)
 
         # Main collection types for multi-tenancy
         self.MEMORY_COLLECTION = f'{self.collection_prefix}_memories'
@@ -111,53 +105,66 @@ class MilvusClient(VectorDBBase):
             return self.KNOWLEDGE_COLLECTION, resource_id
 
     def _create_shared_collection(self, mt_collection_name: str, dimension: int):
-        fields = [
-            FieldSchema(
-                name='id',
-                dtype=DataType.VARCHAR,
-                is_primary=True,
-                auto_id=False,
-                max_length=36,
-            ),
-            FieldSchema(name='vector', dtype=DataType.FLOAT_VECTOR, dim=dimension),
-            FieldSchema(name='text', dtype=DataType.VARCHAR, max_length=65535),
-            FieldSchema(name='metadata', dtype=DataType.JSON),
-            FieldSchema(name=RESOURCE_ID_FIELD, dtype=DataType.VARCHAR, max_length=255),
-        ]
-        schema = CollectionSchema(fields, 'Shared collection for multi-tenancy')
-        collection = Collection(mt_collection_name, schema)
+        schema = self.client.create_schema(auto_id=False, description='Shared collection for multi-tenancy')
+        schema.add_field(field_name='id', datatype=DataType.VARCHAR, is_primary=True, max_length=36)
+        schema.add_field(field_name='vector', datatype=DataType.FLOAT_VECTOR, dim=dimension)
+        schema.add_field(field_name='text', datatype=DataType.VARCHAR, max_length=MILVUS_TEXT_MAX_LENGTH)
+        schema.add_field(field_name='metadata', datatype=DataType.JSON)
+        schema.add_field(field_name=RESOURCE_ID_FIELD, datatype=DataType.VARCHAR, max_length=255)
 
-        index_params = {
-            'metric_type': MILVUS_METRIC_TYPE,
-            'index_type': MILVUS_INDEX_TYPE,
-            'params': {},
-        }
+        index_build_params = {}
         if MILVUS_INDEX_TYPE == 'HNSW':
-            index_params['params'] = {
+            index_build_params = {
                 'M': MILVUS_HNSW_M,
                 'efConstruction': MILVUS_HNSW_EFCONSTRUCTION,
             }
         elif MILVUS_INDEX_TYPE == 'IVF_FLAT':
-            index_params['params'] = {'nlist': MILVUS_IVF_FLAT_NLIST}
+            index_build_params = {'nlist': MILVUS_IVF_FLAT_NLIST}
 
-        collection.create_index('vector', index_params)
-        collection.create_index(RESOURCE_ID_FIELD)
+        vector_index = self.client.prepare_index_params(
+            field_name='vector',
+            index_type=MILVUS_INDEX_TYPE,
+            metric_type=MILVUS_METRIC_TYPE,
+            params=index_build_params,
+        )
+
+        self.client.create_collection(collection_name=mt_collection_name, schema=schema)
+        self.client.create_index(collection_name=mt_collection_name, index_params=vector_index)
+        try:
+            # A Milvus server auto-selects the scalar index type from a parameterless call.
+            self.client.create_index(
+                collection_name=mt_collection_name,
+                index_params=self.client.prepare_index_params(field_name=RESOURCE_ID_FIELD),
+            )
+        except MilvusException:
+            try:
+                self.client.create_index(
+                    collection_name=mt_collection_name,
+                    index_params=self.client.prepare_index_params(field_name=RESOURCE_ID_FIELD, index_type='INVERTED'),
+                )
+            except MilvusException as e:
+                # The index only accelerates resource_id filters; never fail
+                # collection creation over it.
+                log.warning(f'Could not create {RESOURCE_ID_FIELD} index on {mt_collection_name}: {e}')
         log.info(f'Created shared collection: {mt_collection_name}')
-        return collection
 
     def _ensure_collection(self, mt_collection_name: str, dimension: int):
-        if not utility.has_collection(mt_collection_name):
+        if not self.client.has_collection(mt_collection_name):
             self._create_shared_collection(mt_collection_name, dimension)
 
     def has_collection(self, collection_name: str) -> bool:
         mt_collection, resource_id = self._get_collection_and_resource_id(collection_name)
         _validate_resource_id(resource_id)
-        if not utility.has_collection(mt_collection):
+        if not self.client.has_collection(mt_collection):
             return False
 
-        collection = Collection(mt_collection)
-        collection.load()
-        res = collection.query(expr=f"{RESOURCE_ID_FIELD} == '{resource_id}'", limit=1)
+        self.client.load_collection(mt_collection)
+        res = self.client.query(
+            collection_name=mt_collection,
+            filter=f"{RESOURCE_ID_FIELD} == '{resource_id}'",
+            output_fields=['id'],
+            limit=1,
+        )
         return len(res) > 0
 
     def upsert(self, collection_name: str, items: List[VectorItem]):
@@ -167,19 +174,35 @@ class MilvusClient(VectorDBBase):
         _validate_resource_id(resource_id)
         dimension = len(items[0]['vector'])
         self._ensure_collection(mt_collection, dimension)
-        collection = Collection(mt_collection)
 
-        entities = [
-            {
-                'id': item['id'],
-                'vector': item['vector'],
-                'text': item['text'],
-                'metadata': item['metadata'],
-                RESOURCE_ID_FIELD: resource_id,
-            }
-            for item in items
-        ]
-        collection.insert(entities)
+        entities = []
+        for item in items:
+            text = item['text'] or ''
+            if len(text) > MILVUS_TEXT_MAX_LENGTH:
+                log.warning(
+                    f'Milvus: truncating text id={item["id"]} '
+                    f'{len(text)}->{MILVUS_TEXT_MAX_LENGTH} chars '
+                    f'(collection={mt_collection}, resource_id={resource_id})'
+                )
+                text = text[:MILVUS_TEXT_MAX_LENGTH]
+            entities.append(
+                {
+                    'id': item['id'],
+                    'vector': item['vector'],
+                    'text': text,
+                    'metadata': item['metadata'],
+                    RESOURCE_ID_FIELD: resource_id,
+                }
+            )
+
+        try:
+            self.client.insert(collection_name=mt_collection, data=entities)
+        except MilvusException as e:
+            log.error(
+                f'Milvus insert failed (collection={mt_collection}, '
+                f'resource_id={resource_id}, items={len(entities)}): {e}'
+            )
+            raise
 
     def search(
         self,
@@ -193,19 +216,18 @@ class MilvusClient(VectorDBBase):
 
         mt_collection, resource_id = self._get_collection_and_resource_id(collection_name)
         _validate_resource_id(resource_id)
-        if not utility.has_collection(mt_collection):
+        if not self.client.has_collection(mt_collection):
             return None
 
-        collection = Collection(mt_collection)
-        collection.load()
+        self.client.load_collection(mt_collection)
 
-        search_params = {'metric_type': MILVUS_METRIC_TYPE, 'params': {}}
-        results = collection.search(
+        results = self.client.search(
+            collection_name=mt_collection,
             data=vectors,
             anns_field='vector',
-            param=search_params,
+            search_params={'metric_type': MILVUS_METRIC_TYPE, 'params': {}},
             limit=limit,
-            expr=f"{RESOURCE_ID_FIELD} == '{resource_id}'",
+            filter=f"{RESOURCE_ID_FIELD} == '{resource_id}'",
             output_fields=['id', 'text', 'metadata'],
         )
 
@@ -213,10 +235,11 @@ class MilvusClient(VectorDBBase):
         for hits in results:
             batch_ids, batch_docs, batch_metadatas, batch_dists = [], [], [], []
             for hit in hits:
-                batch_ids.append(hit.entity.get('id'))
-                batch_docs.append(hit.entity.get('text'))
-                batch_metadatas.append(hit.entity.get('metadata'))
-                batch_dists.append(hit.distance)
+                entity = hit.get('entity', {})
+                batch_ids.append(entity.get('id'))
+                batch_docs.append(entity.get('text'))
+                batch_metadatas.append(entity.get('metadata'))
+                batch_dists.append(hit.get('distance'))
             ids.append(batch_ids)
             documents.append(batch_docs)
             metadatas.append(batch_metadatas)
@@ -232,11 +255,9 @@ class MilvusClient(VectorDBBase):
     ):
         mt_collection, resource_id = self._get_collection_and_resource_id(collection_name)
         _validate_resource_id(resource_id)
-        if not utility.has_collection(mt_collection):
+        if not self.client.has_collection(mt_collection):
             return
 
-        collection = Collection(mt_collection)
-
         expr = [f"{RESOURCE_ID_FIELD} == '{resource_id}'"]
         if ids:
             # Milvus expects a string list for 'in' operator
@@ -248,30 +269,28 @@ class MilvusClient(VectorDBBase):
                 _validate_metadata_key(key)
                 expr.append(f"metadata['{key}'] == '{_escape_milvus_string(str(value))}'")
 
-        collection.delete(' and '.join(expr))
+        self.client.delete(collection_name=mt_collection, filter=' and '.join(expr))
 
     def reset(self):
         for collection_name in self.shared_collections:
-            if utility.has_collection(collection_name):
-                utility.drop_collection(collection_name)
+            if self.client.has_collection(collection_name):
+                self.client.drop_collection(collection_name)
 
     def delete_collection(self, collection_name: str):
         mt_collection, resource_id = self._get_collection_and_resource_id(collection_name)
         _validate_resource_id(resource_id)
-        if not utility.has_collection(mt_collection):
+        if not self.client.has_collection(mt_collection):
             return
 
-        collection = Collection(mt_collection)
-        collection.delete(f"{RESOURCE_ID_FIELD} == '{resource_id}'")
+        self.client.delete(collection_name=mt_collection, filter=f"{RESOURCE_ID_FIELD} == '{resource_id}'")
 
     def query(self, collection_name: str, filter: Dict[str, Any], limit: Optional[int] = None) -> Optional[GetResult]:
         mt_collection, resource_id = self._get_collection_and_resource_id(collection_name)
         _validate_resource_id(resource_id)
-        if not utility.has_collection(mt_collection):
+        if not self.client.has_collection(mt_collection):
             return None
 
-        collection = Collection(mt_collection)
-        collection.load()
+        self.client.load_collection(mt_collection)
 
         expr = [f"{RESOURCE_ID_FIELD} == '{resource_id}'"]
         if filter:
@@ -286,8 +305,9 @@ class MilvusClient(VectorDBBase):
                 else:
                     raise TypeError(f'Unsupported Milvus filter value type for key {key!r}: {type(value).__name__}')
 
-        iterator = collection.query_iterator(
-            expr=' and '.join(expr),
+        iterator = self.client.query_iterator(
+            collection_name=mt_collection,
+            filter=' and '.join(expr),
             output_fields=['id', 'text', 'metadata'],
             limit=limit if limit else -1,
         )

backend/open_webui/retrieval/vector/dbs/pgvector.py

@@ -24,7 +24,7 @@ from open_webui.retrieval.vector.main import (
     VectorDBBase,
     VectorItem,
 )
-from open_webui.retrieval.vector.utils import process_metadata
+from open_webui.retrieval.vector.utils import merge_hybrid_search_results, process_metadata
 from open_webui.utils.misc import sanitize_text_for_db
 from pgvector.sqlalchemy import HALFVEC, Vector
 from sqlalchemy import (
@@ -153,6 +153,7 @@ class PgvectorClient(VectorDBBase):
 
             index_method, index_options = self._vector_index_configuration()
             self._ensure_vector_index(index_method, index_options)
+            self._ensure_text_search_index()
 
             self.session.execute(
                 text(
@@ -236,6 +237,19 @@ class PgvectorClient(VectorDBBase):
                 f' {index_options}' if index_options else '',
             )
 
+    def _ensure_text_search_index(self) -> None:
+        if PGVECTOR_PGCRYPTO:
+            return
+
+        self.session.execute(
+            text("""
+                CREATE INDEX IF NOT EXISTS idx_document_chunk_text_search
+                ON document_chunk
+                USING GIN (to_tsvector('simple', coalesce(text, '')));
+                """)
+        )
+        log.info("Ensured text search index 'idx_document_chunk_text_search'.")
+
     def check_vector_length(self) -> None:
         """
         Check if the VECTOR_LENGTH matches the existing vector column dimension in the database.
@@ -521,6 +535,71 @@ class PgvectorClient(VectorDBBase):
             log.exception(f'Error during search: {e}')
             return None
 
+    def hybrid_search(
+        self,
+        collection_name: str,
+        query: str,
+        vectors: List[List[float]],
+        filter: Optional[Dict[str, Any]] = None,
+        limit: int = 10,
+        hybrid_bm25_weight: float = 0.5,
+    ) -> Optional[SearchResult]:
+        if PGVECTOR_PGCRYPTO or filter:
+            return None
+
+        try:
+            limit = max(1, limit)
+            vectors = [self.adjust_vector_length(vector) for vector in vectors] if vectors else []
+            num_queries = len(vectors) if vectors else 1
+            bm25_weight = min(max(hybrid_bm25_weight, 0.0), 1.0)
+            vector_weight = 1.0 - bm25_weight
+
+            vector_result = None
+            if vector_weight > 0 and vectors:
+                vector_result = self.search(collection_name=collection_name, vectors=vectors, limit=limit)
+
+            fts_results = []
+            if bm25_weight > 0 and query and query.strip():
+                fts_rows = self.session.execute(
+                    text("""
+                        WITH fts_query AS (
+                            SELECT plainto_tsquery('simple', :query) AS query
+                        )
+                        SELECT
+                            document_chunk.id AS id,
+                            document_chunk.text AS text,
+                            document_chunk.vmetadata AS vmetadata,
+                            ts_rank_cd(
+                                to_tsvector('simple', coalesce(document_chunk.text, '')),
+                                fts_query.query
+                            ) AS rank
+                        FROM document_chunk, fts_query
+                        WHERE document_chunk.collection_name = :collection_name
+                          AND to_tsvector('simple', coalesce(document_chunk.text, '')) @@ fts_query.query
+                        ORDER BY rank DESC
+                        LIMIT :limit
+                    """),
+                    {
+                        'collection_name': collection_name,
+                        'query': query,
+                        'limit': limit,
+                    },
+                )
+                fts_results = [dict(row) for row in fts_rows.mappings().all()]
+                self.session.rollback()
+
+            return merge_hybrid_search_results(
+                vector_result=vector_result,
+                fts_results=fts_results,
+                num_queries=num_queries,
+                limit=limit,
+                hybrid_bm25_weight=hybrid_bm25_weight,
+            )
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f'Error during hybrid search: {e}')
+            return None
+
     def query(self, collection_name: str, filter: Dict[str, Any], limit: Optional[int] = None) -> Optional[GetResult]:
         try:
             if PGVECTOR_PGCRYPTO:

backend/open_webui/retrieval/vector/main.py

@@ -63,6 +63,18 @@ class VectorDBBase(ABC):
         """Search for similar vectors in a collection."""
         pass
 
+    def hybrid_search(
+        self,
+        collection_name: str,
+        query: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[Dict] = None,
+        limit: int = 10,
+        hybrid_bm25_weight: float = 0.5,
+    ) -> Optional[SearchResult]:
+        """Search using a backend-native hybrid keyword/vector implementation when available."""
+        return None
+
     @abstractmethod
     def query(self, collection_name: str, filter: Dict, limit: Optional[int] = None) -> Optional[GetResult]:
         """Query vectors from a collection using metadata filter."""

backend/open_webui/retrieval/vector/utils.py

@@ -1,5 +1,7 @@
-from datetime import datetime
+import datetime as dt
+from typing import Any
 
+from open_webui.retrieval.vector.main import SearchResult
 from open_webui.utils.misc import sanitize_text_for_db
 
 KEYS_TO_EXCLUDE = ['content', 'pages', 'tables', 'paragraphs', 'sections', 'figures']
@@ -21,9 +23,71 @@ def process_metadata(
         # Skip large fields
         if key in KEYS_TO_EXCLUDE:
             continue
+        if value is None:
+            continue
         # Convert non-serializable fields to strings
-        if isinstance(value, (datetime, list, dict)):
+        if isinstance(value, (dt.datetime, list, dict)):
             result[key] = sanitize_text_for_db(str(value))
         else:
             result[key] = sanitize_text_for_db(value)
     return result
+
+
+def merge_hybrid_search_results(
+    vector_result: SearchResult | None,
+    fts_results: list[dict[str, Any]],
+    num_queries: int,
+    limit: int,
+    hybrid_bm25_weight: float,
+) -> SearchResult:
+    rank_constant = 60.0
+    bm25_weight = min(max(hybrid_bm25_weight, 0.0), 1.0)
+    vector_weight = 1.0 - bm25_weight
+
+    ids = [[] for _ in range(num_queries)]
+    distances = [[] for _ in range(num_queries)]
+    documents = [[] for _ in range(num_queries)]
+    metadatas = [[] for _ in range(num_queries)]
+
+    for qid in range(num_queries):
+        candidates: dict[str, dict[str, Any]] = {}
+
+        if vector_result and vector_result.ids and qid < len(vector_result.ids):
+            for rank, item_id in enumerate(vector_result.ids[qid] or [], start=1):
+                score = vector_weight / (rank_constant + rank) if vector_weight > 0 else 0
+                if score <= 0:
+                    continue
+
+                candidate = candidates.setdefault(
+                    item_id,
+                    {
+                        'score': 0.0,
+                        'document': vector_result.documents[qid][rank - 1],
+                        'metadata': vector_result.metadatas[qid][rank - 1],
+                    },
+                )
+                candidate['score'] += score
+
+        for rank, row in enumerate(fts_results, start=1):
+            score = bm25_weight / (rank_constant + rank) if bm25_weight > 0 else 0
+            if score <= 0:
+                continue
+
+            item_id = row['id']
+            candidate = candidates.setdefault(
+                item_id,
+                {
+                    'score': 0.0,
+                    'document': row['text'],
+                    'metadata': row['vmetadata'],
+                },
+            )
+            candidate['score'] += score
+
+        ranked = sorted(candidates.items(), key=lambda item: item[1]['score'], reverse=True)[:limit]
+        ids[qid] = [item_id for item_id, _ in ranked]
+        distances[qid] = [candidate['score'] for _, candidate in ranked]
+        documents[qid] = [candidate['document'] for _, candidate in ranked]
+        metadatas[qid] = [candidate['metadata'] for _, candidate in ranked]
+
+    return SearchResult(ids=ids, distances=distances, documents=documents, metadatas=metadatas)

backend/open_webui/retrieval/web/firecrawl.py

@@ -28,10 +28,10 @@ def build_firecrawl_url(base_url: str | None, path: str) -> str:
 
 
 def build_firecrawl_headers(api_key: str | None) -> dict[str, str]:
-    return {
-        'Content-Type': 'application/json',
-        'Authorization': f'Bearer {api_key or ""}',
-    }
+    headers = {'Content-Type': 'application/json'}
+    if api_key:
+        headers['Authorization'] = f'Bearer {api_key}'
+    return headers
 
 
 def get_firecrawl_timeout_seconds(timeout: Any) -> float | None:

backend/open_webui/retrieval/web/main.py

@@ -1,10 +1,11 @@
 from __future__ import annotations
 
+import ipaddress
 from urllib.parse import urlparse
 
 import validators
 from open_webui.retrieval.web.utils import resolve_hostname
-from open_webui.utils.misc import is_string_allowed
+from open_webui.utils.misc import get_allow_block_lists, is_host_allowed
 from pydantic import BaseModel
 
 
@@ -12,6 +13,16 @@ def get_filtered_results(results, filter_list):
     if not filter_list:
         return results
 
+    allow_list, block_list = get_allow_block_lists(filter_list)
+    resolve_ips = False
+    for entry in allow_list + block_list:
+        try:
+            ipaddress.ip_address(entry)
+        except ValueError:
+            continue
+        resolve_ips = True
+        break
+
     filtered_results = []
 
     for result in results:
@@ -19,20 +30,21 @@ def get_filtered_results(results, filter_list):
         if not validators.url(url):
             continue
 
-        domain = urlparse(url).netloc
+        domain = urlparse(url).hostname
         if not domain:
             continue
 
         hostnames = [domain]
 
-        try:
-            ipv4_addresses, ipv6_addresses = resolve_hostname(domain)
-            hostnames.extend(ipv4_addresses)
-            hostnames.extend(ipv6_addresses)
-        except Exception:
-            pass
+        if resolve_ips:
+            try:
+                ipv4_addresses, ipv6_addresses = resolve_hostname(domain)
+                hostnames.extend(ipv4_addresses)
+                hostnames.extend(ipv6_addresses)
+            except Exception:
+                pass
 
-        if is_string_allowed(hostnames, filter_list):
+        if is_host_allowed(hostnames, filter_list):
             filtered_results.append(result)
             continue
 

backend/open_webui/retrieval/web/microsoft_web_iq.py

@@ -0,0 +1,60 @@
+from __future__ import annotations
+
+import logging
+from urllib.parse import urlparse
+
+import requests
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.utils.headers import include_user_info_headers
+
+log = logging.getLogger(__name__)
+
+DEFAULT_MICROSOFT_WEB_IQ_API_BASE_URL = 'https://api.microsoft.ai/v3'
+
+
+def search_microsoft_web_iq(
+    api_base_url: str,
+    api_key: str,
+    query: str,
+    count: int,
+    filter_list: list[str | None] | None = None,
+    language: str = 'en',
+    user=None,
+) -> list[SearchResult]:
+    try:
+        api_base_url = (api_base_url or DEFAULT_MICROSOFT_WEB_IQ_API_BASE_URL).rstrip('/')
+        headers = {
+            'host': urlparse(api_base_url).netloc or 'api.microsoft.ai',
+            'x-apikey': api_key,
+            'content-type': 'application/json',
+        }
+        if user is not None:
+            headers = include_user_info_headers(headers, user)
+
+        response = requests.post(
+            f'{api_base_url}/search/web',
+            json={
+                'query': query,
+                'maxResults': count,
+                'language': language,
+                'contentFormat': 'passage',
+            },
+            headers=headers,
+        )
+        response.raise_for_status()
+
+        results = response.json().get('webResults', [])
+        if filter_list:
+            results = get_filtered_results(results, filter_list)
+
+        return [
+            SearchResult(
+                link=result['url'],
+                title=result.get('title'),
+                snippet=result.get('content'),
+            )
+            for result in results
+        ]
+    except Exception as e:
+        log.error(f'Error searching with Microsoft Web IQ API: {e}')
+        return []

backend/open_webui/retrieval/web/openserp.py

@@ -0,0 +1,45 @@
+from __future__ import annotations
+
+import logging
+
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.utils.session_pool import get_session
+
+log = logging.getLogger(__name__)
+
+
+async def search_openserp(
+    base_url: str,
+    query: str,
+    count: int,
+    filter_list: list[str | None] | None = None,
+) -> list[SearchResult]:
+    """Query an OpenSERP instance and return normalised results.
+
+    OpenSERP aggregates results from 6 engines (google, bing, yandex,
+    baidu, duckduckgo, ecosia) at once via ``/mega/search``.
+
+    No API key is required -- only a reachable OpenSERP base URL.
+    """
+    url = f'{base_url.rstrip("/")}/mega/search'
+    params = {'text': query, 'limit': count}
+
+    log.debug('searching OpenSERP at %s', url)
+
+    session = await get_session()
+    async with session.get(url, params=params) as response:
+        response.raise_for_status()
+        payload = await response.json()
+
+    results = payload.get('results', [])
+    if filter_list:
+        results = get_filtered_results(results, filter_list)
+
+    return [
+        SearchResult(
+            link=item.get('url', ''),
+            title=item.get('title'),
+            snippet=item.get('snippet'),
+        )
+        for item in results[:count]
+    ]

backend/open_webui/retrieval/web/perplexity.py

@@ -38,9 +38,7 @@ def search_perplexity(
 
     """
 
-    # Handle ConfigVar object
-    if hasattr(api_key, '__str__'):
-        api_key = str(api_key)
+    api_key = str(api_key)
 
     try:
         url = 'https://api.perplexity.ai/chat/completions'

backend/open_webui/retrieval/web/perplexity_search.py

@@ -29,12 +29,8 @@ def search_perplexity_search(
 
     """
 
-    # Handle ConfigVar object
-    if hasattr(api_key, '__str__'):
-        api_key = str(api_key)
-
-    if hasattr(api_url, '__str__'):
-        api_url = str(api_url)
+    api_key = str(api_key)
+    api_url = str(api_url)
 
     try:
         url = api_url

backend/open_webui/retrieval/web/searxng.py

@@ -1,7 +1,10 @@
 from __future__ import annotations
 
 import logging
+import ssl
+from functools import lru_cache
 
+from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, SEARXNG_CLIENT_CERT_FILE, SEARXNG_CLIENT_KEY_FILE
 from open_webui.retrieval.web.main import SearchResult, get_filtered_results
 from open_webui.utils.session_pool import get_session
 
@@ -17,6 +20,19 @@ _SEARXNG_HEADERS = {
 }
 
 
+@lru_cache
+def _get_ssl_context() -> bool | ssl.SSLContext:
+    if not SEARXNG_CLIENT_CERT_FILE:
+        return AIOHTTP_CLIENT_SESSION_SSL
+
+    ssl_context = ssl.create_default_context()
+    ssl_context.load_cert_chain(
+        certfile=SEARXNG_CLIENT_CERT_FILE,
+        keyfile=SEARXNG_CLIENT_KEY_FILE or None,
+    )
+    return ssl_context
+
+
 async def search_searxng(
     query_url: str,
     query: str,
@@ -48,7 +64,12 @@ async def search_searxng(
     log.debug('searching %s', query_url)
 
     session = await get_session()
-    async with session.get(query_url, headers=_SEARXNG_HEADERS, params=params) as response:
+    async with session.get(
+        query_url,
+        headers=_SEARXNG_HEADERS,
+        params=params,
+        ssl=_get_ssl_context(),
+    ) as response:
         response.raise_for_status()
         payload = await response.json()
 

backend/open_webui/retrieval/web/serphouse.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+from open_webui.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.utils.session_pool import get_session
+
+
+async def search_serphouse(
+    api_key: str,
+    domain: str,
+    query: str,
+    count: int,
+    filter_list: list[str | None] | None = None,
+) -> list[SearchResult]:
+    """Query SERPHouse and return normalised organic results."""
+    session = await get_session()
+    async with session.get(
+        'https://api.serphouse.com/serp/live',
+        params={
+            'q': query,
+            'domain': (domain or 'google.com').strip() or 'google.com',
+            'device': 'desktop',
+            'serp_type': 'web',
+            'page': 1,
+            'num_result': count,
+        },
+        headers={'Authorization': f'Bearer {api_key}', 'Accept': 'application/json'},
+    ) as response:
+        response.raise_for_status()
+        payload = await response.json()
+
+    organic = payload.get('results', {}).get('results', {}).get('organic', [])
+    organic = sorted(organic, key=lambda item: item.get('position', 0))
+    if filter_list:
+        organic = get_filtered_results(organic, filter_list)
+
+    return [
+        SearchResult(
+            link=item.get('link', ''),
+            title=item.get('title'),
+            snippet=item.get('snippet'),
+        )
+        for item in organic[:count]
+        if item.get('link')
+    ]

backend/open_webui/retrieval/web/utils.py

@@ -3,9 +3,10 @@ import ipaddress
 import logging
 import socket
 import ssl
+import time
 import urllib.parse
 import urllib.request
-from datetime import datetime, time, timedelta
+from datetime import datetime, timedelta
 from typing import (
     Any,
     AsyncIterator,
@@ -21,7 +22,6 @@ from typing import (
 import aiohttp
 import aiohttp.resolver
 import certifi
-import requests
 import urllib3.connection
 import urllib3.connectionpool
 import validators
@@ -31,12 +31,15 @@ from langchain_community.document_loaders import PlaywrightURLLoader, WebBaseLoa
 from langchain_community.document_loaders.base import BaseLoader
 from langchain_core.documents import Document
 from open_webui.config import (
-    ENABLE_RAG_LOCAL_WEB_FETCH,
+    ENABLE_LOCAL_WEB_FETCH,
     EXTERNAL_WEB_LOADER_API_KEY,
     EXTERNAL_WEB_LOADER_URL,
     FIRECRAWL_API_BASE_URL,
     FIRECRAWL_API_KEY,
     FIRECRAWL_TIMEOUT,
+    MICROSOFT_WEB_IQ_API_BASE_URL,
+    MICROSOFT_WEB_IQ_API_KEY,
+    MICROSOFT_WEB_IQ_LANGUAGE,
     PLAYWRIGHT_TIMEOUT,
     PLAYWRIGHT_WS_URL,
     TAVILY_API_KEY,
@@ -46,11 +49,17 @@ from open_webui.config import (
     WEB_LOADER_TIMEOUT,
 )
 from open_webui.constants import ERROR_MESSAGES
-from open_webui.env import AIOHTTP_CLIENT_ALLOW_REDIRECTS, AIOHTTP_CLIENT_SESSION_SSL, USER_AGENT
+from open_webui.env import (
+    AIOHTTP_CLIENT_ALLOW_REDIRECTS,
+    AIOHTTP_CLIENT_SESSION_SSL,
+    AIOHTTP_CLIENT_TIMEOUT,
+    USER_AGENT,
+)
 from open_webui.retrieval.loaders.external_web import ExternalWebLoader
+from open_webui.retrieval.loaders.microsoft_web_iq import MicrosoftWebIQLoader
 from open_webui.retrieval.loaders.tavily import TavilyLoader
 from open_webui.retrieval.web.firecrawl import scrape_firecrawl_url
-from open_webui.utils.misc import is_string_allowed
+from open_webui.utils.misc import is_host_allowed
 
 log = logging.getLogger(__name__)
 
@@ -66,6 +75,34 @@ def resolve_hostname(hostname):
     return ipv4_addresses, ipv6_addresses
 
 
+def _is_global_addr(ip: str) -> bool:
+    addr = ipaddress.ip_address(ip)
+    if not addr.is_global:
+        return False
+    if not isinstance(addr, ipaddress.IPv6Address):
+        return True
+
+    embedded = []
+    if addr.ipv4_mapped:
+        embedded.append(addr.ipv4_mapped)
+    if addr.sixtofour:
+        embedded.append(addr.sixtofour)
+    if addr.teredo:
+        embedded.extend(addr.teredo)
+
+    b = addr.packed
+    if b[:12] == b'\x00' * 12:
+        embedded.append(ipaddress.IPv4Address(b[12:]))
+    elif b[:12] == b'\x00\x64\xff\x9b' + b'\x00' * 8:
+        embedded.append(ipaddress.IPv4Address(b[12:]))
+    elif b[:6] == b'\x00\x64\xff\x9b\x00\x01':
+        if b[8] != 0:
+            return False
+        embedded.append(ipaddress.IPv4Address(bytes((b[6], b[7], b[9], b[10]))))
+
+    return all(ip.is_global for ip in embedded)
+
+
 def validate_url(url: Union[str, Sequence[str]]):
     if isinstance(url, str):
         if isinstance(validators.url(url), validators.ValidationError):
@@ -88,20 +125,21 @@ def validate_url(url: Union[str, Sequence[str]]):
 
         # Blocklist check using unified filtering logic
         if WEB_FETCH_FILTER_LIST:
-            if not is_string_allowed(url, WEB_FETCH_FILTER_LIST):
+            # Match on the parsed hostname, not the full URL: a path component would
+            # otherwise let any URL slip past a hostname-based block/allow entry.
+            if not is_host_allowed(parsed_url.hostname, WEB_FETCH_FILTER_LIST):
                 log.warning(f'URL blocked by filter list: {url}')
                 raise ValueError(ERROR_MESSAGES.INVALID_URL)
 
-        if not ENABLE_RAG_LOCAL_WEB_FETCH:
-            # Local web fetch is disabled, filter out any URLs that resolve to private IP addresses
+        if not ENABLE_LOCAL_WEB_FETCH:
+            # Local web fetch is disabled, filter out URLs that resolve to non-global IP addresses.
             parsed_url = urllib.parse.urlparse(url)
             # Get IPv4 and IPv6 addresses
             ipv4_addresses, ipv6_addresses = resolve_hostname(parsed_url.hostname)
             # Check if any of the resolved addresses are private
             # DNS rebinding is mitigated at the connection layer; see _SSRFSafeResolver / _SSRFSafeAdapter
             for ip in ipv4_addresses + ipv6_addresses:
-                addr = ipaddress.ip_address(ip)
-                if not addr.is_global:
+                if not _is_global_addr(ip):
                     raise ValueError(ERROR_MESSAGES.INVALID_URL)
         return True
     elif isinstance(url, Sequence):
@@ -134,9 +172,9 @@ def _ssrf_safe_new_conn(self):
     infos = socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM)
     if not infos:
         raise OSError(f'getaddrinfo for {host!r} returned empty list')
-    if not ENABLE_RAG_LOCAL_WEB_FETCH:
+    if not ENABLE_LOCAL_WEB_FETCH:
         for _, _, _, _, sa in infos:
-            if not ipaddress.ip_address(sa[0]).is_global:
+            if not _is_global_addr(sa[0]):
                 raise ValueError(ERROR_MESSAGES.INVALID_URL)
     err = None
     for fam, typ, proto, _, sa in infos:
@@ -148,6 +186,11 @@ def _ssrf_safe_new_conn(self):
             if getattr(self, 'source_address', None):
                 sock.bind(self.source_address)
             for opt in getattr(self, 'socket_options', None) or ():
+                if len(opt) == 4 and isinstance(opt[3], str):
+                    # urllib3-future per-protocol form: (level, optname, value, "tcp"/"udp")
+                    if opt[3].lower() == 'tcp':
+                        sock.setsockopt(*opt[:3])
+                    continue
                 sock.setsockopt(*opt)
             sock.connect(sa)
             return sock
@@ -190,13 +233,26 @@ class _SSRFSafeResolver(aiohttp.resolver.DefaultResolver):
 
     async def resolve(self, host, port=0, family=socket.AF_INET):
         results = await super().resolve(host, port, family)
-        if not ENABLE_RAG_LOCAL_WEB_FETCH:
+        if not ENABLE_LOCAL_WEB_FETCH:
             for entry in results:
-                if not ipaddress.ip_address(entry['host']).is_global:
+                if not _is_global_addr(entry['host']):
                     raise ValueError(ERROR_MESSAGES.INVALID_URL)
         return results
 
 
+def get_ssrf_safe_session() -> aiohttp.ClientSession:
+    """A one-off aiohttp session that re-validates the connect-time IP via _SSRFSafeResolver,
+    defeating DNS rebinding. Use for validate_url-gated fetches of user-supplied URLs that must
+    not use the shared (rebinding-vulnerable) pool. Use as a context manager so it is closed:
+    ``async with get_ssrf_safe_session() as session: ...``.
+    """
+    return aiohttp.ClientSession(
+        connector=aiohttp.TCPConnector(resolver=_SSRFSafeResolver()),
+        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT),
+        trust_env=True,
+    )
+
+
 def extract_metadata(soup, url):
     metadata = {'source': url}
     if title := soup.find('title'):
@@ -303,8 +359,9 @@ class SafeFireCrawlLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
         self.params = params or {}
 
     def lazy_load(self) -> Iterator[Document]:
-        try:
-            for url in self.web_paths:
+        for url in self.web_paths:
+            try:
+                self._sync_wait_for_rate_limit()
                 doc = scrape_firecrawl_url(
                     self.api_url,
                     self.api_key,
@@ -315,28 +372,39 @@ class SafeFireCrawlLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
                 )
                 if doc is not None:
                     yield doc
-        except Exception as e:
-            if self.continue_on_failure:
-                log.warning(f'Error extracting content from URLs with Firecrawl: {e}')
-            else:
-                raise e
+            except Exception as e:
+                if self.continue_on_failure:
+                    log.warning(f'Error extracting content from {url} with Firecrawl: {e}')
+                    continue
+                raise
 
     async def alazy_load(self):
-        try:
-            docs = await run_in_threadpool(lambda: list(self.lazy_load()))
-            for doc in docs:
-                yield doc
-        except Exception as e:
-            if self.continue_on_failure:
-                log.warning(f'Error extracting content from URLs with Firecrawl: {e}')
-            else:
-                raise e
+        for url in self.web_paths:
+            try:
+                await self._wait_for_rate_limit()
+                doc = await run_in_threadpool(
+                    scrape_firecrawl_url,
+                    self.api_url,
+                    self.api_key,
+                    url,
+                    verify_ssl=self.verify_ssl,
+                    timeout=self.timeout,
+                    params=self.params,
+                )
+                if doc is not None:
+                    yield doc
+            except Exception as e:
+                if self.continue_on_failure:
+                    log.warning(f'Error extracting content from {url} with Firecrawl: {e}')
+                    continue
+                raise
 
 
 class SafeTavilyLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
     def __init__(
         self,
         web_paths: Union[str, List[str]],
+        api_base_url: str,
         api_key: str,
         extract_depth: Literal['basic', 'advanced'] = 'basic',
         continue_on_failure: bool = True,
@@ -370,6 +438,7 @@ class SafeTavilyLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
 
         # Store parameters for creating TavilyLoader instances
         self.web_paths = web_paths if isinstance(web_paths, list) else [web_paths]
+        self.api_base_url = api_base_url
         self.api_key = api_key
         self.extract_depth = extract_depth
         self.continue_on_failure = continue_on_failure
@@ -445,6 +514,67 @@ class SafeTavilyLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
                 raise e
 
 
+class SafeMicrosoftWebIQLoader(BaseLoader, RateLimitMixin, URLProcessingMixin):
+    def __init__(
+        self,
+        web_paths: Union[str, List[str]],
+        api_key: str,
+        language: str = 'en',
+        verify_ssl: bool = True,
+        trust_env: bool = False,
+        requests_per_second: Optional[float] = None,
+        continue_on_failure: bool = True,
+        timeout: Optional[int] = None,
+    ):
+        self.web_paths = web_paths if isinstance(web_paths, list) else [web_paths]
+        self.api_key = api_key
+        self.language = language
+        self.verify_ssl = verify_ssl
+        self.trust_env = trust_env
+        self.requests_per_second = requests_per_second
+        self.last_request_time = None
+        self.continue_on_failure = continue_on_failure
+        self.timeout = timeout
+
+    def lazy_load(self) -> Iterator[Document]:
+        valid_urls = []
+        for url in self.web_paths:
+            try:
+                self._safe_process_url_sync(url)
+                valid_urls.append(url)
+            except Exception as e:
+                log.warning(f'SSL verification failed for {url}: {str(e)}')
+                if not self.continue_on_failure:
+                    raise e
+        if not valid_urls:
+            if self.continue_on_failure:
+                log.warning('No valid URLs to process after SSL verification')
+                return
+            raise ValueError('No valid URLs to process after SSL verification')
+
+        loader = MicrosoftWebIQLoader(
+            urls=valid_urls,
+            api_base_url=self.api_base_url,
+            api_key=self.api_key,
+            language=self.language,
+            verify_ssl=self.verify_ssl,
+            timeout=self.timeout,
+            continue_on_failure=self.continue_on_failure,
+        )
+        yield from loader.lazy_load()
+
+    async def alazy_load(self) -> AsyncIterator[Document]:
+        try:
+            docs = await run_in_threadpool(lambda: list(self.lazy_load()))
+            for doc in docs:
+                yield doc
+        except Exception as e:
+            if self.continue_on_failure:
+                log.warning(f'Error browsing URLs with Microsoft Web IQ: {e}')
+            else:
+                raise e
+
+
 class SafePlaywrightURLLoader(PlaywrightURLLoader, RateLimitMixin, URLProcessingMixin):
     """Load HTML pages safely with Playwright, supporting SSL verification, rate limiting, and remote browser connection.
 
@@ -503,57 +633,63 @@ class SafePlaywrightURLLoader(PlaywrightURLLoader, RateLimitMixin, URLProcessing
     def _intercept_navigation_sync(self, route, request=None):
         req = request or route.request
 
-        if req.resource_type != 'document':
-            route.continue_()
-            return
-
         try:
             validate_url(req.url)
+            resp = route.fetch(max_redirects=0)
+
+            if 300 <= resp.status < 400:
+                for _ in range(20):
+                    if not AIOHTTP_CLIENT_ALLOW_REDIRECTS:
+                        route.abort()
+                        return
+
+                    location = resp.headers.get('location')
+                    if not location:
+                        break
+
+                    url = urllib.parse.urljoin(resp.url, location)
+                    validate_url(url)
+                    resp = route.fetch(url=url, max_redirects=0)
+                    if not 300 <= resp.status < 400:
+                        break
+                else:
+                    route.abort()
+                    return
         except Exception:
             route.abort()
             return
 
-        if AIOHTTP_CLIENT_ALLOW_REDIRECTS:
-            resp = route.fetch()
-        else:
-            try:
-                resp = route.fetch(max_redirects=0)
-            except TypeError:
-                route.abort()
-                return
-
-            if 300 <= resp.status < 400:
-                route.abort()
-                return
-
         route.fulfill(response=resp)
 
     async def _intercept_navigation(self, route, request=None):
         req = request or route.request
 
-        if req.resource_type != 'document':
-            await route.continue_()
-            return
-
         try:
             await run_in_threadpool(validate_url, req.url)
+            resp = await route.fetch(max_redirects=0)
+
+            if 300 <= resp.status < 400:
+                for _ in range(20):
+                    if not AIOHTTP_CLIENT_ALLOW_REDIRECTS:
+                        await route.abort()
+                        return
+
+                    location = resp.headers.get('location')
+                    if not location:
+                        break
+
+                    url = urllib.parse.urljoin(resp.url, location)
+                    await run_in_threadpool(validate_url, url)
+                    resp = await route.fetch(url=url, max_redirects=0)
+                    if not 300 <= resp.status < 400:
+                        break
+                else:
+                    await route.abort()
+                    return
         except Exception:
             await route.abort()
             return
 
-        if AIOHTTP_CLIENT_ALLOW_REDIRECTS:
-            resp = await route.fetch()
-        else:
-            try:
-                resp = await route.fetch(max_redirects=0)
-            except TypeError:
-                await route.abort()
-                return
-
-            if 300 <= resp.status < 400:
-                await route.abort()
-                return
-
         await route.fulfill(response=resp)
 
     def lazy_load(self) -> Iterator[Document]:
@@ -567,24 +703,25 @@ class SafePlaywrightURLLoader(PlaywrightURLLoader, RateLimitMixin, URLProcessing
             else:
                 browser = p.chromium.launch(headless=self.headless, proxy=self.proxy)
 
-            for url in self.urls:
-                try:
-                    self._safe_process_url_sync(url)
-                    page = browser.new_page()
-                    page.route('**/*', self._intercept_navigation_sync)
-                    response = page.goto(url, timeout=self.playwright_timeout)
-                    if response is None:
-                        raise ValueError(f'page.goto() returned None for url {url}')
+            with browser:
+                for url in self.urls:
+                    try:
+                        self._safe_process_url_sync(url)
+                        with browser.new_page(service_workers='block') as page:
+                            page.route('**/*', self._intercept_navigation_sync)
+                            page.route_web_socket('**/*', lambda ws_route: ws_route.close())
+                            response = page.goto(url, timeout=self.playwright_timeout)
+                            if response is None:
+                                raise ValueError(f'page.goto() returned None for url {url}')
 
-                    text = self.evaluator.evaluate(page, browser, response)
-                    metadata = {'source': url}
-                    yield Document(page_content=text, metadata=metadata)
-                except Exception as e:
-                    if self.continue_on_failure:
-                        log.exception(f'Error loading {url}: {e}')
-                        continue
-                    raise e
-            browser.close()
+                            text = self.evaluator.evaluate(page, browser, response)
+                            metadata = {'source': url}
+                            yield Document(page_content=text, metadata=metadata)
+                    except Exception as e:
+                        if self.continue_on_failure:
+                            log.exception(f'Error loading {url}: {e}')
+                            continue
+                        raise e
 
     async def alazy_load(self) -> AsyncIterator[Document]:
         """Safely load URLs asynchronously with support for remote browser."""
@@ -597,24 +734,25 @@ class SafePlaywrightURLLoader(PlaywrightURLLoader, RateLimitMixin, URLProcessing
             else:
                 browser = await p.chromium.launch(headless=self.headless, proxy=self.proxy)
 
-            for url in self.urls:
-                try:
-                    await self._safe_process_url(url)
-                    page = await browser.new_page()
-                    await page.route('**/*', self._intercept_navigation)
-                    response = await page.goto(url, timeout=self.playwright_timeout)
-                    if response is None:
-                        raise ValueError(f'page.goto() returned None for url {url}')
+            async with browser:
+                for url in self.urls:
+                    try:
+                        await self._safe_process_url(url)
+                        async with await browser.new_page(service_workers='block') as page:
+                            await page.route('**/*', self._intercept_navigation)
+                            await page.route_web_socket('**/*', lambda ws_route: ws_route.close())
+                            response = await page.goto(url, timeout=self.playwright_timeout)
+                            if response is None:
+                                raise ValueError(f'page.goto() returned None for url {url}')
 
-                    text = await self.evaluator.evaluate_async(page, browser, response)
-                    metadata = {'source': url}
-                    yield Document(page_content=text, metadata=metadata)
-                except Exception as e:
-                    if self.continue_on_failure:
-                        log.exception(f'Error loading {url}: {e}')
-                        continue
-                    raise e
-            await browser.close()
+                            text = await self.evaluator.evaluate_async(page, browser, response)
+                            metadata = {'source': url}
+                            yield Document(page_content=text, metadata=metadata)
+                    except Exception as e:
+                        if self.continue_on_failure:
+                            log.exception(f'Error loading {url}: {e}')
+                            continue
+                        raise e
 
 
 class SafeWebBaseLoader(WebBaseLoader):
@@ -626,6 +764,8 @@ class SafeWebBaseLoader(WebBaseLoader):
             trust_env (bool, optional): set to True if using proxy to make web requests, for example
                 using http(s)_proxy environment variables. Defaults to False.
         """
+        # lxml parses scraped pages far faster than the html.parser default
+        kwargs.setdefault('default_parser', 'lxml')
         super().__init__(*args, **kwargs)
         self.trust_env = trust_env
 
@@ -688,20 +828,13 @@ class SafeWebBaseLoader(WebBaseLoader):
         final_results = []
         for i, result in enumerate(results):
             url = urls[i]
-            if parser is None:
-                if url.endswith('.xml'):
-                    parser = 'xml'
-                else:
-                    parser = self.default_parser
-                self._check_parser(parser)
-            final_results.append(BeautifulSoup(result, parser, **self.bs_kwargs))
+            url_parser = parser
+            if url_parser is None:
+                url_parser = 'xml' if url.endswith('.xml') else self.default_parser
+                self._check_parser(url_parser)
+            final_results.append(BeautifulSoup(result, url_parser, **self.bs_kwargs))
         return final_results
 
-    async def ascrape_all(self, urls: List[str], parser: Union[str, None] = None) -> List[Any]:
-        """Async fetch all urls, then return soups for all results."""
-        results = await self.fetch_all(urls)
-        return self._unpack_fetch_results(results, urls, parser=parser)
-
     def lazy_load(self) -> Iterator[Document]:
         """Lazy load text from the url(s) in web_path with error handling."""
         for path in self.web_paths:
@@ -717,19 +850,20 @@ class SafeWebBaseLoader(WebBaseLoader):
                 # Log the error and continue with the next URL
                 log.exception(f'Error loading {path}: {e}')
 
+    def _document_from_html(self, html: str, url: str) -> Document:
+        """Build one Document."""
+        soup = self._unpack_fetch_results([html], [url])[0]
+        return Document(
+            page_content=soup.get_text(**self.bs_get_text_kwargs),
+            metadata=extract_metadata(soup, url),
+        )
+
     async def alazy_load(self) -> AsyncIterator[Document]:
         """Async lazy load text from the url(s) in web_path."""
-        results = await self.ascrape_all(self.web_paths)
-        for path, soup in zip(self.web_paths, results):
-            text = soup.get_text(**self.bs_get_text_kwargs)
-            metadata = {'source': path}
-            if title := soup.find('title'):
-                metadata['title'] = title.get_text()
-            if description := soup.find('meta', attrs={'name': 'description'}):
-                metadata['description'] = description.get('content', 'No description found.')
-            if html := soup.find('html'):
-                metadata['language'] = html.get('lang', 'No language found.')
-            yield Document(page_content=text, metadata=metadata)
+        results = await self.fetch_all(self.web_paths)
+        for path, html in zip(self.web_paths, results):
+            # parsing a large page costs hundreds of ms, keep it off the event loop
+            yield await asyncio.to_thread(self._document_from_html, html, path)
 
     async def aload(self) -> list[Document]:
         """Load data into Document objects."""
@@ -741,6 +875,7 @@ def get_web_loader(
     verify_ssl: bool = True,
     requests_per_second: int = 2,
     trust_env: bool = False,
+    loader_config: Optional[dict] = None,
 ):
     # Check if the URLs are valid
     safe_urls = safe_validate_urls([urls] if isinstance(urls, str) else urls)
@@ -749,6 +884,16 @@ def get_web_loader(
         log.warning(f'All provided URLs were blocked or invalid: {urls}')
         raise ValueError(ERROR_MESSAGES.INVALID_URL)
 
+    loader_config = loader_config or {}
+
+    def cfg(key, env_value):
+        # Admin-saved DB value wins; env constant covers keys never saved.
+        value = loader_config.get(key)
+        return env_value if value is None else value
+
+    engine = cfg('web_loader_engine', WEB_LOADER_ENGINE)
+    web_loader_timeout = cfg('web_loader_timeout', WEB_LOADER_TIMEOUT)
+
     web_loader_args = {
         'web_paths': safe_urls,
         'verify_ssl': verify_ssl,
@@ -757,13 +902,15 @@ def get_web_loader(
         'trust_env': trust_env,
     }
 
-    if WEB_LOADER_ENGINE.value == '' or WEB_LOADER_ENGINE.value == 'safe_web':
+    WebLoaderClass = None
+
+    if engine == '' or engine == 'safe_web':
         WebLoaderClass = SafeWebBaseLoader
 
         request_kwargs = {}
-        if WEB_LOADER_TIMEOUT.value:
+        if web_loader_timeout:
             try:
-                timeout_value = float(WEB_LOADER_TIMEOUT.value)
+                timeout_value = float(web_loader_timeout)
             except ValueError:
                 timeout_value = None
 
@@ -773,31 +920,44 @@ def get_web_loader(
         if request_kwargs:
             web_loader_args['requests_kwargs'] = request_kwargs
 
-    if WEB_LOADER_ENGINE.value == 'playwright':
+    if engine == 'playwright':
         WebLoaderClass = SafePlaywrightURLLoader
-        web_loader_args['playwright_timeout'] = PLAYWRIGHT_TIMEOUT.value
-        if PLAYWRIGHT_WS_URL.value:
-            web_loader_args['playwright_ws_url'] = PLAYWRIGHT_WS_URL.value
+        web_loader_args['playwright_timeout'] = cfg('playwright_timeout', PLAYWRIGHT_TIMEOUT)
+        playwright_ws_url = cfg('playwright_ws_url', PLAYWRIGHT_WS_URL)
+        if playwright_ws_url:
+            web_loader_args['playwright_ws_url'] = playwright_ws_url
 
-    if WEB_LOADER_ENGINE.value == 'firecrawl':
+    if engine == 'firecrawl':
         WebLoaderClass = SafeFireCrawlLoader
-        web_loader_args['api_key'] = FIRECRAWL_API_KEY.value
-        web_loader_args['api_url'] = FIRECRAWL_API_BASE_URL.value
-        if FIRECRAWL_TIMEOUT.value:
+        web_loader_args['api_key'] = cfg('firecrawl_api_key', FIRECRAWL_API_KEY)
+        web_loader_args['api_url'] = cfg('firecrawl_api_url', FIRECRAWL_API_BASE_URL)
+        firecrawl_timeout = cfg('firecrawl_timeout', FIRECRAWL_TIMEOUT)
+        if firecrawl_timeout:
             try:
-                web_loader_args['timeout'] = int(FIRECRAWL_TIMEOUT.value)
+                web_loader_args['timeout'] = int(firecrawl_timeout)
             except ValueError:
                 pass
 
-    if WEB_LOADER_ENGINE.value == 'tavily':
+    if engine == 'tavily':
         WebLoaderClass = SafeTavilyLoader
-        web_loader_args['api_key'] = TAVILY_API_KEY.value
-        web_loader_args['extract_depth'] = TAVILY_EXTRACT_DEPTH.value
+        web_loader_args['api_key'] = cfg('tavily_api_key', TAVILY_API_KEY)
+        web_loader_args['extract_depth'] = cfg('tavily_extract_depth', TAVILY_EXTRACT_DEPTH)
 
-    if WEB_LOADER_ENGINE.value == 'external':
+    if engine == 'microsoft_web_iq':
+        WebLoaderClass = SafeMicrosoftWebIQLoader
+        web_loader_args['api_base_url'] = cfg('microsoft_web_iq_api_base_url', MICROSOFT_WEB_IQ_API_BASE_URL)
+        web_loader_args['api_key'] = cfg('microsoft_web_iq_api_key', MICROSOFT_WEB_IQ_API_KEY)
+        web_loader_args['language'] = cfg('microsoft_web_iq_language', MICROSOFT_WEB_IQ_LANGUAGE)
+        if web_loader_timeout:
+            try:
+                web_loader_args['timeout'] = int(web_loader_timeout)
+            except ValueError:
+                pass
+
+    if engine == 'external':
         WebLoaderClass = ExternalWebLoader
-        web_loader_args['external_url'] = EXTERNAL_WEB_LOADER_URL.value
-        web_loader_args['external_api_key'] = EXTERNAL_WEB_LOADER_API_KEY.value
+        web_loader_args['external_url'] = cfg('external_web_loader_url', EXTERNAL_WEB_LOADER_URL)
+        web_loader_args['external_api_key'] = cfg('external_web_loader_api_key', EXTERNAL_WEB_LOADER_API_KEY)
 
     if WebLoaderClass:
         web_loader = WebLoaderClass(**web_loader_args)
@@ -811,6 +971,6 @@ def get_web_loader(
         return web_loader
     else:
         raise ValueError(
-            f'Invalid WEB_LOADER_ENGINE: {WEB_LOADER_ENGINE.value}. '
-            "Please set it to 'safe_web', 'playwright', 'firecrawl', or 'tavily'."
+            f'Invalid WEB_LOADER_ENGINE: {engine}. '
+            "Please set it to 'safe_web', 'playwright', 'firecrawl', 'tavily', 'external', or 'microsoft_web_iq'."
         )

backend/open_webui/routers/analytics.py

@@ -28,6 +28,8 @@ router = APIRouter()
 class ModelAnalyticsEntry(BaseModel):
     model_id: str
     count: int
+    unique_users: int = 0
+    unique_chats: int = 0
 
 
 class ModelAnalyticsResponse(BaseModel):
@@ -65,8 +67,16 @@ async def get_model_analytics(
     counts = await ChatMessages.get_message_count_by_model(
         start_date=start_date, end_date=end_date, group_id=group_id, db=db
     )
+    unique_counts = await ChatMessages.get_unique_counts_by_model(
+        start_date=start_date, end_date=end_date, group_id=group_id, db=db
+    )
     models = [
-        ModelAnalyticsEntry(model_id=model_id, count=count)
+        ModelAnalyticsEntry(
+            model_id=model_id,
+            count=count,
+            unique_users=unique_counts.get(model_id, {}).get('unique_users', 0),
+            unique_chats=unique_counts.get(model_id, {}).get('unique_chats', 0),
+        )
         for model_id, count in sorted(counts.items(), key=lambda x: -x[1])
     ]
     return ModelAnalyticsResponse(models=models)
@@ -269,6 +279,9 @@ class ModelChatsResponse(BaseModel):
     total: int
 
 
+MODEL_CHAT_ORDER_FIELDS = {'title', 'updated_at', 'user_name'}
+
+
 @router.get('/models/{model_id:path}/chats', response_model=ModelChatsResponse)
 async def get_model_chats(
     model_id: str,
@@ -276,65 +289,34 @@ async def get_model_chats(
     end_date: Optional[int] = Query(None),
     skip: int = Query(0),
     limit: int = Query(50, le=100),
+    order_by: str = Query('updated_at'),
+    direction: str = Query('desc'),
     user=Depends(get_admin_user),
     db: AsyncSession = Depends(get_async_session),
 ):
     """Get chats that used a specific model, with preview and feedback info."""
+    filter = {}
+    if start_date:
+        filter['start_date'] = start_date
+    if end_date:
+        filter['end_date'] = end_date
+    if order_by in MODEL_CHAT_ORDER_FIELDS:
+        filter['order_by'] = order_by
+    if direction in {'asc', 'desc'}:
+        filter['direction'] = direction
 
-    # Get chat IDs that used this model
-    chat_ids = await ChatMessages.get_chat_ids_by_model_id(
+    result = await Chats.get_chats_by_model_id(
         model_id=model_id,
-        start_date=start_date,
-        end_date=end_date,
+        filter=filter,
         skip=skip,
         limit=limit,
         db=db,
     )
 
-    if not chat_ids:
-        return ModelChatsResponse(chats=[], total=0)
-
-    # Get chat details from messages only
-    chats_data = []
-    for chat_id in chat_ids:
-        messages = await ChatMessages.get_messages_by_chat_id(chat_id, db=db)
-        if not messages:
-            continue
-
-        # Get user_id from first user message
-        first_user_msg = next((m for m in messages if m.role == 'user'), None)
-        user_id = first_user_msg.user_id if first_user_msg else None
-
-        # Extract first message content as preview
-        first_message = None
-        if first_user_msg and first_user_msg.content:
-            content = first_user_msg.content
-            if isinstance(content, str):
-                first_message = content[:200]
-            elif isinstance(content, list):
-                text_parts = [b.get('text', '') for b in content if isinstance(b, dict)]
-                first_message = ' '.join(text_parts)[:200]
-
-        # Get user info
-        user_name = None
-        if user_id:
-            user_info = await Users.get_user_by_id(user_id, db=db)
-            user_name = user_info.name if user_info else None
-
-        # Timestamps from messages
-        updated_at = max(m.created_at for m in messages) if messages else 0
-
-        chats_data.append(
-            ModelChatEntry(
-                chat_id=chat_id,
-                user_id=user_id,
-                user_name=user_name,
-                first_message=first_message,
-                updated_at=updated_at,
-            )
-        )
-
-    return ModelChatsResponse(chats=chats_data, total=len(chats_data))
+    return ModelChatsResponse(
+        chats=[ModelChatEntry.model_validate(chat) for chat in result['items']],
+        total=result['total'] or 0,
+    )
 
 
 ####################
@@ -367,6 +349,12 @@ async def get_model_overview(
 ):
     """Get model overview with feedback history and chat tags."""
 
+    # Calculate start date for history
+    now = datetime.now()
+    start_dt = None
+    if days > 0:
+        start_dt = now - timedelta(days=days)
+
     # Get chat IDs that used this model
     chat_ids = await ChatMessages.get_chat_ids_by_model_id(
         model_id=model_id,
@@ -377,31 +365,18 @@ async def get_model_overview(
         db=db,
     )
 
-    # Get feedback history per day
-    history_counts: dict[str, dict] = defaultdict(lambda: {'won': 0, 'lost': 0})
-
-    # Calculate start date for history
-    now = datetime.now()
-    start_dt = None
-    if days > 0:
-        start_dt = now - timedelta(days=days)
-
-    for chat_id in chat_ids:
-        feedbacks = await Feedbacks.get_feedbacks_by_chat_id(chat_id, db=db)
-        for fb in feedbacks:
-            if fb.data and 'rating' in fb.data:
-                rating = fb.data['rating']
-                fb_date = datetime.fromtimestamp(fb.created_at)
-
-                # Filter by date range
-                if start_dt and fb_date < start_dt:
-                    continue
-
-                date_str = fb_date.strftime('%Y-%m-%d')
-                if rating == 1:
-                    history_counts[date_str]['won'] += 1
-                elif rating == -1:
-                    history_counts[date_str]['lost'] += 1
+    history_rows = await Feedbacks.get_model_feedback_counts_by_day(
+        model_id=model_id,
+        start_date=int(start_dt.timestamp()) if start_dt else None,
+        db=db,
+    )
+    history_counts = {
+        entry.date: {
+            'won': entry.won,
+            'lost': entry.lost,
+        }
+        for entry in history_rows
+    }
 
     # Fill in missing days
     history = []
@@ -430,10 +405,14 @@ async def get_model_overview(
 
     # Get chat tags
     tag_counts: dict[str, int] = defaultdict(int)
-    for chat_id in chat_ids:
-        chat = await Chats.get_chat_by_id(chat_id, db=db)
-        if chat and chat.meta:
-            for tag in chat.meta.get('tags', []):
+    if chat_ids:
+        chat_metas = await Chats.get_chat_metas_by_chat_ids(
+            chat_ids,
+            include_archived=True,
+            db=db,
+        )
+        for meta in chat_metas:
+            for tag in meta.get('tags', []):
                 tag_counts[tag] += 1
 
     # Sort by count and take top 10

backend/open_webui/routers/audio.py

@@ -28,6 +28,8 @@ from fastapi import (
 )
 from fastapi.responses import FileResponse
 from pydantic import BaseModel
+
+# pydub needs stdlib audioop (gone in 3.13); keep requires-python capped < 3.13
 from pydub import AudioSegment
 from pydub.silence import split_on_silence
 from pydub.utils import mediainfo
@@ -47,11 +49,14 @@ from open_webui.env import (
     AIOHTTP_CLIENT_SESSION_SSL,
     AIOHTTP_CLIENT_TIMEOUT,
     AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST,
+    AIOHTTP_FILE_STREAM_CHUNK_SIZE,
     BYPASS_PYDUB_PREPROCESSING,
     DEVICE_TYPE,
     ENABLE_FORWARD_USER_INFO_HEADERS,
     ENV,
 )
+from open_webui.events import EVENTS, publish_event
+from open_webui.models.config import Config
 from open_webui.utils.access_control import has_permission
 from open_webui.utils.auth import get_admin_user, get_verified_user
 from open_webui.utils.headers import include_user_info_headers
@@ -71,6 +76,51 @@ AZURE_MAX_FILE_SIZE: int = AZURE_MAX_FILE_SIZE_MB * 1024 * 1024
 SPEECH_CACHE_DIR = CACHE_DIR / 'audio' / 'speech'
 SPEECH_CACHE_DIR.mkdir(parents=True, exist_ok=True)
 
+TTS_CONFIG_KEYS = {
+    'OPENAI_API_BASE_URL': 'audio.tts.openai.api_base_url',
+    'OPENAI_API_KEY': 'audio.tts.openai.api_key',
+    'OPENAI_PARAMS': 'audio.tts.openai.params',
+    'API_KEY': 'audio.tts.api_key',
+    'ENGINE': 'audio.tts.engine',
+    'MODEL': 'audio.tts.model',
+    'VOICE': 'audio.tts.voice',
+    'SPLIT_ON': 'audio.tts.split_on',
+    'AZURE_SPEECH_REGION': 'audio.tts.azure.speech_region',
+    'AZURE_SPEECH_BASE_URL': 'audio.tts.azure.speech_base_url',
+    'AZURE_SPEECH_OUTPUT_FORMAT': 'audio.tts.azure.speech_output_format',
+    'MISTRAL_API_KEY': 'audio.tts.mistral.api_key',
+    'MISTRAL_API_BASE_URL': 'audio.tts.mistral.api_base_url',
+}
+
+STT_CONFIG_KEYS = {
+    'OPENAI_API_BASE_URL': 'audio.stt.openai.api_base_url',
+    'OPENAI_API_KEY': 'audio.stt.openai.api_key',
+    'OPENAI_API_REQUEST_FORMAT': 'audio.stt.openai.api_request_format',
+    'ENGINE': 'audio.stt.engine',
+    'MODEL': 'audio.stt.model',
+    'SUPPORTED_CONTENT_TYPES': 'audio.stt.supported_content_types',
+    'ALLOWED_EXTENSIONS': 'audio.stt.allowed_extensions',
+    'WHISPER_MODEL': 'audio.stt.whisper_model',
+    'DEEPGRAM_API_KEY': 'audio.stt.deepgram.api_key',
+    'AZURE_API_KEY': 'audio.stt.azure.api_key',
+    'AZURE_REGION': 'audio.stt.azure.region',
+    'AZURE_LOCALES': 'audio.stt.azure.locales',
+    'AZURE_BASE_URL': 'audio.stt.azure.base_url',
+    'AZURE_MAX_SPEAKERS': 'audio.stt.azure.max_speakers',
+    'MISTRAL_API_KEY': 'audio.stt.mistral.api_key',
+    'MISTRAL_API_BASE_URL': 'audio.stt.mistral.api_base_url',
+    'MISTRAL_USE_CHAT_COMPLETIONS': 'audio.stt.mistral.use_chat_completions',
+}
+
+
+async def get_config_values(key_map: dict[str, str]) -> dict:
+    values = await Config.get_many(*key_map.values())
+    return {field: values[storage_key] for field, storage_key in key_map.items() if storage_key in values}
+
+
+def config_updates(data: dict, key_map: dict[str, str]) -> dict:
+    return {key_map[field]: value for field, value in data.items() if field in key_map}
+
 
 def is_audio_conversion_required(file_path):
     """
@@ -204,6 +254,7 @@ class TTSConfigForm(BaseModel):
 class STTConfigForm(BaseModel):
     OPENAI_API_BASE_URL: str
     OPENAI_API_KEY: str
+    OPENAI_API_REQUEST_FORMAT: str = 'multipart'
     ENGINE: str
     MODEL: str
     SUPPORTED_CONTENT_TYPES: list[str] = []
@@ -228,119 +279,39 @@ class AudioConfigUpdateForm(BaseModel):
 @router.get('/config')
 async def get_audio_config(request: Request, user=Depends(get_admin_user)):
     return {
-        'tts': {
-            'OPENAI_API_BASE_URL': request.app.state.config.TTS_OPENAI_API_BASE_URL,
-            'OPENAI_API_KEY': request.app.state.config.TTS_OPENAI_API_KEY,
-            'OPENAI_PARAMS': request.app.state.config.TTS_OPENAI_PARAMS,
-            'API_KEY': request.app.state.config.TTS_API_KEY,
-            'ENGINE': request.app.state.config.TTS_ENGINE,
-            'MODEL': request.app.state.config.TTS_MODEL,
-            'VOICE': request.app.state.config.TTS_VOICE,
-            'SPLIT_ON': request.app.state.config.TTS_SPLIT_ON,
-            'AZURE_SPEECH_REGION': request.app.state.config.TTS_AZURE_SPEECH_REGION,
-            'AZURE_SPEECH_BASE_URL': request.app.state.config.TTS_AZURE_SPEECH_BASE_URL,
-            'AZURE_SPEECH_OUTPUT_FORMAT': request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
-            'MISTRAL_API_KEY': request.app.state.config.TTS_MISTRAL_API_KEY,
-            'MISTRAL_API_BASE_URL': request.app.state.config.TTS_MISTRAL_API_BASE_URL,
-        },
-        'stt': {
-            'OPENAI_API_BASE_URL': request.app.state.config.STT_OPENAI_API_BASE_URL,
-            'OPENAI_API_KEY': request.app.state.config.STT_OPENAI_API_KEY,
-            'ENGINE': request.app.state.config.STT_ENGINE,
-            'MODEL': request.app.state.config.STT_MODEL,
-            'SUPPORTED_CONTENT_TYPES': request.app.state.config.STT_SUPPORTED_CONTENT_TYPES,
-            'ALLOWED_EXTENSIONS': request.app.state.config.STT_ALLOWED_EXTENSIONS,
-            'WHISPER_MODEL': request.app.state.config.WHISPER_MODEL,
-            'DEEPGRAM_API_KEY': request.app.state.config.DEEPGRAM_API_KEY,
-            'AZURE_API_KEY': request.app.state.config.AUDIO_STT_AZURE_API_KEY,
-            'AZURE_REGION': request.app.state.config.AUDIO_STT_AZURE_REGION,
-            'AZURE_LOCALES': request.app.state.config.AUDIO_STT_AZURE_LOCALES,
-            'AZURE_BASE_URL': request.app.state.config.AUDIO_STT_AZURE_BASE_URL,
-            'AZURE_MAX_SPEAKERS': request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS,
-            'MISTRAL_API_KEY': request.app.state.config.AUDIO_STT_MISTRAL_API_KEY,
-            'MISTRAL_API_BASE_URL': request.app.state.config.AUDIO_STT_MISTRAL_API_BASE_URL,
-            'MISTRAL_USE_CHAT_COMPLETIONS': request.app.state.config.AUDIO_STT_MISTRAL_USE_CHAT_COMPLETIONS,
-        },
+        'tts': await get_config_values(TTS_CONFIG_KEYS),
+        'stt': await get_config_values(STT_CONFIG_KEYS),
     }
 
 
 @router.post('/config/update')
 async def update_audio_config(request: Request, form_data: AudioConfigUpdateForm, user=Depends(get_admin_user)):
-    # TTS settings
-    request.app.state.config.TTS_OPENAI_API_BASE_URL = form_data.tts.OPENAI_API_BASE_URL
-    request.app.state.config.TTS_OPENAI_API_KEY = form_data.tts.OPENAI_API_KEY
-    request.app.state.config.TTS_OPENAI_PARAMS = form_data.tts.OPENAI_PARAMS
-    request.app.state.config.TTS_API_KEY = form_data.tts.API_KEY
-    request.app.state.config.TTS_ENGINE = form_data.tts.ENGINE
-    request.app.state.config.TTS_MODEL = form_data.tts.MODEL
-    request.app.state.config.TTS_VOICE = form_data.tts.VOICE
-    request.app.state.config.TTS_SPLIT_ON = form_data.tts.SPLIT_ON
-    request.app.state.config.TTS_AZURE_SPEECH_REGION = form_data.tts.AZURE_SPEECH_REGION
-    request.app.state.config.TTS_AZURE_SPEECH_BASE_URL = form_data.tts.AZURE_SPEECH_BASE_URL
-    request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = form_data.tts.AZURE_SPEECH_OUTPUT_FORMAT
-    request.app.state.config.TTS_MISTRAL_API_KEY = form_data.tts.MISTRAL_API_KEY
-    request.app.state.config.TTS_MISTRAL_API_BASE_URL = form_data.tts.MISTRAL_API_BASE_URL
+    await Config.upsert(
+        {
+            **config_updates(form_data.tts.model_dump(), TTS_CONFIG_KEYS),
+            **config_updates(form_data.stt.model_dump(), STT_CONFIG_KEYS),
+        }
+    )
 
-    # STT settings
-    request.app.state.config.STT_OPENAI_API_BASE_URL = form_data.stt.OPENAI_API_BASE_URL
-    request.app.state.config.STT_OPENAI_API_KEY = form_data.stt.OPENAI_API_KEY
-    request.app.state.config.STT_ENGINE = form_data.stt.ENGINE
-    request.app.state.config.STT_MODEL = form_data.stt.MODEL
-    request.app.state.config.STT_SUPPORTED_CONTENT_TYPES = form_data.stt.SUPPORTED_CONTENT_TYPES
-    request.app.state.config.STT_ALLOWED_EXTENSIONS = form_data.stt.ALLOWED_EXTENSIONS
-    request.app.state.config.WHISPER_MODEL = form_data.stt.WHISPER_MODEL
-    request.app.state.config.DEEPGRAM_API_KEY = form_data.stt.DEEPGRAM_API_KEY
-    request.app.state.config.AUDIO_STT_AZURE_API_KEY = form_data.stt.AZURE_API_KEY
-    request.app.state.config.AUDIO_STT_AZURE_REGION = form_data.stt.AZURE_REGION
-    request.app.state.config.AUDIO_STT_AZURE_LOCALES = form_data.stt.AZURE_LOCALES
-    request.app.state.config.AUDIO_STT_AZURE_BASE_URL = form_data.stt.AZURE_BASE_URL
-    request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS = form_data.stt.AZURE_MAX_SPEAKERS
-    request.app.state.config.AUDIO_STT_MISTRAL_API_KEY = form_data.stt.MISTRAL_API_KEY
-    request.app.state.config.AUDIO_STT_MISTRAL_API_BASE_URL = form_data.stt.MISTRAL_API_BASE_URL
-    request.app.state.config.AUDIO_STT_MISTRAL_USE_CHAT_COMPLETIONS = form_data.stt.MISTRAL_USE_CHAT_COMPLETIONS
-
-    if request.app.state.config.STT_ENGINE == '':
-        request.app.state.faster_whisper_model = set_faster_whisper_model(
-            form_data.stt.WHISPER_MODEL, WHISPER_MODEL_AUTO_UPDATE
+    if form_data.stt.ENGINE == '':
+        request.app.state.faster_whisper_model = await asyncio.to_thread(
+            set_faster_whisper_model, form_data.stt.WHISPER_MODEL, WHISPER_MODEL_AUTO_UPDATE
         )
     else:
         request.app.state.faster_whisper_model = None
 
-    return {
-        'tts': {
-            'ENGINE': request.app.state.config.TTS_ENGINE,
-            'MODEL': request.app.state.config.TTS_MODEL,
-            'VOICE': request.app.state.config.TTS_VOICE,
-            'OPENAI_API_BASE_URL': request.app.state.config.TTS_OPENAI_API_BASE_URL,
-            'OPENAI_API_KEY': request.app.state.config.TTS_OPENAI_API_KEY,
-            'OPENAI_PARAMS': request.app.state.config.TTS_OPENAI_PARAMS,
-            'API_KEY': request.app.state.config.TTS_API_KEY,
-            'SPLIT_ON': request.app.state.config.TTS_SPLIT_ON,
-            'AZURE_SPEECH_REGION': request.app.state.config.TTS_AZURE_SPEECH_REGION,
-            'AZURE_SPEECH_BASE_URL': request.app.state.config.TTS_AZURE_SPEECH_BASE_URL,
-            'AZURE_SPEECH_OUTPUT_FORMAT': request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
-            'MISTRAL_API_KEY': request.app.state.config.TTS_MISTRAL_API_KEY,
-            'MISTRAL_API_BASE_URL': request.app.state.config.TTS_MISTRAL_API_BASE_URL,
+    config = await get_audio_config(request, user)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_UPDATED,
+        actor=user,
+        subject_id='audio',
+        data={
+            'tts_engine': config.get('tts', {}).get('ENGINE'),
+            'stt_engine': config.get('stt', {}).get('ENGINE'),
         },
-        'stt': {
-            'OPENAI_API_BASE_URL': request.app.state.config.STT_OPENAI_API_BASE_URL,
-            'OPENAI_API_KEY': request.app.state.config.STT_OPENAI_API_KEY,
-            'ENGINE': request.app.state.config.STT_ENGINE,
-            'MODEL': request.app.state.config.STT_MODEL,
-            'SUPPORTED_CONTENT_TYPES': request.app.state.config.STT_SUPPORTED_CONTENT_TYPES,
-            'ALLOWED_EXTENSIONS': request.app.state.config.STT_ALLOWED_EXTENSIONS,
-            'WHISPER_MODEL': request.app.state.config.WHISPER_MODEL,
-            'DEEPGRAM_API_KEY': request.app.state.config.DEEPGRAM_API_KEY,
-            'AZURE_API_KEY': request.app.state.config.AUDIO_STT_AZURE_API_KEY,
-            'AZURE_REGION': request.app.state.config.AUDIO_STT_AZURE_REGION,
-            'AZURE_LOCALES': request.app.state.config.AUDIO_STT_AZURE_LOCALES,
-            'AZURE_BASE_URL': request.app.state.config.AUDIO_STT_AZURE_BASE_URL,
-            'AZURE_MAX_SPEAKERS': request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS,
-            'MISTRAL_API_KEY': request.app.state.config.AUDIO_STT_MISTRAL_API_KEY,
-            'MISTRAL_API_BASE_URL': request.app.state.config.AUDIO_STT_MISTRAL_API_BASE_URL,
-            'MISTRAL_USE_CHAT_COMPLETIONS': request.app.state.config.AUDIO_STT_MISTRAL_USE_CHAT_COMPLETIONS,
-        },
-    }
+    )
+    return config
 
 
 def load_speech_pipeline(request):
@@ -388,14 +359,16 @@ async def _write_tts_cache(
 
 async def _tts_openai(request, payload, file_path, file_body_path, user):
     """Generate speech via an OpenAI-compatible TTS endpoint."""
-    payload['model'] = request.app.state.config.TTS_MODEL
+    payload['model'] = await Config.get('audio.tts.model')
     if not payload.get('voice'):
-        payload['voice'] = request.app.state.config.TTS_VOICE
-    payload = {**payload, **(request.app.state.config.TTS_OPENAI_PARAMS or {})}
+        payload['voice'] = await Config.get('audio.tts.voice')
+    payload = {**payload, **(await Config.get('audio.tts.openai.params') or {})}
+    api_key = await Config.get('audio.tts.openai.api_key')
+    api_base_url = await Config.get('audio.tts.openai.api_base_url')
 
     headers = {
         'Content-Type': 'application/json',
-        'Authorization': f'Bearer {request.app.state.config.TTS_OPENAI_API_KEY}',
+        'Authorization': f'Bearer {api_key}',
     }
     if ENABLE_FORWARD_USER_INFO_HEADERS:
         headers = include_user_info_headers(headers, user)
@@ -404,7 +377,7 @@ async def _tts_openai(request, payload, file_path, file_body_path, user):
     try:
         session = await get_session()
         r = await session.post(
-            url=f'{request.app.state.config.TTS_OPENAI_API_BASE_URL}/audio/speech',
+            url=f'{api_base_url}/audio/speech',
             json=payload,
             headers=headers,
             ssl=AIOHTTP_CLIENT_SESSION_SSL,
@@ -429,8 +402,12 @@ async def _tts_openai(request, payload, file_path, file_body_path, user):
 
 async def _tts_elevenlabs(request, payload, file_path, file_body_path, user):
     """Generate speech via the ElevenLabs TTS API."""
-    voice_id = payload.get('voice', '')
-    if voice_id not in await get_available_voices(request):
+    voice_id = (payload.get('voice') or '').strip()
+    if not voice_id:
+        raise HTTPException(status_code=400, detail='Invalid voice id')
+
+    available_voices = await get_available_voices(request)
+    if available_voices and voice_id not in available_voices:
         raise HTTPException(status_code=400, detail='Invalid voice id')
 
     r = None
@@ -440,13 +417,13 @@ async def _tts_elevenlabs(request, payload, file_path, file_body_path, user):
             f'{ELEVENLABS_API_BASE_URL}/v1/text-to-speech/{voice_id}',
             json={
                 'text': payload['input'],
-                'model_id': request.app.state.config.TTS_MODEL,
+                'model_id': await Config.get('audio.tts.model'),
                 'voice_settings': {'stability': 0.5, 'similarity_boost': 0.5},
             },
             headers={
                 'Accept': 'audio/mpeg',
                 'Content-Type': 'application/json',
-                'xi-api-key': request.app.state.config.TTS_API_KEY,
+                'xi-api-key': await Config.get('audio.tts.api_key'),
             },
             ssl=AIOHTTP_CLIENT_SESSION_SSL,
         ) as r:
@@ -460,15 +437,15 @@ async def _tts_elevenlabs(request, payload, file_path, file_body_path, user):
 
 async def _tts_azure(request, payload, file_path, file_body_path, user):
     """Generate speech via Azure Cognitive Services TTS."""
-    az_region = request.app.state.config.TTS_AZURE_SPEECH_REGION or 'eastus'
-    az_base = request.app.state.config.TTS_AZURE_SPEECH_BASE_URL
-    language = payload.get('voice') or request.app.state.config.TTS_VOICE
+    az_region = await Config.get('audio.tts.azure.speech_region') or 'eastus'
+    az_base = await Config.get('audio.tts.azure.speech_base_url')
+    language = payload.get('voice') or await Config.get('audio.tts.voice')
     locale = '-'.join(language.split('-')[:2])
-    output_format = request.app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT
+    output_format = await Config.get('audio.tts.azure.speech_output_format')
 
     ssml = (
-        f'<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="{locale}">'
-        f'<voice name="{language}">{html.escape(payload["input"])}</voice>'
+        f'<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="{html.escape(locale)}">'
+        f'<voice name="{html.escape(language)}">{html.escape(payload["input"])}</voice>'
         f'</speak>'
     )
 
@@ -478,7 +455,7 @@ async def _tts_azure(request, payload, file_path, file_body_path, user):
         async with session.post(
             (az_base or f'https://{az_region}.tts.speech.microsoft.com') + '/cognitiveservices/v1',
             headers={
-                'Ocp-Apim-Subscription-Key': request.app.state.config.TTS_API_KEY,
+                'Ocp-Apim-Subscription-Key': await Config.get('audio.tts.api_key'),
                 'Content-Type': 'application/ssml+xml',
                 'X-Microsoft-OutputFormat': output_format,
             },
@@ -498,10 +475,10 @@ async def _tts_transformers(request, payload, file_path, file_body_path, user):
     import soundfile as sf
     import torch
 
-    load_speech_pipeline(request)
+    await asyncio.to_thread(load_speech_pipeline, request)
 
     embeddings = request.app.state.speech_speaker_embeddings_dataset
-    model_name = request.app.state.config.TTS_MODEL
+    model_name = await Config.get('audio.tts.model')
 
     idx = 6799
     try:
@@ -529,8 +506,8 @@ async def _tts_transformers(request, payload, file_path, file_body_path, user):
 
 async def _tts_mistral(request, payload, file_path, file_body_path, user):
     """Generate speech via the Mistral TTS API."""
-    api_key = request.app.state.config.TTS_MISTRAL_API_KEY
-    api_base_url = request.app.state.config.TTS_MISTRAL_API_BASE_URL or 'https://api.mistral.ai/v1'
+    api_key = await Config.get('audio.tts.mistral.api_key')
+    api_base_url = await Config.get('audio.tts.mistral.api_base_url') or 'https://api.mistral.ai/v1'
 
     if not api_key:
         raise HTTPException(status_code=400, detail='Mistral API key is required for Mistral TTS')
@@ -542,7 +519,7 @@ async def _tts_mistral(request, payload, file_path, file_body_path, user):
             url=f'{api_base_url}/audio/speech',
             json={
                 'input': payload.get('input', ''),  # text to synthesize
-                'model': request.app.state.config.TTS_MODEL or 'voxtral-mini-tts-2603',
+                'model': await Config.get('audio.tts.model') or 'voxtral-mini-tts-2603',
                 'voice_id': payload.get('voice', ''),
                 'response_format': 'mp3',
             },
@@ -578,16 +555,14 @@ _TTS_ENGINES = {
 
 @router.post('/speech')
 async def speech(request: Request, user=Depends(get_verified_user)):
-    engine = request.app.state.config.TTS_ENGINE
+    engine = await Config.get('audio.tts.engine')
     if engine == '':
         raise HTTPException(
             status_code=status.HTTP_404_NOT_FOUND,
             detail=ERROR_MESSAGES.NOT_FOUND,
         )
 
-    if user.role != 'admin' and not await has_permission(
-        user.id, 'chat.tts', request.app.state.config.USER_PERMISSIONS
-    ):
+    if user.role != 'admin' and not await has_permission(user.id, 'chat.tts', await Config.get('user.permissions')):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
             detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
@@ -595,7 +570,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
 
     body = await request.body()
     name = hashlib.sha256(
-        body + str(engine).encode('utf-8') + str(request.app.state.config.TTS_MODEL).encode('utf-8')
+        body + str(engine).encode('utf-8') + str(await Config.get('audio.tts.model')).encode('utf-8')
     ).hexdigest()
 
     file_path = SPEECH_CACHE_DIR.joinpath(f'{name}.mp3')
@@ -603,6 +578,13 @@ async def speech(request: Request, user=Depends(get_verified_user)):
 
     # Return cached result if available
     if file_path.is_file():
+        await publish_event(
+            request,
+            EVENTS.AUDIO_SPEECH_REQUESTED,
+            actor=user,
+            subject_id=name,
+            data={'engine': engine, 'cached': True},
+        )
         return FileResponse(file_path)
 
     try:
@@ -615,12 +597,27 @@ async def speech(request: Request, user=Depends(get_verified_user)):
     if handler is None:
         raise HTTPException(status_code=400, detail=f'Unsupported TTS engine: {engine}')
 
-    return await handler(request, payload, file_path, file_body_path, user)
+    response = await handler(request, payload, file_path, file_body_path, user)
+    await publish_event(
+        request,
+        EVENTS.AUDIO_SPEECH_REQUESTED,
+        actor=user,
+        subject_id=name,
+        data={
+            'engine': engine,
+            'model': payload.get('model'),
+            'input_preview': str(payload.get('input', ''))[:300],
+            'cached': False,
+        },
+    )
+    return response
 
 
 async def _transcribe_whisper(request, file_path, languages, file_dir, id):
     if request.app.state.faster_whisper_model is None:
-        request.app.state.faster_whisper_model = set_faster_whisper_model(request.app.state.config.WHISPER_MODEL)
+        request.app.state.faster_whisper_model = await asyncio.to_thread(
+            set_faster_whisper_model, await Config.get('audio.stt.whisper_model')
+        )
 
     model = request.app.state.faster_whisper_model
 
@@ -650,26 +647,51 @@ async def _transcribe_openai(request, file_path, filename, languages, file_dir,
     r = None
     try:
         session = await get_session()
+        api_key = await Config.get('audio.stt.openai.api_key')
+        api_base_url = await Config.get('audio.stt.openai.api_base_url')
+        request_format = (await Config.get('audio.stt.openai.api_request_format') or 'multipart').lower()
+
+        headers = {'Authorization': f'Bearer {api_key}'}
+        if user and ENABLE_FORWARD_USER_INFO_HEADERS:
+            headers = include_user_info_headers(headers, user)
+
         for language in languages:
-            payload = {'model': request.app.state.config.STT_MODEL}
+            payload = {'model': await Config.get('audio.stt.model')}
             if language:
                 payload['language'] = language
 
-            headers = {'Authorization': f'Bearer {request.app.state.config.STT_OPENAI_API_KEY}'}
-            if user and ENABLE_FORWARD_USER_INFO_HEADERS:
-                headers = include_user_info_headers(headers, user)
+            if request_format == 'json':
+                ext = os.path.splitext(filename)[1].lower().lstrip('.') or 'wav'
+                async with aiofiles.open(file_path, 'rb') as f:
+                    payload['input_audio'] = {
+                        'data': base64.b64encode(await f.read()).decode('utf-8'),
+                        'format': 'ogg' if ext == 'oga' else ext,
+                    }
 
-            form_data = aiohttp.FormData()
-            for key, value in payload.items():
-                form_data.add_field(key, str(value))
-            form_data.add_field('file', open(file_path, 'rb'), filename=filename)
+                r = await session.post(
+                    url=f'{api_base_url}/audio/transcriptions',
+                    headers={**headers, 'Content-Type': 'application/json'},
+                    json=payload,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
+                )
+            else:
+                form_data = aiohttp.FormData()
+                for key, value in payload.items():
+                    form_data.add_field(key, str(value))
 
-            r = await session.post(
-                url=f'{request.app.state.config.STT_OPENAI_API_BASE_URL}/audio/transcriptions',
-                headers=headers,
-                data=form_data,
-                ssl=AIOHTTP_CLIENT_SESSION_SSL,
-            )
+                async def audio_chunks():
+                    async with aiofiles.open(file_path, 'rb') as audio_file:
+                        while chunk := await audio_file.read(AIOHTTP_FILE_STREAM_CHUNK_SIZE):
+                            yield chunk
+
+                form_data.add_field('file', audio_chunks(), filename=filename)
+
+                r = await session.post(
+                    url=f'{api_base_url}/audio/transcriptions',
+                    headers=headers,
+                    data=form_data,
+                    ssl=AIOHTTP_CLIENT_SESSION_SSL,
+                )
             if r.status == 200:
                 break
 
@@ -699,8 +721,8 @@ async def _transcribe_deepgram(request, file_path, languages, file_dir, id):
     async with aiofiles.open(file_path, 'rb') as f:
         audio_bytes = await f.read()
 
-    api_key = request.app.state.config.DEEPGRAM_API_KEY
-    stt_model = request.app.state.config.STT_MODEL
+    api_key = await Config.get('audio.stt.deepgram.api_key')
+    stt_model = await Config.get('audio.stt.model')
 
     r = None
     try:
@@ -767,11 +789,11 @@ async def _transcribe_azure(request, file_path, filename, file_dir, id):
             detail=f'File size ({audio_size // (1024 * 1024)}MB) exceeds Azure limit of {AZURE_MAX_FILE_SIZE_MB}MB',
         )
 
-    api_key = request.app.state.config.AUDIO_STT_AZURE_API_KEY
-    region = request.app.state.config.AUDIO_STT_AZURE_REGION or 'eastus'
-    locale_str = request.app.state.config.AUDIO_STT_AZURE_LOCALES
-    base_url = request.app.state.config.AUDIO_STT_AZURE_BASE_URL
-    max_speakers = request.app.state.config.AUDIO_STT_AZURE_MAX_SPEAKERS or 3
+    api_key = await Config.get('audio.stt.azure.api_key')
+    region = await Config.get('audio.stt.azure.region') or 'eastus'
+    locale_str = await Config.get('audio.stt.azure.locales')
+    base_url = await Config.get('audio.stt.azure.base_url')
+    max_speakers = await Config.get('audio.stt.azure.max_speakers') or 3
 
     # Default to a broad set of locales when none are configured
     if len(locale_str) < 2:
@@ -806,13 +828,18 @@ async def _transcribe_azure(request, file_path, filename, file_dir, id):
         base_url or f'https://{region}.api.cognitive.microsoft.com'
     ) + '/speechtotext/transcriptions:transcribe?api-version=2024-11-15'
 
-    form_data = aiohttp.FormData()
-    form_data.add_field('definition', definition)
-    form_data.add_field('audio', open(file_path, 'rb'), filename=filename)
-
     r = None
     try:
         session = await get_session()
+        form_data = aiohttp.FormData()
+        form_data.add_field('definition', definition)
+
+        async def audio_chunks():
+            async with aiofiles.open(file_path, 'rb') as audio_file:
+                while chunk := await audio_file.read(AIOHTTP_FILE_STREAM_CHUNK_SIZE):
+                    yield chunk
+
+        form_data.add_field('audio', audio_chunks(), filename=filename)
         r = await session.post(
             url=endpoint,
             data=form_data,
@@ -881,16 +908,16 @@ async def transcription_handler(request, file_path, metadata, user=None):
         None,  # Always fallback to None in case transcription fails
     ]
 
-    if request.app.state.config.STT_ENGINE == '':
+    if await Config.get('audio.stt.engine') == '':
         return await _transcribe_whisper(request, file_path, languages, file_dir, id)
-    elif request.app.state.config.STT_ENGINE == 'openai':
+    elif await Config.get('audio.stt.engine') == 'openai':
         return await _transcribe_openai(request, file_path, filename, languages, file_dir, id, user)
-    elif request.app.state.config.STT_ENGINE == 'deepgram':
+    elif await Config.get('audio.stt.engine') == 'deepgram':
         return await _transcribe_deepgram(request, file_path, languages, file_dir, id)
-    elif request.app.state.config.STT_ENGINE == 'azure':
+    elif await Config.get('audio.stt.engine') == 'azure':
         return await _transcribe_azure(request, file_path, filename, file_dir, id)
 
-    elif request.app.state.config.STT_ENGINE == 'mistral':
+    elif await Config.get('audio.stt.engine') == 'mistral':
         return await _transcribe_mistral(request, file_path, filename, metadata, file_dir, id)
 
 
@@ -903,16 +930,16 @@ async def _transcribe_mistral(request, file_path, filename, metadata, file_dir,
     if file_size > MAX_FILE_SIZE:
         raise HTTPException(status_code=400, detail=f'File size exceeds limit of {MAX_FILE_SIZE_MB}MB')
 
-    api_key = request.app.state.config.AUDIO_STT_MISTRAL_API_KEY
-    api_base_url = request.app.state.config.AUDIO_STT_MISTRAL_API_BASE_URL or 'https://api.mistral.ai/v1'
-    use_chat_completions = request.app.state.config.AUDIO_STT_MISTRAL_USE_CHAT_COMPLETIONS
+    api_key = await Config.get('audio.stt.mistral.api_key')
+    api_base_url = await Config.get('audio.stt.mistral.api_base_url') or 'https://api.mistral.ai/v1'
+    use_chat_completions = await Config.get('audio.stt.mistral.use_chat_completions')
 
     if not api_key:
         raise HTTPException(status_code=400, detail='Mistral API key is required for Mistral STT')
 
     r = None
     try:
-        model = request.app.state.config.STT_MODEL or 'voxtral-mini-latest'
+        model = await Config.get('audio.stt.model') or 'voxtral-mini-latest'
         log.info(
             f'Mistral STT - model: {model}, method: {"chat_completions" if use_chat_completions else "transcriptions"}'
         )
@@ -985,7 +1012,12 @@ async def _transcribe_mistral(request, file_path, filename, metadata, file_dir,
             if language:
                 form_data.add_field('language', language)
 
-            form_data.add_field('file', open(file_path, 'rb'), filename=filename, content_type=mime_type)
+            async def audio_chunks():
+                async with aiofiles.open(file_path, 'rb') as audio_file:
+                    while chunk := await audio_file.read(AIOHTTP_FILE_STREAM_CHUNK_SIZE):
+                        yield chunk
+
+            form_data.add_field('file', audio_chunks(), filename=filename, content_type=mime_type)
 
             r = await session.post(
                 url=f'{api_base_url}/audio/transcriptions',
@@ -1056,28 +1088,26 @@ async def transcribe(request: Request, file_path: str, metadata: Optional[dict]
             log.exception(e)
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.DEFAULT(e),
+                detail=ERROR_MESSAGES.DEFAULT(e, 'Error processing audio file'),
             )
 
-    results = []
     try:
         tasks = [transcription_handler(request, chunk_path, metadata, user) for chunk_path in chunk_paths]
-        for coro in asyncio.as_completed(tasks):
-            try:
-                results.append(await coro)
-            except HTTPException:
-                raise
-            except Exception as transcribe_exc:
-                raise HTTPException(
-                    status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-                    detail=f'Error transcribing chunk: {transcribe_exc}',
-                )
+        # gather keeps results in chunk order, unlike as_completed
+        results = await asyncio.gather(*tasks)
+    except HTTPException:
+        raise
+    except Exception as transcribe_exc:
+        raise HTTPException(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            detail=f'Error transcribing chunk: {transcribe_exc}',
+        )
     finally:
         # Clean up only the temporary chunks, never the original file
         for chunk_path in chunk_paths:
             if chunk_path != file_path and os.path.isfile(chunk_path):
                 try:
-                    os.remove(chunk_path)
+                    await asyncio.to_thread(os.remove, chunk_path)
                 except Exception:
                     pass
 
@@ -1153,15 +1183,13 @@ async def transcription(
     language: Optional[str] = Form(None),
     user=Depends(get_verified_user),
 ):
-    if user.role != 'admin' and not await has_permission(
-        user.id, 'chat.stt', request.app.state.config.USER_PERMISSIONS
-    ):
+    if user.role != 'admin' and not await has_permission(user.id, 'chat.stt', await Config.get('user.permissions')):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
             detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
         )
     log.info(f'file.content_type: {file.content_type}')
-    stt_supported_content_types = getattr(request.app.state.config, 'STT_SUPPORTED_CONTENT_TYPES', [])
+    stt_supported_content_types = await Config.get('audio.stt.supported_content_types', [])
 
     if not strict_match_mime_type(stt_supported_content_types, file.content_type):
         raise HTTPException(
@@ -1173,7 +1201,7 @@ async def transcription(
         safe_name = os.path.basename(file.filename) if file.filename else ''
         ext = safe_name.rsplit('.', 1)[-1].lower() if '.' in safe_name else ''
 
-        allowed_extensions = getattr(request.app.state.config, 'STT_ALLOWED_EXTENSIONS', [])
+        allowed_extensions = await Config.get('audio.stt.allowed_extensions', [])
         if allowed_extensions and ext not in allowed_extensions:
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
@@ -1193,8 +1221,8 @@ async def transcription(
         if not os.path.realpath(file_path).startswith(os.path.realpath(file_dir)):
             raise ValueError('Invalid file path detected')
 
-        with open(file_path, 'wb') as f:
-            f.write(contents)
+        async with aiofiles.open(file_path, 'wb') as f:
+            await f.write(contents)
 
         try:
             metadata = None
@@ -1204,6 +1232,17 @@ async def transcription(
 
             result = await transcribe(request, file_path, metadata, user)
 
+            await publish_event(
+                request,
+                EVENTS.AUDIO_TRANSCRIPTION_REQUESTED,
+                actor=user,
+                subject_id=str(id),
+                data={
+                    'filename': safe_name,
+                    'content_type': file.content_type,
+                    'language': language,
+                },
+            )
             return {
                 **result,
                 'filename': os.path.basename(file_path),
@@ -1233,11 +1272,11 @@ async def transcription(
 async def get_available_models(request: Request) -> list[dict]:
     """Return the list of available TTS models for the configured engine."""
     available_models = []
-    engine = request.app.state.config.TTS_ENGINE
+    engine = await Config.get('audio.tts.engine')
     _timeout = aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST)
 
     if engine == 'openai':
-        base_url = request.app.state.config.TTS_OPENAI_API_BASE_URL
+        base_url = await Config.get('audio.tts.openai.api_base_url')
         if not base_url.startswith('https://api.openai.com'):
             session = await get_session()
             try:
@@ -1272,7 +1311,7 @@ async def get_available_models(request: Request) -> list[dict]:
             async with session.get(
                 f'{ELEVENLABS_API_BASE_URL}/v1/models',
                 headers={
-                    'xi-api-key': request.app.state.config.TTS_API_KEY,
+                    'xi-api-key': await Config.get('audio.tts.api_key'),
                     'Content-Type': 'application/json',
                 },
                 ssl=AIOHTTP_CLIENT_SESSION_SSL,
@@ -1307,11 +1346,11 @@ _OPENAI_DEFAULT_VOICES = {
 
 async def get_available_voices(request) -> dict:
     """Return ``{voice_id: voice_name}`` for the configured TTS engine."""
-    engine = request.app.state.config.TTS_ENGINE
+    engine = await Config.get('audio.tts.engine')
     _timeout = aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_MODEL_LIST)
 
     if engine == 'openai':
-        base_url = request.app.state.config.TTS_OPENAI_API_BASE_URL
+        base_url = await Config.get('audio.tts.openai.api_base_url')
         if not base_url.startswith('https://api.openai.com'):
             try:
                 session = await get_session()
@@ -1334,7 +1373,7 @@ async def get_available_voices(request) -> dict:
             async with session.get(
                 f'{ELEVENLABS_API_BASE_URL}/v1/voices',
                 headers={
-                    'xi-api-key': request.app.state.config.TTS_API_KEY,
+                    'xi-api-key': await Config.get('audio.tts.api_key'),
                     'Content-Type': 'application/json',
                 },
                 ssl=AIOHTTP_CLIENT_SESSION_SSL,
@@ -1344,19 +1383,19 @@ async def get_available_voices(request) -> dict:
                 voices_data = await resp.json()
                 return {v['voice_id']: v['name'] for v in voices_data.get('voices', [])}
         except Exception as e:
-            log.error(f'Error fetching ElevenLabs voices: {e}')
+            log.warning(f'Error fetching ElevenLabs voices: {e}')
             return {}
 
     if engine == 'azure':
         try:
-            region = request.app.state.config.TTS_AZURE_SPEECH_REGION
-            base_url = request.app.state.config.TTS_AZURE_SPEECH_BASE_URL
+            region = await Config.get('audio.tts.azure.speech_region')
+            base_url = await Config.get('audio.tts.azure.speech_base_url')
             url = (base_url or f'https://{region}.tts.speech.microsoft.com') + '/cognitiveservices/voices/list'
 
             session = await get_session()
             async with session.get(
                 url,
-                headers={'Ocp-Apim-Subscription-Key': request.app.state.config.TTS_API_KEY},
+                headers={'Ocp-Apim-Subscription-Key': await Config.get('audio.tts.api_key')},
                 ssl=AIOHTTP_CLIENT_SESSION_SSL,
                 timeout=_timeout,
             ) as resp:
@@ -1368,8 +1407,8 @@ async def get_available_voices(request) -> dict:
             return {}
 
     if engine == 'mistral':
-        api_key = request.app.state.config.TTS_MISTRAL_API_KEY
-        api_base_url = request.app.state.config.TTS_MISTRAL_API_BASE_URL or 'https://api.mistral.ai/v1'
+        api_key = await Config.get('audio.tts.mistral.api_key')
+        api_base_url = await Config.get('audio.tts.mistral.api_base_url') or 'https://api.mistral.ai/v1'
         if api_key:
             try:
                 session = await get_session()

backend/open_webui/routers/automations.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.internal.db import get_async_session
 from open_webui.models.automations import (
     AutomationForm,
@@ -14,6 +15,8 @@ from open_webui.models.automations import (
     AutomationRuns,
     Automations,
 )
+from open_webui.models.config import Config
+from open_webui.models.folders import Folders
 from open_webui.utils.access_control import has_permission
 from open_webui.utils.auth import get_admin_user, get_verified_user
 from open_webui.utils.automations import (
@@ -38,13 +41,14 @@ PAGE_ITEM_COUNT = 30
 
 
 async def check_automations_permission(request, user):
-    if not request.app.state.config.ENABLE_AUTOMATIONS:
+    config = await Config.get_many('automations.enable', 'user.permissions')
+    if not config.get('automations.enable'):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
             detail=ERROR_MESSAGES.UNAUTHORIZED,
         )
     if user.role != 'admin' and not await has_permission(
-        user.id, 'features.automations', request.app.state.config.USER_PERMISSIONS
+        user.id, 'features.automations', config.get('user.permissions')
     ):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
@@ -53,16 +57,11 @@ async def check_automations_permission(request, user):
 
 
 def check_automation_access(automation, user):
-    if not automation:
+    if not automation or user.id != automation.user_id:
         raise HTTPException(
             status_code=status.HTTP_404_NOT_FOUND,
             detail=ERROR_MESSAGES.NOT_FOUND,
         )
-    if user.role != 'admin' and user.id != automation.user_id:
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.UNAUTHORIZED,
-        )
 
 
 async def check_automation_limits(request, user, rrule_str: str, db, is_create: bool = False):
@@ -72,7 +71,7 @@ async def check_automation_limits(request, user, rrule_str: str, db, is_create:
 
     # Max count (create only)
     if is_create:
-        max_count = request.app.state.config.AUTOMATION_MAX_COUNT
+        max_count = await Config.get('automations.max_count')
         if max_count:
             max_count = int(max_count)
             if max_count > 0 and await Automations.count_by_user(user.id, db=db) >= max_count:
@@ -82,7 +81,7 @@ async def check_automation_limits(request, user, rrule_str: str, db, is_create:
                 )
 
     # Min interval (create + update)
-    min_interval = request.app.state.config.AUTOMATION_MIN_INTERVAL
+    min_interval = await Config.get('automations.min_interval')
     if min_interval:
         min_interval = int(min_interval)
         if min_interval > 0:
@@ -94,6 +93,17 @@ async def check_automation_limits(request, user, rrule_str: str, db, is_create:
                 )
 
 
+async def check_automation_folder_access(folder_id: Optional[str], user, db: AsyncSession):
+    if folder_id is None:
+        return
+    folder = await Folders.get_folder_by_id_and_user_id(folder_id, user.id, db=db)
+    if not folder:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+
 async def enrich_automation(automation: AutomationModel, db: AsyncSession, tz: str = None) -> AutomationResponse:
     """Full enrichment for single-item views (includes next_runs computation)."""
     last_run = await AutomationRuns.get_latest(automation.id, db=db)
@@ -114,6 +124,7 @@ async def get_automation_items(
     request: Request,
     query: Optional[str] = None,
     status: Optional[str] = None,
+    folder_id: Optional[str] = None,
     page: Optional[int] = 1,
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
@@ -127,6 +138,7 @@ async def get_automation_items(
         user_id=user.id,
         query=query,
         status=status,
+        folder_id=folder_id,
         skip=skip,
         limit=limit,
         db=db,
@@ -161,6 +173,7 @@ async def create_new_automation(
     db: AsyncSession = Depends(get_async_session),
 ):
     await check_automations_permission(request, user)
+    await check_automation_folder_access(form_data.folder_id, user, db)
     try:
         validate_rrule(form_data.data.rrule, tz=user.timezone)
     except ValueError as e:
@@ -173,7 +186,15 @@ async def create_new_automation(
 
     tz = user.timezone
     automation = await Automations.insert(user.id, form_data, next_run_ns(form_data.data.rrule, tz=tz), db=db)
-    return await enrich_automation(automation, db, tz=tz)
+    response = await enrich_automation(automation, db, tz=tz)
+    await publish_event(
+        request,
+        EVENTS.AUTOMATION_CREATED,
+        actor=user,
+        subject_id=automation.id,
+        data={'name': automation.name, 'is_active': automation.is_active, 'folder_id': automation.folder_id},
+    )
+    return response
 
 
 ############################
@@ -210,6 +231,7 @@ async def update_automation_by_id(
     await check_automations_permission(request, user)
     automation = await Automations.get_by_id(id, db=db)
     check_automation_access(automation, user)
+    await check_automation_folder_access(form_data.folder_id, user, db)
 
     try:
         validate_rrule(form_data.data.rrule, tz=user.timezone)
@@ -223,7 +245,15 @@ async def update_automation_by_id(
 
     tz = user.timezone
     updated = await Automations.update_by_id(id, form_data, next_run_ns(form_data.data.rrule, tz=tz), db=db)
-    return await enrich_automation(updated, db, tz=tz)
+    response = await enrich_automation(updated, db, tz=tz)
+    await publish_event(
+        request,
+        EVENTS.AUTOMATION_UPDATED,
+        actor=user,
+        subject_id=updated.id,
+        data={'name': updated.name, 'is_active': updated.is_active, 'folder_id': updated.folder_id},
+    )
+    return response
 
 
 ############################
@@ -242,7 +272,16 @@ async def toggle_automation_by_id(
     automation = await Automations.get_by_id(id, db=db)
     check_automation_access(automation, user)
     toggled = await Automations.toggle(id, next_run_ns(automation.data['rrule'], tz=user.timezone), db=db)
-    return await enrich_automation(toggled, db, tz=user.timezone)
+    response = await enrich_automation(toggled, db, tz=user.timezone)
+    await publish_event(
+        request,
+        EVENTS.AUTOMATION_ENABLED if toggled.is_active else EVENTS.AUTOMATION_DISABLED,
+        actor=user,
+        subject_id=toggled.id,
+        subject_type='automation',
+        data={'name': toggled.name},
+    )
+    return response
 
 
 ############################
@@ -261,6 +300,13 @@ async def run_automation_by_id(
     automation = await Automations.get_by_id(id, db=db)
     check_automation_access(automation, user)
     asyncio.create_task(execute_automation(request.app, automation))
+    await publish_event(
+        request,
+        EVENTS.AUTOMATION_RUN_STARTED,
+        actor=user,
+        subject_id=automation.id,
+        data={'name': automation.name},
+    )
     return await enrich_automation(automation, db, tz=user.timezone)
 
 
@@ -280,7 +326,16 @@ async def delete_automation_by_id(
     automation = await Automations.get_by_id(id, db=db)
     check_automation_access(automation, user)
     await AutomationRuns.delete_by_automation(id, db=db)
-    return await Automations.delete(id, db=db)
+    result = await Automations.delete(id, db=db)
+    if result:
+        await publish_event(
+            request,
+            EVENTS.AUTOMATION_DELETED,
+            actor=user,
+            subject_id=id,
+            data={'name': automation.name},
+        )
+    return result
 
 
 ############################

backend/open_webui/routers/calendar.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.models.access_grants import AccessGrants
 from open_webui.models.calendar import (
     CalendarEventAttendees,
@@ -19,6 +20,7 @@ from open_webui.models.calendar import (
     CalendarUpdateForm,
     RSVPForm,
 )
+from open_webui.models.config import Config
 from open_webui.models.groups import Groups
 from open_webui.models.users import UserModel
 from open_webui.utils.access_control import filter_allowed_access_grants, has_permission
@@ -34,14 +36,13 @@ SCHEDULED_TASKS_CALENDAR_ID = '__scheduled_tasks__'
 
 async def check_calendar_permission(request: Request, user):
     """Check global feature flag AND per-user permission for calendar access."""
-    if not request.app.state.config.ENABLE_CALENDAR:
+    config = await Config.get_many('calendar.enable', 'user.permissions')
+    if not config.get('calendar.enable'):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
             detail=ERROR_MESSAGES.UNAUTHORIZED,
         )
-    if user.role != 'admin' and not await has_permission(
-        user.id, 'features.calendar', request.app.state.config.USER_PERMISSIONS
-    ):
+    if user.role != 'admin' and not await has_permission(user.id, 'features.calendar', config.get('user.permissions')):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
             detail=ERROR_MESSAGES.UNAUTHORIZED,
@@ -50,11 +51,12 @@ async def check_calendar_permission(request: Request, user):
 
 async def _user_has_automations(request: Request, user) -> bool:
     """Check if automations feature is available to this user."""
-    if not getattr(request.app.state.config, 'ENABLE_AUTOMATIONS', False):
+    config = await Config.get_many('automations.enable', 'user.permissions')
+    if not config.get('automations.enable', False):
         return False
     if user.role == 'admin':
         return True
-    return await has_permission(user.id, 'features.automations', request.app.state.config.USER_PERMISSIONS)
+    return await has_permission(user.id, 'features.automations', config.get('user.permissions'))
 
 
 async def _check_calendar_access(calendar_id: str, user: UserModel, permission: str = 'write') -> CalendarModel:
@@ -116,13 +118,21 @@ async def create_calendar(request: Request, form_data: CalendarForm, user: UserM
     # could create a calendar with `principal_id='*' permission='read'|'write'`,
     # making their events readable or writable by any other verified user.
     form_data.access_grants = await filter_allowed_access_grants(
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         user.id,
         user.role,
         form_data.access_grants,
         'sharing.public_calendars',
     )
-    return await Calendars.insert_new_calendar(user.id, form_data)
+    calendar = await Calendars.insert_new_calendar(user.id, form_data)
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_CREATED,
+        actor=user,
+        subject_id=calendar.id,
+        data={'name': calendar.name},
+    )
+    return calendar
 
 
 ####################
@@ -263,7 +273,15 @@ async def get_events(
 async def create_event(request: Request, form_data: CalendarEventForm, user: UserModel = Depends(get_verified_user)):
     await check_calendar_permission(request, user)
     await _check_calendar_access(form_data.calendar_id, user, 'write')
-    return await CalendarEvents.insert_new_event(user.id, form_data)
+    event = await CalendarEvents.insert_new_event(user.id, form_data)
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_EVENT_CREATED,
+        actor=user,
+        subject_id=event.id,
+        data={'calendar_id': event.calendar_id, 'title': event.title},
+    )
+    return event
 
 
 @router.get('/events/search', response_model=CalendarEventListResponse)
@@ -310,6 +328,13 @@ async def update_event(
     updated = await CalendarEvents.update_event_by_id(event_id, form_data)
     if not updated:
         raise HTTPException(status_code=500, detail='Failed to update')
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_EVENT_UPDATED,
+        actor=user,
+        subject_id=updated.id,
+        data={'calendar_id': updated.calendar_id, 'title': updated.title},
+    )
     return updated
 
 
@@ -325,6 +350,13 @@ async def delete_event(request: Request, event_id: str, user: UserModel = Depend
     result = await CalendarEvents.delete_event_by_id(event_id)
     if not result:
         raise HTTPException(status_code=500, detail='Failed to delete')
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_EVENT_DELETED,
+        actor=user,
+        subject_id=event_id,
+        data={'calendar_id': event.calendar_id, 'title': event.title},
+    )
     return {'status': True}
 
 
@@ -340,6 +372,13 @@ async def rsvp_event(
     result = await CalendarEventAttendees.update_rsvp(event_id, user.id, form_data.status)
     if not result:
         raise HTTPException(status_code=404, detail='Not an attendee of this event')
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_EVENT_RSVP_UPDATED,
+        actor=user,
+        subject_id=event_id,
+        data={'status': result.status},
+    )
     return {'status': True, 'rsvp': result.status}
 
 
@@ -373,7 +412,7 @@ async def update_calendar(
     # publicly readable/writable without the corresponding sharing permission.
     if form_data.access_grants is not None:
         form_data.access_grants = await filter_allowed_access_grants(
-            request.app.state.config.USER_PERMISSIONS,
+            await Config.get('user.permissions'),
             user.id,
             user.role,
             form_data.access_grants,
@@ -383,6 +422,13 @@ async def update_calendar(
     updated = await Calendars.update_calendar_by_id(calendar_id, form_data)
     if not updated:
         raise HTTPException(status_code=500, detail='Failed to update')
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_UPDATED,
+        actor=user,
+        subject_id=updated.id,
+        data={'name': updated.name},
+    )
     return updated
 
 
@@ -407,6 +453,13 @@ async def delete_calendar(request: Request, calendar_id: str, user: UserModel =
     result = await Calendars.delete_calendar_by_id(calendar_id)
     if not result:
         raise HTTPException(status_code=500, detail='Failed to delete')
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_DELETED,
+        actor=user,
+        subject_id=calendar_id,
+        data={'name': cal.name},
+    )
     return {'status': True}
 
 
@@ -416,4 +469,11 @@ async def set_default_calendar(request: Request, calendar_id: str, user: UserMod
     cal = await Calendars.set_default_calendar(user.id, calendar_id)
     if not cal:
         raise HTTPException(status_code=404, detail='Calendar not found')
+    await publish_event(
+        request,
+        EVENTS.CALENDAR_DEFAULT_UPDATED,
+        actor=user,
+        subject_id=cal.id,
+        data={'name': cal.name},
+    )
     return cal

backend/open_webui/routers/channels.py

@@ -8,9 +8,11 @@ from fastapi import APIRouter, BackgroundTasks, Depends, HTTPException, Request,
 from fastapi.responses import FileResponse, Response, StreamingResponse
 from open_webui.config import ENABLE_ADMIN_CHAT_ACCESS, ENABLE_ADMIN_EXPORT
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.env import STATIC_DIR
 from open_webui.internal.db import get_async_session
 from open_webui.models.access_grants import AccessGrants, has_public_read_access_grant, has_public_write_access_grant
+from open_webui.models.config import Config
 from open_webui.models.channels import (
     ChannelForm,
     ChannelModel,
@@ -31,9 +33,7 @@ from open_webui.models.messages import (
 from open_webui.models.users import (
     UserIdNameResponse,
     UserIdNameStatusResponse,
-    UserListResponse,
     UserModel,
-    UserModelResponse,
     UserNameResponse,
     Users,
 )
@@ -51,7 +51,6 @@ from open_webui.utils.models import (
     get_all_models,
     get_filtered_models,
 )
-from open_webui.utils.webhook import post_webhook
 from pydantic import BaseModel, field_validator
 from sqlalchemy.ext.asyncio import AsyncSession
 
@@ -125,7 +124,7 @@ def get_channel_permitted_group_and_user_ids(
 
 async def check_channels_access(request: Request, user: Optional[UserModel] = None):
     """Dependency to ensure channels are globally enabled."""
-    if not request.app.state.config.ENABLE_CHANNELS:
+    if not await Config.get('channels.enable'):
         raise HTTPException(
             status_code=status.HTTP_403_FORBIDDEN,
             detail=ERROR_MESSAGES.FEATURE_DISABLED('Channels'),
@@ -133,7 +132,7 @@ async def check_channels_access(request: Request, user: Optional[UserModel] = No
 
     if user:
         if user.role != 'admin' and not await has_permission(
-            user.id, 'features.channels', request.app.state.config.USER_PERMISSIONS
+            user.id, 'features.channels', await Config.get('user.permissions')
         ):
             raise HTTPException(
                 status_code=status.HTTP_401_UNAUTHORIZED,
@@ -294,7 +293,7 @@ async def create_new_channel(
         )
 
     form_data.access_grants = await filter_allowed_access_grants(
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         user.id,
         user.role,
         form_data.access_grants,
@@ -316,6 +315,13 @@ async def create_new_channel(
                 await enter_room_for_users(f'channel:{existing_channel.id}', participant_ids)
 
                 await Channels.update_member_active_status(existing_channel.id, user.id, True, db=db)
+                await publish_event(
+                    request,
+                    EVENTS.CHANNEL_MEMBER_ACTIVE_UPDATED,
+                    actor=user,
+                    subject_id=existing_channel.id,
+                    data={'is_active': True},
+                )
                 return ChannelModel(**existing_channel.model_dump())
 
         channel = await Channels.insert_new_channel(form_data, user.id, db=db)
@@ -330,6 +336,13 @@ async def create_new_channel(
             )
             await enter_room_for_users(f'channel:{channel.id}', participant_ids)
 
+            await publish_event(
+                request,
+                EVENTS.CHANNEL_CREATED,
+                actor=user,
+                subject_id=channel.id,
+                data={'type': channel.type, 'name': channel.name},
+            )
             return ChannelModel(**channel.model_dump())
         else:
             raise Exception('Error creating channel')
@@ -440,7 +453,40 @@ async def get_channel_by_id(
 PAGE_ITEM_COUNT = 30
 
 
-@router.get('/{id}/members', response_model=UserListResponse)
+class ChannelMemberResponse(BaseModel):
+    id: str
+    email: str
+    name: str
+    role: str
+    profile_image_url: str | None = None
+    presence_state: str | None = None
+    status_emoji: str | None = None
+    status_message: str | None = None
+    status_expires_at: int | None = None
+    is_active: bool = False
+
+
+class ChannelMemberListResponse(BaseModel):
+    users: list[ChannelMemberResponse]
+    total: int
+
+
+def serialize_channel_member(user: UserModel) -> ChannelMemberResponse:
+    return ChannelMemberResponse(
+        id=user.id,
+        email=user.email,
+        name=user.name,
+        role=user.role,
+        profile_image_url=user.profile_image_url,
+        presence_state=user.presence_state,
+        status_emoji=user.status_emoji,
+        status_message=user.status_message,
+        status_expires_at=user.status_expires_at,
+        is_active=Users.is_active(user),
+    )
+
+
+@router.get('/{id}/members', response_model=ChannelMemberListResponse)
 async def get_channel_members_by_id(
     request: Request,
     id: str,
@@ -475,7 +521,7 @@ async def get_channel_members_by_id(
         total = len(fetched_users)
 
         return {
-            'users': [UserModelResponse(**u.model_dump(), is_active=Users.is_active(u)) for u in fetched_users],
+            'users': [serialize_channel_member(u) for u in fetched_users],
             'total': total,
         }
     else:
@@ -503,7 +549,7 @@ async def get_channel_members_by_id(
         total = result['total']
 
         return {
-            'users': [UserModelResponse(**u.model_dump(), is_active=Users.is_active(u)) for u in fetched_users],
+            'users': [serialize_channel_member(u) for u in fetched_users],
             'total': total,
         }
 
@@ -534,6 +580,13 @@ async def update_is_active_member_by_id_and_user_id(
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND)
 
     await Channels.update_member_active_status(channel.id, user.id, form_data.is_active, db=db)
+    await publish_event(
+        request,
+        EVENTS.CHANNEL_MEMBER_ACTIVE_UPDATED,
+        actor=user,
+        subject_id=channel.id,
+        data={'is_active': form_data.is_active},
+    )
     return True
 
 
@@ -568,6 +621,13 @@ async def add_members_by_id(
             channel.id, user.id, form_data.user_ids, form_data.group_ids, db=db
         )
 
+        await publish_event(
+            request,
+            EVENTS.CHANNEL_MEMBER_ADDED,
+            actor=user,
+            subject_id=channel.id,
+            data={'user_ids': form_data.user_ids, 'group_ids': form_data.group_ids},
+        )
         return memberships
     except Exception as e:
         log.exception(e)
@@ -603,6 +663,13 @@ async def remove_members_by_id(
     try:
         deleted = await Channels.remove_members_from_channel(channel.id, form_data.user_ids, db=db)
 
+        await publish_event(
+            request,
+            EVENTS.CHANNEL_MEMBER_REMOVED,
+            actor=user,
+            subject_id=channel.id,
+            data={'user_ids': form_data.user_ids},
+        )
         return deleted
     except Exception as e:
         log.exception(e)
@@ -632,7 +699,7 @@ async def update_channel_by_id(
         raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
 
     form_data.access_grants = await filter_allowed_access_grants(
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         user.id,
         user.role,
         form_data.access_grants,
@@ -641,6 +708,13 @@ async def update_channel_by_id(
 
     try:
         channel = await Channels.update_channel_by_id(id, form_data, db=db)
+        await publish_event(
+            request,
+            EVENTS.CHANNEL_UPDATED,
+            actor=user,
+            subject_id=id,
+            data={'name': channel.name, 'type': channel.type},
+        )
         return ChannelModel(**channel.model_dump())
     except Exception as e:
         log.exception(e)
@@ -670,6 +744,13 @@ async def delete_channel_by_id(
 
     try:
         await Channels.delete_channel_by_id(id, db=db)
+        await publish_event(
+            request,
+            EVENTS.CHANNEL_DELETED,
+            actor=user,
+            subject_id=id,
+            data={'name': channel.name, 'type': channel.type},
+        )
         return True
     except Exception as e:
         log.exception(e)
@@ -726,10 +807,14 @@ async def get_channel_messages(
     user_ids = list(set(m.user_id for m in message_list))
     fetched_users = {u.id: u for u in await Users.get_users_by_user_ids(user_ids, db=db)}
 
+    # Batch fetch reactions and reply counts in 2 queries (fixes N+1)
+    message_ids = [m.id for m in message_list]
+    all_reactions = await Messages.get_reactions_by_message_ids(message_ids, db=db)
+    all_reply_counts = await Messages.get_thread_reply_counts_by_message_ids(message_ids, db=db)
+
     messages = []
     for message in message_list:
-        thread_replies = await Messages.get_thread_replies_by_message_id(message.id, db=db)
-        latest_thread_reply_at = thread_replies[0].created_at if thread_replies else None
+        reply_count, latest_reply_at = all_reply_counts.get(message.id, (0, None))
 
         # Use message.user if present (for webhooks), otherwise look up by user_id
         user_info = message.user
@@ -740,9 +825,9 @@ async def get_channel_messages(
             MessageUserResponse(
                 **{
                     **message.model_dump(),
-                    'reply_count': len(thread_replies),
-                    'latest_reply_at': latest_thread_reply_at,
-                    'reactions': await Messages.get_reactions_by_message_id(message.id, db=db),
+                    'reply_count': reply_count,
+                    'latest_reply_at': latest_reply_at,
+                    'reactions': all_reactions.get(message.id, []),
                     'user': user_info,
                 }
             )
@@ -791,6 +876,10 @@ async def get_pinned_channel_messages(
     user_ids = list(set(m.user_id for m in message_list))
     fetched_users = {u.id: u for u in await Users.get_users_by_user_ids(user_ids, db=db)}
 
+    # Batch fetch reactions in 1 query (fixes N+1)
+    message_ids = [m.id for m in message_list]
+    all_reactions = await Messages.get_reactions_by_message_ids(message_ids, db=db)
+
     messages = []
     for message in message_list:
         # Check for webhook identity in meta
@@ -810,7 +899,7 @@ async def get_pinned_channel_messages(
             MessageWithReactionsResponse(
                 **{
                     **message.model_dump(),
-                    'reactions': await Messages.get_reactions_by_message_id(message.id, db=db),
+                    'reactions': all_reactions.get(message.id, []),
                     'user': user_info,
                 }
             )
@@ -825,28 +914,36 @@ async def get_pinned_channel_messages(
 
 
 async def send_notification(request, channel, message, active_user_ids, db=None):
-    name = request.app.state.WEBUI_NAME
-    webui_url = request.app.state.config.WEBUI_URL
-    enable_user_webhooks = request.app.state.config.ENABLE_USER_WEBHOOKS
+    webui_url = await Config.get('webui.url')
+    enable_user_webhooks = await Config.get('ui.enable_user_webhooks')
 
     users = await get_channel_users_with_access(channel, 'read', db=db)
 
+    # Batch fetch channel members in 1 query (fixes N+1)
+    member_ids = {m.user_id for m in await Channels.get_members_by_channel_id(channel.id, db=db)}
+    url = f'{webui_url}/channels/{channel.id}'
+
     for u in users:
-        if (u.id not in active_user_ids) and await Channels.is_user_channel_member(channel.id, u.id, db=db):
+        if (u.id not in active_user_ids) and u.id in member_ids:
             if enable_user_webhooks and u.settings:
-                webhook_url = u.settings.ui.get('notifications', {}).get('webhook_url', None)
-                if webhook_url:
-                    await post_webhook(
-                        name,
-                        webhook_url,
-                        f'#{channel.name} - {webui_url}/channels/{channel.id}\n\n{message.content}',
-                        {
-                            'action': 'channel',
-                            'message': message.content,
-                            'title': channel.name,
-                            'url': f'{webui_url}/channels/{channel.id}',
-                        },
-                    )
+                await publish_event(
+                    request,
+                    EVENTS.CHANNEL_MESSAGE,
+                    subject_id=channel.id,
+                    subject_type='channel',
+                    data={
+                        'user_id': u.id,
+                        'channel_id': channel.id,
+                        'message_id': message.id,
+                        'sender_id': message.user_id,
+                        'content': message.content,
+                        'message': f'#{channel.name} - {url}\n\n{message.content}',
+                        'content_preview': message.content[:300],
+                        'title': channel.name,
+                        'url': url,
+                    },
+                    message=channel.name,
+                )
 
     return True
 
@@ -890,13 +987,20 @@ async def model_response_handler(request, channel, message, user, db=None):
                         db=db,
                     )
                 )[::-1]
+                response_parent_id = (
+                    message.parent_id
+                    if message.parent_id
+                    else (
+                        message.id if await Config.get('channels.model_response_mode', 'thread') == 'thread' else None
+                    )
+                )
 
                 response_message, channel = await new_message_handler(
                     request,
                     channel.id,
                     MessageForm(
                         **{
-                            'parent_id': (message.parent_id if message.parent_id else message.id),
+                            'parent_id': response_parent_id,
                             'content': f'',
                             'data': {},
                             'meta': {
@@ -978,7 +1082,7 @@ async def model_response_handler(request, channel, message, user, db=None):
                 )
 
                 tool_ids = _resolve_model_tool_ids(request.app, model_id)
-                features = _resolve_model_features(request.app, model_id)
+                features = await _resolve_model_features(request.app, model_id)
                 filter_ids = _resolve_model_filter_ids(request.app, model_id)
 
                 # Build full form_data — same shape as frontend POST.
@@ -1036,6 +1140,13 @@ async def new_message_handler(request: Request, id: str, form_data: MessageForm,
         ):
             raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
 
+    # Thread parent / reply target must belong to this channel (no cross-channel binding).
+    for ref_id in (form_data.parent_id, form_data.reply_to_id):
+        if ref_id:
+            ref = await Messages.get_message_by_id(ref_id, include_thread_replies=False, db=db)
+            if not ref or ref.channel_id != channel.id:
+                raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT())
+
     try:
         message = await Messages.insert_new_message(form_data, channel.id, user.id, db=db)
         if message:
@@ -1128,6 +1239,16 @@ async def post_new_message(
 
         background_tasks.add_task(background_handler)
 
+        await publish_event(
+            request,
+            EVENTS.MESSAGE_CREATED,
+            actor=user,
+            subject_id=message.id,
+            data={
+                'channel_id': channel.id,
+                'content_preview': message.content[:300],
+            },
+        )
         return message
 
     except HTTPException as e:
@@ -1255,12 +1376,37 @@ async def pin_channel_message(
         await Messages.update_is_pinned_by_id(message_id, form_data.is_pinned, user.id, db=db)
         message = await Messages.get_message_by_id(message_id, db=db)
         message_user = await Users.get_user_by_id(message.user_id, db=db)
-        return MessageUserResponse(
+        message_data = MessageUserResponse(
             **{
                 **message.model_dump(),
                 'user': UserNameResponse(**message_user.model_dump()) if message_user else None,
             }
         )
+
+        await sio.emit(
+            'events:channel',
+            {
+                'channel_id': channel.id,
+                'message_id': message.id,
+                'data': {
+                    'type': 'message:update',
+                    'data': message_data.model_dump(),
+                },
+                'user': UserNameResponse(**user.model_dump()).model_dump(),
+                'channel': channel.model_dump(),
+            },
+            to=f'channel:{channel.id}',
+        )
+
+        await publish_event(
+            request,
+            EVENTS.MESSAGE_PINNED if form_data.is_pinned else EVENTS.MESSAGE_UNPINNED,
+            actor=user,
+            subject_id=message_id,
+            subject_type='message',
+            data={'channel_id': id},
+        )
+        return message_data
     except Exception as e:
         log.exception(e)
         raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT())
@@ -1302,6 +1448,10 @@ async def get_channel_thread_messages(
     user_ids = list(set(m.user_id for m in message_list))
     fetched_users = {u.id: u for u in await Users.get_users_by_user_ids(user_ids, db=db)}
 
+    # Batch fetch reactions in 1 query (fixes N+1)
+    message_ids = [m.id for m in message_list]
+    all_reactions = await Messages.get_reactions_by_message_ids(message_ids, db=db)
+
     messages = []
     for message in message_list:
         # Use message.user if present (for webhooks), otherwise look up by user_id
@@ -1315,7 +1465,7 @@ async def get_channel_thread_messages(
                     **message.model_dump(),
                     'reply_count': 0,
                     'latest_reply_at': None,
-                    'reactions': await Messages.get_reactions_by_message_id(message.id, db=db),
+                    'reactions': all_reactions.get(message.id, []),
                     'user': user_info,
                 }
             )
@@ -1357,12 +1507,13 @@ async def update_message_by_id(
         if user.role != 'admin' and message.user_id != user.id:
             raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
     else:
-        if (
-            user.role != 'admin'
-            and message.user_id != user.id
-            and not await channel_has_access(user.id, channel, permission='write', strict=False, db=db)
+        if user.role != 'admin' and not await channel_has_access(
+            user.id, channel, permission='write', strict=False, db=db
         ):
             raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
+        # Write access is not authorship — block cross-member edits.
+        if user.role != 'admin' and message.user_id != user.id:
+            raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
 
     try:
         await Messages.update_message_by_id(message_id, form_data, db=db)
@@ -1384,6 +1535,13 @@ async def update_message_by_id(
                 to=f'channel:{channel.id}',
             )
 
+        await publish_event(
+            request,
+            EVENTS.MESSAGE_UPDATED,
+            actor=user,
+            subject_id=message_id,
+            data={'channel_id': id, 'content_preview': form_data.content[:300]},
+        )
         return MessageModel(**message.model_dump())
     except Exception as e:
         log.exception(e)
@@ -1455,6 +1613,13 @@ async def add_reaction_to_message(
             to=f'channel:{channel.id}',
         )
 
+        await publish_event(
+            request,
+            EVENTS.MESSAGE_REACTION_ADDED,
+            actor=user,
+            subject_id=message_id,
+            data={'channel_id': id, 'reaction': form_data.name},
+        )
         return True
     except Exception as e:
         log.exception(e)
@@ -1523,6 +1688,13 @@ async def remove_reaction_by_id_and_user_id_and_name(
             to=f'channel:{channel.id}',
         )
 
+        await publish_event(
+            request,
+            EVENTS.MESSAGE_REACTION_REMOVED,
+            actor=user,
+            subject_id=message_id,
+            data={'channel_id': id, 'reaction': form_data.name},
+        )
         return True
     except Exception as e:
         log.exception(e)
@@ -1561,18 +1733,17 @@ async def delete_message_by_id(
         if user.role != 'admin' and message.user_id != user.id:
             raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
     else:
-        if (
-            user.role != 'admin'
-            and message.user_id != user.id
-            and not await channel_has_access(
-                user.id,
-                channel,
-                permission='write',
-                strict=False,
-                db=db,
-            )
+        if user.role != 'admin' and not await channel_has_access(
+            user.id,
+            channel,
+            permission='write',
+            strict=False,
+            db=db,
         ):
             raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
+        # Write access is not authorship — block cross-member deletes.
+        if user.role != 'admin' and message.user_id != user.id:
+            raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.DEFAULT())
 
     try:
         await Messages.delete_message_by_id(message_id, db=db)
@@ -1614,6 +1785,13 @@ async def delete_message_by_id(
                     to=f'channel:{channel.id}',
                 )
 
+        await publish_event(
+            request,
+            EVENTS.MESSAGE_DELETED,
+            actor=user,
+            subject_id=message_id,
+            data={'channel_id': id},
+        )
         return True
     except Exception as e:
         log.exception(e)
@@ -1699,6 +1877,13 @@ async def create_channel_webhook(
     if not webhook:
         raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT())
 
+    await publish_event(
+        request,
+        EVENTS.CHANNEL_WEBHOOK_CREATED,
+        actor=user,
+        subject_id=webhook.id,
+        data={'channel_id': id, 'name': webhook.name},
+    )
     return webhook
 
 
@@ -1728,6 +1913,13 @@ async def update_channel_webhook(
     if not updated:
         raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT())
 
+    await publish_event(
+        request,
+        EVENTS.CHANNEL_WEBHOOK_UPDATED,
+        actor=user,
+        subject_id=webhook_id,
+        data={'channel_id': id, 'name': updated.name},
+    )
     return updated
 
 
@@ -1752,7 +1944,16 @@ async def delete_channel_webhook(
     if not webhook or webhook.channel_id != id:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND)
 
-    return await Channels.delete_webhook_by_id(webhook_id, db=db)
+    deleted = await Channels.delete_webhook_by_id(webhook_id, db=db)
+    if deleted:
+        await publish_event(
+            request,
+            EVENTS.CHANNEL_WEBHOOK_DELETED,
+            actor=user,
+            subject_id=webhook_id,
+            data={'channel_id': id},
+        )
+    return deleted
 
 
 ############################
@@ -1835,4 +2036,12 @@ async def post_webhook_message(
         to=f'channel:{channel.id}',
     )
 
+    await publish_event(
+        request,
+        EVENTS.MESSAGE_CREATED,
+        actor={'id': webhook.id, 'name': webhook.name, 'role': 'webhook', 'type': 'webhook'},
+        subject_id=message.id,
+        source='channel_webhook',
+        data={'channel_id': channel.id, 'content_preview': form_data.content[:300]},
+    )
     return {'success': True, 'message_id': message.id}

backend/open_webui/routers/configs.py

@@ -7,22 +7,27 @@ from typing import Optional
 import aiohttp
 from fastapi import APIRouter, Depends, HTTPException, Request
 from mcp.shared.auth import OAuthMetadata
-from open_webui.config import BannerModel, async_save_config, get_config, save_config
+from open_webui.config import BannerModel
 from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, AIOHTTP_CLIENT_TIMEOUT
+from open_webui.events import EVENTS, publish_event
+from open_webui.models.config import Config
 from open_webui.models.oauth_sessions import OAuthSessions
 from open_webui.utils.auth import get_admin_user, get_verified_user
 from open_webui.utils.headers import get_custom_headers
 from open_webui.utils.mcp.client import MCPClient
 from open_webui.utils.oauth import (
     OAuthClientInformationFull,
+    apply_connection_oauth_options,
     decrypt_data,
     encrypt_data,
     get_discovery_urls,
     get_oauth_client_info_with_dynamic_client_registration,
     get_oauth_client_info_with_static_credentials,
+    recover_static_oauth_client_metadata,
     resolve_oauth_client_info,
 )
 from open_webui.utils.tools import (
+    bearer_auth_header,
     get_tool_server_data,
     get_tool_server_url,
     set_terminal_servers,
@@ -34,6 +39,53 @@ router = APIRouter()
 
 log = logging.getLogger(__name__)
 
+CONNECTIONS_CONFIG_KEYS = {
+    'ENABLE_DIRECT_CONNECTIONS': 'direct.enable',
+    'ENABLE_BASE_MODELS_CACHE': 'models.base_models_cache',
+}
+CODE_EXECUTION_CONFIG_KEYS = {
+    'ENABLE_CODE_EXECUTION': 'code_execution.enable',
+    'CODE_EXECUTION_ENGINE': 'code_execution.engine',
+    'CODE_EXECUTION_JUPYTER_URL': 'code_execution.jupyter.url',
+    'CODE_EXECUTION_JUPYTER_AUTH': 'code_execution.jupyter.auth',
+    'CODE_EXECUTION_JUPYTER_AUTH_TOKEN': 'code_execution.jupyter.auth_token',
+    'CODE_EXECUTION_JUPYTER_AUTH_PASSWORD': 'code_execution.jupyter.auth_password',
+    'CODE_EXECUTION_JUPYTER_TIMEOUT': 'code_execution.jupyter.timeout',
+    'ENABLE_CODE_INTERPRETER': 'code_interpreter.enable',
+    'CODE_INTERPRETER_ENGINE': 'code_interpreter.engine',
+    'CODE_INTERPRETER_PROMPT_TEMPLATE': 'code_interpreter.prompt_template',
+    'CODE_INTERPRETER_JUPYTER_URL': 'code_interpreter.jupyter.url',
+    'CODE_INTERPRETER_JUPYTER_AUTH': 'code_interpreter.jupyter.auth',
+    'CODE_INTERPRETER_JUPYTER_AUTH_TOKEN': 'code_interpreter.jupyter.auth_token',
+    'CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD': 'code_interpreter.jupyter.auth_password',
+    'CODE_INTERPRETER_JUPYTER_TIMEOUT': 'code_interpreter.jupyter.timeout',
+}
+MODELS_CONFIG_KEYS = {
+    'DEFAULT_MODELS': 'ui.default_models',
+    'DEFAULT_PINNED_MODELS': 'ui.default_pinned_models',
+    'MODEL_ORDER_LIST': 'ui.model_order_list',
+    'DEFAULT_MODEL_METADATA': 'models.default_metadata',
+    'DEFAULT_MODEL_PARAMS': 'models.default_params',
+}
+SUBAGENTS_CONFIG_KEYS = {
+    'ENABLE_SUBAGENTS': 'subagents.enable',
+    'SUBAGENTS_BACKGROUND_ENABLED': 'subagents.background_enabled',
+    'SUBAGENTS_MAX_CONCURRENT': 'subagents.max_concurrent',
+    'SUBAGENTS_MAX_ASYNC': 'subagents.max_async',
+    'SUBAGENTS_MAX_ITERATIONS': 'subagents.max_iterations',
+    'SUBAGENTS_MAX_OUTPUT': 'subagents.max_output',
+    'SUBAGENTS_SYSTEM_PROMPT': 'subagents.system_prompt',
+}
+
+
+async def get_config_values(key_map: dict[str, str]) -> dict:
+    values = await Config.get_many(*key_map.values())
+    return {field: values[storage_key] for field, storage_key in key_map.items() if storage_key in values}
+
+
+def config_updates(data: dict, key_map: dict[str, str]) -> dict:
+    return {key_map[field]: value for field, value in data.items() if field in key_map}
+
 
 ############################
 # ImportConfig
@@ -48,9 +100,15 @@ class ImportConfigForm(BaseModel):
 
 @router.post('/import', response_model=dict)
 async def import_config(request: Request, form_data: ImportConfigForm, user=Depends(get_admin_user)):
-    await async_save_config(form_data.config)
-    request.app.state.config._sync_to_redis()
-    return get_config()
+    await Config.upsert(form_data.config)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_IMPORTED,
+        actor=user,
+        subject_id='import',
+        data={'keys': list(form_data.config.keys())},
+    )
+    return await Config.get_all()
 
 
 ############################
@@ -60,7 +118,12 @@ async def import_config(request: Request, form_data: ImportConfigForm, user=Depe
 
 @router.get('/export', response_model=dict)
 async def export_config(user=Depends(get_admin_user)):
-    return get_config()
+    return await Config.get_all()
+
+
+@router.get('/namespace/{namespace}', response_model=dict)
+async def get_config_namespace(namespace: str, user=Depends(get_admin_user)):
+    return await Config.get_namespace(namespace)
 
 
 ############################
@@ -75,10 +138,7 @@ class ConnectionsConfigForm(BaseModel):
 
 @router.get('/connections', response_model=ConnectionsConfigForm)
 async def get_connections_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        'ENABLE_DIRECT_CONNECTIONS': request.app.state.config.ENABLE_DIRECT_CONNECTIONS,
-        'ENABLE_BASE_MODELS_CACHE': request.app.state.config.ENABLE_BASE_MODELS_CACHE,
-    }
+    return await get_config_values(CONNECTIONS_CONFIG_KEYS)
 
 
 @router.post('/connections', response_model=ConnectionsConfigForm)
@@ -87,13 +147,17 @@ async def set_connections_config(
     form_data: ConnectionsConfigForm,
     user=Depends(get_admin_user),
 ):
-    request.app.state.config.ENABLE_DIRECT_CONNECTIONS = form_data.ENABLE_DIRECT_CONNECTIONS
-    request.app.state.config.ENABLE_BASE_MODELS_CACHE = form_data.ENABLE_BASE_MODELS_CACHE
-
-    return {
-        'ENABLE_DIRECT_CONNECTIONS': request.app.state.config.ENABLE_DIRECT_CONNECTIONS,
-        'ENABLE_BASE_MODELS_CACHE': request.app.state.config.ENABLE_BASE_MODELS_CACHE,
-    }
+    await Config.upsert(config_updates(form_data.model_dump(), CONNECTIONS_CONFIG_KEYS))
+    values = await get_config_values(CONNECTIONS_CONFIG_KEYS)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_CONNECTIONS_UPDATED,
+        actor=user,
+        subject_id='connections',
+        subject_type='config',
+        data=values,
+    )
+    return values
 
 
 class OAuthClientRegistrationForm(BaseModel):
@@ -102,6 +166,7 @@ class OAuthClientRegistrationForm(BaseModel):
     client_name: str | None = None
     client_secret: str | None = None
     oauth_server_url: str | None = None
+    oauth_scope: str | None = None
 
 
 @router.post('/oauth/clients/register')
@@ -126,10 +191,11 @@ async def register_oauth_client(
                 oauth_server_url,
                 oauth_client_id=form_data.client_id,
                 oauth_client_secret=form_data.client_secret,
+                oauth_scope=form_data.oauth_scope,
             )
         else:
             oauth_client_info = await get_oauth_client_info_with_dynamic_client_registration(
-                request, oauth_client_id, oauth_server_url
+                request, oauth_client_id, oauth_server_url, oauth_scope=form_data.oauth_scope
             )
         return {
             'status': True,
@@ -139,7 +205,7 @@ async def register_oauth_client(
         log.debug(f'Failed to register OAuth client: {e}')
         raise HTTPException(
             status_code=400,
-            detail=f'Failed to register OAuth client',
+            detail=f'Failed to register OAuth client: {e}',
         )
 
 
@@ -167,9 +233,7 @@ class ToolServersConfigForm(BaseModel):
 
 @router.get('/tool_servers', response_model=ToolServersConfigForm)
 async def get_tool_servers_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        'TOOL_SERVER_CONNECTIONS': request.app.state.config.TOOL_SERVER_CONNECTIONS,
-    }
+    return {'TOOL_SERVER_CONNECTIONS': await Config.get('tool_server.connections')}
 
 
 @router.post('/tool_servers', response_model=ToolServersConfigForm)
@@ -178,13 +242,14 @@ async def set_tool_servers_config(
     form_data: ToolServersConfigForm,
     user=Depends(get_admin_user),
 ):
-    for connection in request.app.state.config.TOOL_SERVER_CONNECTIONS:
+    existing_connections = await Config.get('tool_server.connections', []) or []
+    for connection in existing_connections:
         server_type = connection.get('type', 'openapi')
         auth_type = connection.get('auth_type', 'none')
 
         if auth_type in ('oauth_2.1', 'oauth_2.1_static'):
             # Remove existing OAuth clients for tool servers
-            server_id = connection.get('info', {}).get('id')
+            server_id = (connection.get('info') or {}).get('id')
             client_key = f'{server_type}:{server_id}'
 
             try:
@@ -193,21 +258,22 @@ async def set_tool_servers_config(
                 pass
 
     # Set new tool server connections
-    request.app.state.config.TOOL_SERVER_CONNECTIONS = [
-        connection.model_dump() for connection in form_data.TOOL_SERVER_CONNECTIONS
-    ]
+    connections = [connection.model_dump() for connection in form_data.TOOL_SERVER_CONNECTIONS]
+    await Config.upsert({'tool_server.connections': connections})
 
     await set_tool_servers(request)
 
-    for connection in request.app.state.config.TOOL_SERVER_CONNECTIONS:
+    for connection in connections:
         server_type = connection.get('type', 'openapi')
         if server_type == 'mcp':
-            server_id = connection.get('info', {}).get('id')
+            server_id = (connection.get('info') or {}).get('id')
             auth_type = connection.get('auth_type', 'none')
 
             if auth_type in ('oauth_2.1', 'oauth_2.1_static') and server_id:
                 try:
                     oauth_client_info = resolve_oauth_client_info(connection)
+                    oauth_client_info = await recover_static_oauth_client_metadata(connection, oauth_client_info)
+                    oauth_client_info = apply_connection_oauth_options(connection, oauth_client_info)
                     request.app.state.oauth_client_manager.add_client(
                         f'{server_type}:{server_id}',
                         OAuthClientInformationFull(**oauth_client_info),
@@ -216,9 +282,15 @@ async def set_tool_servers_config(
                     log.debug(f'Failed to add OAuth client for MCP tool server: {e}')
                     continue
 
-    return {
-        'TOOL_SERVER_CONNECTIONS': request.app.state.config.TOOL_SERVER_CONNECTIONS,
-    }
+    await publish_event(
+        request,
+        EVENTS.CONFIG_TOOL_SERVERS_UPDATED,
+        actor=user,
+        subject_id='tool_server.connections',
+        subject_type='config',
+        data={'count': len(connections), 'types': [connection.get('type', 'openapi') for connection in connections]},
+    )
+    return {'TOOL_SERVER_CONNECTIONS': connections}
 
 
 class TerminalServerConnection(BaseModel):
@@ -235,10 +307,8 @@ class TerminalServerConnection(BaseModel):
 
     config: dict | None = None
 
-    # Orchestrator policy fields
-    server_type: str | None = None  # "orchestrator", "terminal"
+    server_type: str | None = None
     policy_id: str | None = None
-    policy: dict | None = None  # cached policy data
 
     model_config = ConfigDict(extra='allow')
 
@@ -249,9 +319,7 @@ class TerminalServersConfigForm(BaseModel):
 
 @router.get('/terminal_servers')
 async def get_terminal_servers_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        'TERMINAL_SERVER_CONNECTIONS': request.app.state.config.TERMINAL_SERVER_CONNECTIONS,
-    }
+    return {'TERMINAL_SERVER_CONNECTIONS': await Config.get('terminal_server.connections')}
 
 
 @router.post('/terminal_servers')
@@ -260,15 +328,22 @@ async def set_terminal_servers_config(
     form_data: TerminalServersConfigForm,
     user=Depends(get_admin_user),
 ):
-    request.app.state.config.TERMINAL_SERVER_CONNECTIONS = [
-        connection.model_dump() for connection in form_data.TERMINAL_SERVER_CONNECTIONS
+    connections = [
+        connection.model_dump(exclude={'policy', 'lifecycle'}) for connection in form_data.TERMINAL_SERVER_CONNECTIONS
     ]
+    await Config.upsert({'terminal_server.connections': connections})
 
     await set_terminal_servers(request)
 
-    return {
-        'TERMINAL_SERVER_CONNECTIONS': request.app.state.config.TERMINAL_SERVER_CONNECTIONS,
-    }
+    await publish_event(
+        request,
+        EVENTS.CONFIG_TERMINAL_SERVERS_UPDATED,
+        actor=user,
+        subject_id='terminal_server.connections',
+        subject_type='config',
+        data={'count': len(connections)},
+    )
+    return {'TERMINAL_SERVER_CONNECTIONS': connections}
 
 
 @router.post('/terminal_servers/verify')
@@ -287,7 +362,7 @@ async def verify_terminal_server_connection(
 
     headers = {}
     if form_data.auth_type == 'bearer' and form_data.key:
-        headers['Authorization'] = f'Bearer {form_data.key}'
+        headers.update(bearer_auth_header(form_data.key))
 
     try:
         async with aiohttp.ClientSession(
@@ -325,23 +400,39 @@ class TerminalServerPolicyForm(BaseModel):
     key: str | None = ''
     auth_type: str | None = 'bearer'
     policy_id: str
-    policy_data: dict
+    policy_data: dict | None = None
+
+
+class TerminalServerLifecycleForm(BaseModel):
+    url: str
+    key: str | None = ''
+    auth_type: str | None = 'bearer'
+    policy_id: str
+    lifecycle_data: dict | None = None
+
+
+class TerminalServerRefreshForm(BaseModel):
+    url: str
+    key: str | None = ''
+    auth_type: str | None = 'bearer'
+    user_id: str | None = None
+    policy_id: str | None = None
+    only_idle: bool = True
+    reset: bool = False
 
 
 @router.post('/terminal_servers/policy')
 async def put_terminal_server_policy(
     request: Request, form_data: TerminalServerPolicyForm, user=Depends(get_admin_user)
 ):
-    """
-    Proxy a policy PUT to an orchestrator terminal server.
-    """
+    """Proxy a policy read or update to an orchestrator terminal server."""
     base_url = (form_data.url or '').rstrip('/')
     if not base_url:
         raise HTTPException(status_code=400, detail='Terminal server URL is required')
 
     headers = {'Content-Type': 'application/json'}
     if form_data.auth_type == 'bearer' and form_data.key:
-        headers['Authorization'] = f'Bearer {form_data.key}'
+        headers.update(bearer_auth_header(form_data.key))
 
     try:
         async with aiohttp.ClientSession(
@@ -349,8 +440,12 @@ async def put_terminal_server_policy(
             timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT),
         ) as session:
             policy_url = f'{base_url}/api/v1/policies/{form_data.policy_id}'
-            async with session.put(
-                policy_url, headers=headers, json=form_data.policy_data, ssl=AIOHTTP_CLIENT_SESSION_SSL
+            async with session.request(
+                'GET' if form_data.policy_data is None else 'PUT',
+                policy_url,
+                headers=headers,
+                json=form_data.policy_data,
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
             ) as resp:
                 if resp.ok:
                     return await resp.json()
@@ -359,8 +454,92 @@ async def put_terminal_server_policy(
     except HTTPException:
         raise
     except Exception as e:
-        log.debug(f'Failed to save policy to terminal server: {e}')
-        raise HTTPException(status_code=400, detail='Failed to save policy to terminal server')
+        log.debug(f'Failed to access policy on terminal server: {e}')
+        raise HTTPException(status_code=400, detail='Failed to access policy on terminal server')
+
+
+@router.post('/terminal_servers/lifecycle')
+async def put_terminal_server_lifecycle(
+    request: Request, form_data: TerminalServerLifecycleForm, user=Depends(get_admin_user)
+):
+    """Proxy a lifecycle read or update to an orchestrator terminal server."""
+    base_url = (form_data.url or '').rstrip('/')
+    if not base_url:
+        raise HTTPException(status_code=400, detail='Terminal server URL is required')
+
+    headers = {'Content-Type': 'application/json'}
+    if form_data.auth_type == 'bearer' and form_data.key:
+        headers.update(bearer_auth_header(form_data.key))
+
+    try:
+        async with aiohttp.ClientSession(
+            trust_env=True,
+            timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT),
+        ) as session:
+            lifecycle_url = f'{base_url}/api/v1/policies/{form_data.policy_id}/lifecycle'
+            async with session.request(
+                'GET' if form_data.lifecycle_data is None else 'PUT',
+                lifecycle_url,
+                headers=headers,
+                json=form_data.lifecycle_data,
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
+            ) as resp:
+                if resp.ok:
+                    return await resp.json()
+                detail = await resp.text()
+                raise HTTPException(status_code=resp.status, detail=detail)
+    except HTTPException:
+        raise
+    except Exception as e:
+        log.debug(f'Failed to access lifecycle on terminal server: {e}')
+        raise HTTPException(status_code=400, detail='Failed to access lifecycle on terminal server')
+
+
+@router.post('/terminal_servers/refresh')
+async def refresh_terminal_server_terminals(
+    request: Request, form_data: TerminalServerRefreshForm, user=Depends(get_admin_user)
+):
+    """
+    Proxy a terminal refresh request to an orchestrator terminal server.
+    """
+    base_url = (form_data.url or '').rstrip('/')
+    if not base_url:
+        raise HTTPException(status_code=400, detail='Terminal server URL is required')
+
+    headers = {'Content-Type': 'application/json'}
+    if form_data.auth_type == 'bearer' and form_data.key:
+        headers.update(bearer_auth_header(form_data.key))
+
+    body = {
+        'only_idle': form_data.only_idle,
+        'reset': form_data.reset,
+    }
+    if form_data.user_id:
+        body['user_id'] = form_data.user_id
+    if form_data.policy_id:
+        body['policy_id'] = form_data.policy_id
+
+    try:
+        async with aiohttp.ClientSession(
+            trust_env=True,
+            timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT),
+        ) as session:
+            refresh_url = f'{base_url}/api/v1/terminals/refresh'
+            async with session.post(
+                refresh_url,
+                headers=headers,
+                json=body,
+                ssl=AIOHTTP_CLIENT_SESSION_SSL,
+            ) as resp:
+                if resp.ok:
+                    return await resp.json()
+                detail = await resp.text()
+                raise HTTPException(status_code=resp.status, detail=detail)
+    except HTTPException:
+        raise
+    except Exception as e:
+        log.debug(f'Failed to refresh terminals: {e}')
+        raise HTTPException(status_code=400, detail='Failed to refresh terminals')
 
 
 @router.post('/tool_servers/verify')
@@ -435,7 +614,7 @@ async def verify_tool_servers_config(request: Request, form_data: ToolServerConn
                     if form_data.headers and isinstance(form_data.headers, dict):
                         if headers is None:
                             headers = {}
-                        custom_headers = get_custom_headers(form_data.headers, user)
+                        custom_headers = await get_custom_headers(form_data.headers, user)
                         headers.update(custom_headers)
 
                     await client.connect(form_data.url, headers=headers)
@@ -480,7 +659,7 @@ async def verify_tool_servers_config(request: Request, form_data: ToolServerConn
             if form_data.headers and isinstance(form_data.headers, dict):
                 if headers is None:
                     headers = {}
-                custom_headers = get_custom_headers(form_data.headers, user)
+                custom_headers = await get_custom_headers(form_data.headers, user)
                 headers.update(custom_headers)
 
             url = get_tool_server_url(form_data.url, form_data.path)
@@ -518,67 +697,29 @@ class CodeInterpreterConfigForm(BaseModel):
 
 @router.get('/code_execution', response_model=CodeInterpreterConfigForm)
 async def get_code_execution_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        'ENABLE_CODE_EXECUTION': request.app.state.config.ENABLE_CODE_EXECUTION,
-        'CODE_EXECUTION_ENGINE': request.app.state.config.CODE_EXECUTION_ENGINE,
-        'CODE_EXECUTION_JUPYTER_URL': request.app.state.config.CODE_EXECUTION_JUPYTER_URL,
-        'CODE_EXECUTION_JUPYTER_AUTH': request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH,
-        'CODE_EXECUTION_JUPYTER_AUTH_TOKEN': request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH_TOKEN,
-        'CODE_EXECUTION_JUPYTER_AUTH_PASSWORD': request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH_PASSWORD,
-        'CODE_EXECUTION_JUPYTER_TIMEOUT': request.app.state.config.CODE_EXECUTION_JUPYTER_TIMEOUT,
-        'ENABLE_CODE_INTERPRETER': request.app.state.config.ENABLE_CODE_INTERPRETER,
-        'CODE_INTERPRETER_ENGINE': request.app.state.config.CODE_INTERPRETER_ENGINE,
-        'CODE_INTERPRETER_PROMPT_TEMPLATE': request.app.state.config.CODE_INTERPRETER_PROMPT_TEMPLATE,
-        'CODE_INTERPRETER_JUPYTER_URL': request.app.state.config.CODE_INTERPRETER_JUPYTER_URL,
-        'CODE_INTERPRETER_JUPYTER_AUTH': request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH,
-        'CODE_INTERPRETER_JUPYTER_AUTH_TOKEN': request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH_TOKEN,
-        'CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD': request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD,
-        'CODE_INTERPRETER_JUPYTER_TIMEOUT': request.app.state.config.CODE_INTERPRETER_JUPYTER_TIMEOUT,
-    }
+    return await get_config_values(CODE_EXECUTION_CONFIG_KEYS)
 
 
 @router.post('/code_execution', response_model=CodeInterpreterConfigForm)
 async def set_code_execution_config(
     request: Request, form_data: CodeInterpreterConfigForm, user=Depends(get_admin_user)
 ):
-    request.app.state.config.ENABLE_CODE_EXECUTION = form_data.ENABLE_CODE_EXECUTION
-
-    request.app.state.config.CODE_EXECUTION_ENGINE = form_data.CODE_EXECUTION_ENGINE
-    request.app.state.config.CODE_EXECUTION_JUPYTER_URL = form_data.CODE_EXECUTION_JUPYTER_URL
-    request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH = form_data.CODE_EXECUTION_JUPYTER_AUTH
-    request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH_TOKEN = form_data.CODE_EXECUTION_JUPYTER_AUTH_TOKEN
-    request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH_PASSWORD = form_data.CODE_EXECUTION_JUPYTER_AUTH_PASSWORD
-    request.app.state.config.CODE_EXECUTION_JUPYTER_TIMEOUT = form_data.CODE_EXECUTION_JUPYTER_TIMEOUT
-
-    request.app.state.config.ENABLE_CODE_INTERPRETER = form_data.ENABLE_CODE_INTERPRETER
-    request.app.state.config.CODE_INTERPRETER_ENGINE = form_data.CODE_INTERPRETER_ENGINE
-    request.app.state.config.CODE_INTERPRETER_PROMPT_TEMPLATE = form_data.CODE_INTERPRETER_PROMPT_TEMPLATE
-
-    request.app.state.config.CODE_INTERPRETER_JUPYTER_URL = form_data.CODE_INTERPRETER_JUPYTER_URL
-
-    request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH = form_data.CODE_INTERPRETER_JUPYTER_AUTH
-
-    request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH_TOKEN = form_data.CODE_INTERPRETER_JUPYTER_AUTH_TOKEN
-    request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD = form_data.CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD
-    request.app.state.config.CODE_INTERPRETER_JUPYTER_TIMEOUT = form_data.CODE_INTERPRETER_JUPYTER_TIMEOUT
-
-    return {
-        'ENABLE_CODE_EXECUTION': request.app.state.config.ENABLE_CODE_EXECUTION,
-        'CODE_EXECUTION_ENGINE': request.app.state.config.CODE_EXECUTION_ENGINE,
-        'CODE_EXECUTION_JUPYTER_URL': request.app.state.config.CODE_EXECUTION_JUPYTER_URL,
-        'CODE_EXECUTION_JUPYTER_AUTH': request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH,
-        'CODE_EXECUTION_JUPYTER_AUTH_TOKEN': request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH_TOKEN,
-        'CODE_EXECUTION_JUPYTER_AUTH_PASSWORD': request.app.state.config.CODE_EXECUTION_JUPYTER_AUTH_PASSWORD,
-        'CODE_EXECUTION_JUPYTER_TIMEOUT': request.app.state.config.CODE_EXECUTION_JUPYTER_TIMEOUT,
-        'ENABLE_CODE_INTERPRETER': request.app.state.config.ENABLE_CODE_INTERPRETER,
-        'CODE_INTERPRETER_ENGINE': request.app.state.config.CODE_INTERPRETER_ENGINE,
-        'CODE_INTERPRETER_PROMPT_TEMPLATE': request.app.state.config.CODE_INTERPRETER_PROMPT_TEMPLATE,
-        'CODE_INTERPRETER_JUPYTER_URL': request.app.state.config.CODE_INTERPRETER_JUPYTER_URL,
-        'CODE_INTERPRETER_JUPYTER_AUTH': request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH,
-        'CODE_INTERPRETER_JUPYTER_AUTH_TOKEN': request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH_TOKEN,
-        'CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD': request.app.state.config.CODE_INTERPRETER_JUPYTER_AUTH_PASSWORD,
-        'CODE_INTERPRETER_JUPYTER_TIMEOUT': request.app.state.config.CODE_INTERPRETER_JUPYTER_TIMEOUT,
-    }
+    await Config.upsert(config_updates(form_data.model_dump(), CODE_EXECUTION_CONFIG_KEYS))
+    values = await get_config_values(CODE_EXECUTION_CONFIG_KEYS)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_CODE_EXECUTION_UPDATED,
+        actor=user,
+        subject_id='code_execution',
+        subject_type='config',
+        data={
+            'code_execution_enabled': values.get('ENABLE_CODE_EXECUTION'),
+            'code_execution_engine': values.get('CODE_EXECUTION_ENGINE'),
+            'code_interpreter_enabled': values.get('ENABLE_CODE_INTERPRETER'),
+            'code_interpreter_engine': values.get('CODE_INTERPRETER_ENGINE'),
+        },
+    )
+    return values
 
 
 ############################
@@ -595,35 +736,66 @@ class ModelsConfigForm(BaseModel):
 @router.get('/models/defaults')
 async def get_models_defaults(request: Request, user=Depends(get_verified_user)):
     return {
-        'DEFAULT_MODEL_METADATA': request.app.state.config.DEFAULT_MODEL_METADATA,
+        'DEFAULT_MODEL_METADATA': await Config.get('models.default_metadata'),
     }
 
 
 @router.get('/models', response_model=ModelsConfigForm)
 async def get_models_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        'DEFAULT_MODELS': request.app.state.config.DEFAULT_MODELS,
-        'DEFAULT_PINNED_MODELS': request.app.state.config.DEFAULT_PINNED_MODELS,
-        'MODEL_ORDER_LIST': request.app.state.config.MODEL_ORDER_LIST,
-        'DEFAULT_MODEL_METADATA': request.app.state.config.DEFAULT_MODEL_METADATA,
-        'DEFAULT_MODEL_PARAMS': request.app.state.config.DEFAULT_MODEL_PARAMS,
-    }
+    return await get_config_values(MODELS_CONFIG_KEYS)
 
 
 @router.post('/models', response_model=ModelsConfigForm)
 async def set_models_config(request: Request, form_data: ModelsConfigForm, user=Depends(get_admin_user)):
-    request.app.state.config.DEFAULT_MODELS = form_data.DEFAULT_MODELS
-    request.app.state.config.DEFAULT_PINNED_MODELS = form_data.DEFAULT_PINNED_MODELS
-    request.app.state.config.MODEL_ORDER_LIST = form_data.MODEL_ORDER_LIST
-    request.app.state.config.DEFAULT_MODEL_METADATA = form_data.DEFAULT_MODEL_METADATA
-    request.app.state.config.DEFAULT_MODEL_PARAMS = form_data.DEFAULT_MODEL_PARAMS
-    return {
-        'DEFAULT_MODELS': request.app.state.config.DEFAULT_MODELS,
-        'DEFAULT_PINNED_MODELS': request.app.state.config.DEFAULT_PINNED_MODELS,
-        'MODEL_ORDER_LIST': request.app.state.config.MODEL_ORDER_LIST,
-        'DEFAULT_MODEL_METADATA': request.app.state.config.DEFAULT_MODEL_METADATA,
-        'DEFAULT_MODEL_PARAMS': request.app.state.config.DEFAULT_MODEL_PARAMS,
-    }
+    await Config.upsert(config_updates(form_data.model_dump(), MODELS_CONFIG_KEYS))
+    values = await get_config_values(MODELS_CONFIG_KEYS)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_MODELS_UPDATED,
+        actor=user,
+        subject_id='models',
+        subject_type='config',
+        data={
+            'default_models': values.get('DEFAULT_MODELS'),
+            'default_pinned_models': values.get('DEFAULT_PINNED_MODELS'),
+            'model_order_count': len(values.get('MODEL_ORDER_LIST') or []),
+        },
+    )
+    return values
+
+
+class SubagentsConfigForm(BaseModel):
+    ENABLE_SUBAGENTS: bool
+    SUBAGENTS_BACKGROUND_ENABLED: bool
+    SUBAGENTS_MAX_CONCURRENT: int
+    SUBAGENTS_MAX_ASYNC: int
+    SUBAGENTS_MAX_ITERATIONS: int
+    SUBAGENTS_MAX_OUTPUT: int
+    SUBAGENTS_SYSTEM_PROMPT: str
+
+
+@router.get('/subagents', response_model=SubagentsConfigForm)
+async def get_subagents_config(user=Depends(get_admin_user)):
+    return await get_config_values(SUBAGENTS_CONFIG_KEYS)
+
+
+@router.post('/subagents', response_model=SubagentsConfigForm)
+async def set_subagents_config(
+    request: Request,
+    form_data: SubagentsConfigForm,
+    user=Depends(get_admin_user),
+):
+    await Config.upsert(config_updates(form_data.model_dump(), SUBAGENTS_CONFIG_KEYS))
+    values = await get_config_values(SUBAGENTS_CONFIG_KEYS)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_UPDATED,
+        actor=user,
+        subject_id='subagents',
+        subject_type='config',
+        data={'enabled': values.get('ENABLE_SUBAGENTS')},
+    )
+    return values
 
 
 class PromptSuggestion(BaseModel):
@@ -642,8 +814,17 @@ async def set_default_suggestions(
     user=Depends(get_admin_user),
 ):
     data = form_data.model_dump()
-    request.app.state.config.DEFAULT_PROMPT_SUGGESTIONS = data['suggestions']
-    return request.app.state.config.DEFAULT_PROMPT_SUGGESTIONS
+    await Config.upsert({'ui.prompt_suggestions': data['suggestions']})
+    suggestions = await Config.get('ui.prompt_suggestions')
+    await publish_event(
+        request,
+        EVENTS.CONFIG_SUGGESTIONS_UPDATED,
+        actor=user,
+        subject_id='ui.prompt_suggestions',
+        subject_type='config',
+        data={'count': len(suggestions or [])},
+    )
+    return suggestions
 
 
 ############################
@@ -662,8 +843,17 @@ async def set_banners(
     user=Depends(get_admin_user),
 ):
     data = form_data.model_dump()
-    request.app.state.config.BANNERS = data['banners']
-    return request.app.state.config.BANNERS
+    await Config.upsert({'ui.banners': data['banners']})
+    banners = await Config.get('ui.banners')
+    await publish_event(
+        request,
+        EVENTS.CONFIG_BANNERS_UPDATED,
+        actor=user,
+        subject_id='ui.banners',
+        subject_type='config',
+        data={'count': len(banners or [])},
+    )
+    return banners
 
 
 @router.get('/banners', response_model=list[BannerModel])
@@ -671,4 +861,4 @@ async def get_banners(
     request: Request,
     user=Depends(get_verified_user),
 ):
-    return request.app.state.config.BANNERS
+    return await Config.get('ui.banners')

backend/open_webui/routers/evaluations.py

@@ -4,7 +4,9 @@ from typing import Optional
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from fastapi.concurrency import run_in_threadpool
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.internal.db import get_async_session
+from open_webui.models.config import Config
 from open_webui.models.feedbacks import (
     FeedbackForm,
     FeedbackIdResponse,
@@ -25,6 +27,16 @@ log = logging.getLogger(__name__)
 
 router = APIRouter()
 
+EVALUATION_CONFIG_KEYS = {
+    'ENABLE_EVALUATION_ARENA_MODELS': 'evaluation.arena.enable',
+    'EVALUATION_ARENA_MODELS': 'evaluation.arena.models',
+}
+
+
+async def get_config_values(key_map: dict[str, str]) -> dict:
+    values = await Config.get_many(*key_map.values())
+    return {field: values[storage_key] for field, storage_key in key_map.items() if storage_key in values}
+
 
 # Leaderboard Elo Rating Computation
 # The judgment has already been rendered with grace;
@@ -255,10 +267,7 @@ async def get_model_history(
 
 @router.get('/config')
 async def get_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        'ENABLE_EVALUATION_ARENA_MODELS': request.app.state.config.ENABLE_EVALUATION_ARENA_MODELS,
-        'EVALUATION_ARENA_MODELS': request.app.state.config.EVALUATION_ARENA_MODELS,
-    }
+    return await get_config_values(EVALUATION_CONFIG_KEYS)
 
 
 ############################
@@ -277,15 +286,25 @@ async def update_config(
     form_data: UpdateConfigForm,
     user=Depends(get_admin_user),
 ):
-    config = request.app.state.config
+    updates = {}
     if form_data.ENABLE_EVALUATION_ARENA_MODELS is not None:
-        config.ENABLE_EVALUATION_ARENA_MODELS = form_data.ENABLE_EVALUATION_ARENA_MODELS
+        updates['evaluation.arena.enable'] = form_data.ENABLE_EVALUATION_ARENA_MODELS
     if form_data.EVALUATION_ARENA_MODELS is not None:
-        config.EVALUATION_ARENA_MODELS = form_data.EVALUATION_ARENA_MODELS
-    return {
-        'ENABLE_EVALUATION_ARENA_MODELS': config.ENABLE_EVALUATION_ARENA_MODELS,
-        'EVALUATION_ARENA_MODELS': config.EVALUATION_ARENA_MODELS,
-    }
+        updates['evaluation.arena.models'] = form_data.EVALUATION_ARENA_MODELS
+    await Config.upsert(updates)
+    values = await get_config_values(EVALUATION_CONFIG_KEYS)
+    await publish_event(
+        request,
+        EVENTS.CONFIG_UPDATED,
+        actor=user,
+        subject_id='evaluation',
+        data={
+            'keys': list(updates.keys()),
+            'arena_enabled': values.get('ENABLE_EVALUATION_ARENA_MODELS'),
+            'arena_model_count': len(values.get('EVALUATION_ARENA_MODELS') or []),
+        },
+    )
+    return values
 
 
 @router.get('/feedbacks/models', response_model=list[str])
@@ -299,8 +318,19 @@ async def get_all_feedback_ids(user=Depends(get_admin_user), db: AsyncSession =
 
 
 @router.delete('/feedbacks/all')
-async def delete_all_feedbacks(user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+async def delete_all_feedbacks(
+    request: Request,
+    user=Depends(get_admin_user),
+    db: AsyncSession = Depends(get_async_session),
+):
     success = await Feedbacks.delete_all_feedbacks(db=db)
+    if success:
+        await publish_event(
+            request,
+            EVENTS.FEEDBACK_DELETED_ALL,
+            actor=user,
+            subject_id='all',
+        )
     return success
 
 
@@ -332,8 +362,20 @@ async def get_user_feedbacks(
 
 
 @router.delete('/feedbacks', response_model=bool)
-async def delete_feedbacks(user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)):
+async def delete_feedbacks(
+    request: Request,
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
+):
     success = await Feedbacks.delete_feedbacks_by_user_id(user.id, db=db)
+    if success:
+        await publish_event(
+            request,
+            EVENTS.FEEDBACK_DELETED_ALL,
+            actor=user,
+            subject_id=user.id,
+            subject_type='user',
+        )
     return success
 
 
@@ -377,6 +419,13 @@ async def create_feedback(
             detail=ERROR_MESSAGES.DEFAULT(),
         )
 
+    await publish_event(
+        request,
+        EVENTS.FEEDBACK_CREATED,
+        actor=user,
+        subject_id=feedback.id,
+        data={'rating': (feedback.data or {}).get('rating')},
+    )
     return feedback
 
 
@@ -395,6 +444,7 @@ async def get_feedback_by_id(id: str, user=Depends(get_verified_user), db: Async
 
 @router.post('/feedback/{id}', response_model=FeedbackModel)
 async def update_feedback_by_id(
+    request: Request,
     id: str,
     form_data: FeedbackForm,
     user=Depends(get_verified_user),
@@ -408,12 +458,22 @@ async def update_feedback_by_id(
     if not feedback:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND)
 
+    await publish_event(
+        request,
+        EVENTS.FEEDBACK_UPDATED,
+        actor=user,
+        subject_id=feedback.id,
+        data={'rating': (feedback.data or {}).get('rating')},
+    )
     return feedback
 
 
 @router.delete('/feedback/{id}')
 async def delete_feedback_by_id(
-    id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)
+    request: Request,
+    id: str,
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
 ):
     if user.role == 'admin':
         success = await Feedbacks.delete_feedback_by_id(id=id, db=db)
@@ -423,4 +483,10 @@ async def delete_feedback_by_id(
     if not success:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=ERROR_MESSAGES.NOT_FOUND)
 
+    await publish_event(
+        request,
+        EVENTS.FEEDBACK_DELETED,
+        actor=user,
+        subject_id=id,
+    )
     return success

backend/open_webui/routers/files.py

@@ -1,4 +1,5 @@
 import asyncio
+import errno
 import hashlib
 import json
 import logging
@@ -23,9 +24,11 @@ from fastapi import (
 from fastapi.responses import FileResponse, StreamingResponse
 from open_webui.config import BYPASS_ADMIN_ACCESS_CONTROL, STORAGE_LOCAL_CACHE, STORAGE_PROVIDER, UPLOAD_DIR
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.internal.db import get_async_db_context, get_async_session
 from open_webui.models.access_grants import AccessGrants
 from open_webui.models.channels import Channels
+from open_webui.models.config import Config
 from open_webui.models.chats import Chats
 from open_webui.models.files import (
     FileForm,
@@ -105,6 +108,23 @@ def _cleanup_local_cache(file_path: str) -> None:
         log.warning(f'Failed to clean up local cache for {file_path}: {e}')
 
 
+def _matches_configured_mime_type(supported: list[str] | str, content_type: str) -> bool:
+    if isinstance(supported, str):
+        supported = supported.split(',')
+    supported = [item.strip() for item in (supported or []) if item.strip()]
+    if not supported:
+        return False
+    return bool(strict_match_mime_type(supported, content_type))
+
+
+def _media_supported_for_extraction(
+    content_extraction_engine: str | None, supported: list[str] | str | None, content_type: str
+) -> bool:
+    if supported is None:
+        return content_extraction_engine == 'external'
+    return bool(content_extraction_engine and _matches_configured_mime_type(supported, content_type))
+
+
 async def process_uploaded_file(
     request,
     file,
@@ -123,7 +143,11 @@ async def process_uploaded_file(
                 if _is_text_file(file_path):
                     content_type = 'text/plain'
 
-            stt_supported = getattr(request.app.state.config, 'STT_SUPPORTED_CONTENT_TYPES', [])
+            stt_supported = await Config.get('audio.stt.supported_content_types', [])
+            content_extraction_engine = await Config.get('rag.content_extraction_engine')
+            content_extraction_supported_media_mime_types = await Config.get(
+                'rag.content_extraction.supported_media_mime_types'
+            )
 
             if content_type and strict_match_mime_type(stt_supported, content_type):
                 # Audio / STT-supported files → transcribe then index
@@ -144,9 +168,10 @@ async def process_uploaded_file(
             elif (
                 content_type
                 and content_type.startswith(('image/', 'video/'))
-                and request.app.state.config.CONTENT_EXTRACTION_ENGINE != 'external'
+                and not _media_supported_for_extraction(
+                    content_extraction_engine, content_extraction_supported_media_mime_types, content_type
+                )
             ):
-                # Media files without an external extraction engine
                 if content_type.startswith('video/'):
                     # Videos are stored as-is for downstream multimodal
                     # processing (Tools, vision models). Attempting text
@@ -162,7 +187,8 @@ async def process_uploaded_file(
                     raise Exception(f'File type {content_type} is not supported for processing')
 
             else:
-                # Documents, or any file when an external engine is configured
+                # Documents, or media files explicitly enabled for the
+                # configured content extraction engine.
                 if not content_type:
                     log.info(f'File type {file.content_type} is not provided, but trying to process anyway')
                 await process_file(
@@ -178,21 +204,48 @@ async def process_uploaded_file(
             knowledge_id = file_metadata.get('knowledge_id')
             if knowledge_id:
                 try:
-                    await Knowledges.add_file_to_knowledge_by_id(
-                        knowledge_id=knowledge_id,
-                        file_id=file_item.id,
-                        user_id=user.id,
-                        directory_id=file_metadata.get('directory_id'),
+                    # Gate like POST /knowledge/{id}/file/add: a client-supplied
+                    # metadata.knowledge_id must not let a non-writer attach files (CWE-862/863).
+                    knowledge = await Knowledges.get_knowledge_by_id(id=knowledge_id, db=db_session)
+                    can_write = bool(knowledge) and (
+                        knowledge.user_id == user.id
+                        or user.role == 'admin'
+                        or await AccessGrants.has_access(
+                            user_id=user.id,
+                            resource_type='knowledge',
+                            resource_id=knowledge.id,
+                            permission='write',
+                            db=db_session,
+                        )
                     )
-                    await process_file(
-                        request,
-                        ProcessFileForm(file_id=file_item.id, collection_name=knowledge_id),
-                        user=user,
-                        db=db_session,
-                    )
-                    log.info(f'Linked file {file_item.id} to knowledge {knowledge_id}')
+                    if not can_write:
+                        log.warning(
+                            f'Refusing to auto-link file {file_item.id} to knowledge '
+                            f'{knowledge_id}: user {user.id} lacks write access'
+                        )
+                    else:
+                        # Keep the generic file status stream open until the
+                        # KB-specific vector write and durable link both finish.
+                        await Files.update_file_data_by_id(file_item.id, {'status': 'processing'}, db=db_session)
+                        await process_file(
+                            request,
+                            ProcessFileForm(file_id=file_item.id, collection_name=knowledge_id),
+                            user=user,
+                            db=db_session,
+                        )
+                        knowledge_file = await Knowledges.add_file_to_knowledge_by_id(
+                            knowledge_id=knowledge_id,
+                            file_id=file_item.id,
+                            user_id=user.id,
+                            directory_id=file_metadata.get('directory_id'),
+                            db=db_session,
+                        )
+                        if not knowledge_file:
+                            raise Exception(f'Failed to link file {file_item.id} to knowledge {knowledge_id}')
+                        log.info(f'Linked file {file_item.id} to knowledge {knowledge_id}')
                 except Exception as e:
                     log.warning(f'Failed to link file {file_item.id} to knowledge {knowledge_id}: {e}')
+                    raise
 
         except Exception as e:
             log.error(f'Error processing file: {file_item.id}')
@@ -226,7 +279,7 @@ async def upload_file(
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    return await upload_file_handler(
+    result = await upload_file_handler(
         request,
         file=file,
         metadata=metadata,
@@ -237,6 +290,27 @@ async def upload_file(
         db=db,
     )
 
+    if isinstance(result, dict):
+        result_id = result.get('id')
+        result_filename = result.get('filename')
+        result_meta = result.get('meta') or {}
+    else:
+        result_id = result.id
+        result_filename = result.filename
+        result_meta = result.meta or {}
+
+    result_content_type = (
+        result_meta.get('content_type') if isinstance(result_meta, dict) else getattr(result_meta, 'content_type', None)
+    )
+    await publish_event(
+        request,
+        EVENTS.FILE_UPLOADED,
+        actor=user,
+        subject_id=result_id,
+        data={'filename': result_filename, 'content_type': result_content_type},
+    )
+    return result
+
 
 async def upload_file_handler(
     request: Request,
@@ -268,36 +342,59 @@ async def upload_file_handler(
         # Remove the leading dot from the file extension and lowercase it
         file_extension = file_extension[1:].lower() if file_extension else ''
 
-        if process and request.app.state.config.ALLOWED_FILE_EXTENSIONS:
-            request.app.state.config.ALLOWED_FILE_EXTENSIONS = [
-                ext for ext in request.app.state.config.ALLOWED_FILE_EXTENSIONS if ext
-            ]
+        allowed_file_extensions = await Config.get('rag.file.allowed_extensions')
+        if process and allowed_file_extensions:
+            allowed_file_extensions = [ext for ext in allowed_file_extensions if ext]
 
-            if file_extension not in request.app.state.config.ALLOWED_FILE_EXTENSIONS:
+            if file_extension not in allowed_file_extensions:
                 raise HTTPException(
                     status_code=status.HTTP_400_BAD_REQUEST,
                     detail=ERROR_MESSAGES.DEFAULT(f'File type {file_extension} is not allowed'),
                 )
 
-        # replace filename with uuid
+        # Prefer readable storage names for admins, but fall back if the filesystem rejects it.
         id = str(uuid.uuid4())
         name = filename
         filename = f'{id}_{filename}'
-        contents, file_path = await asyncio.to_thread(
-            Storage.upload_file,
-            file.file,
-            filename,
-            {
-                'OpenWebUI-User-Email': user.email,
-                'OpenWebUI-User-Id': user.id,
-                'OpenWebUI-User-Name': user.name,
-                'OpenWebUI-File-Id': id,
-            },
-        )
+        tags = {
+            'OpenWebUI-User-Email': user.email,
+            'OpenWebUI-User-Id': user.id,
+            'OpenWebUI-User-Name': user.name,
+            'OpenWebUI-File-Id': id,
+        }
+        try:
+            contents, file_path = await asyncio.to_thread(Storage.upload_file, file.file, filename, tags)
+        except OSError as e:
+            if e.errno != errno.ENAMETOOLONG:
+                log.exception(e)
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=ERROR_MESSAGES.DEFAULT(e.strerror or 'Error uploading file'),
+                )
+
+            file.file.seek(0)
+            filename = f'{id}.{file_extension}' if file_extension else id
+            try:
+                contents, file_path = await asyncio.to_thread(Storage.upload_file, file.file, filename, tags)
+            except OSError as e:
+                log.exception(e)
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=ERROR_MESSAGES.DEFAULT(e.strerror or 'Error uploading file'),
+                )
+        max_size = await Config.get('rag.file.max_size')
+        if max_size and len(contents) > int(max_size) * 1024 * 1024:
+            await asyncio.to_thread(Storage.delete_file, file_path)
+            raise HTTPException(
+                status_code=status.HTTP_413_REQUEST_ENTITY_TOO_LARGE,
+                detail=ERROR_MESSAGES.FILE_TOO_LARGE(size=f'{max_size} MB'),
+            )
 
         # SHA-256 of raw uploaded bytes for incremental sync diffing.
         # If the client pre-computed and sent file_hash, use that.
-        file_hash = file_metadata.get('file_hash') or hashlib.sha256(contents).hexdigest()
+        file_hash = file_metadata.get('file_hash') or await asyncio.to_thread(
+            lambda: hashlib.sha256(contents).hexdigest()
+        )
 
         file_item = await Files.insert_new_file(
             user.id,
@@ -443,13 +540,29 @@ async def search_files(
     return files
 
 
+############################
+# Count Files
+############################
+
+
+@router.get('/count', response_model=int)
+async def count_files(
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
+):
+    user_id = None if (user.role == 'admin' and BYPASS_ADMIN_ACCESS_CONTROL) else user.id
+    return await Files.count_files_by_user_id(user_id=user_id, db=db)
+
+
 ############################
 # Delete All Files
 ############################
 
 
 @router.delete('/all')
-async def delete_all_files(user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+async def delete_all_files(
+    request: Request, user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)
+):
     result = await Files.delete_all_files(db=db)
     if result:
         try:
@@ -462,6 +575,7 @@ async def delete_all_files(user=Depends(get_admin_user), db: AsyncSession = Depe
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error deleting files'),
             )
+        await publish_event(request, EVENTS.FILE_DELETED_ALL, actor=user, subject_type='file')
         return {'message': 'All files deleted successfully'}
     else:
         raise HTTPException(
@@ -605,6 +719,12 @@ async def update_file_data_content_by_id(
         )
 
     if file.user_id == user.id or user.role == 'admin' or await has_access_to_file(id, 'write', user, db=db):
+        max_size = await Config.get('rag.file.max_size')
+        if max_size and len(form_data.content.encode('utf-8')) > int(max_size) * 1024 * 1024:
+            raise HTTPException(
+                status_code=status.HTTP_413_REQUEST_ENTITY_TOO_LARGE,
+                detail=ERROR_MESSAGES.FILE_TOO_LARGE(size=f'{max_size} MB'),
+            )
         try:
             await process_file(
                 request,
@@ -623,18 +743,29 @@ async def update_file_data_content_by_id(
         knowledges = await Knowledges.get_knowledges_by_file_id(id, db=db)
         for knowledge in knowledges:
             try:
-                # Remove old embeddings for this file from the KB collection
-                await ASYNC_VECTOR_DB_CLIENT.delete(collection_name=knowledge.id, filter={'file_id': id})
-                # Re-add from the now-updated file-{file_id} collection
+                old_vectors = await ASYNC_VECTOR_DB_CLIENT.query(collection_name=knowledge.id, filter={'file_id': id})
+                old_vector_ids = old_vectors.ids[0] if old_vectors and old_vectors.ids else []
+
+                # Re-add from the now-updated file-{file_id} collection before
+                # removing old vectors, so a failed reindex keeps the KB usable.
                 await process_file(
                     request,
                     ProcessFileForm(file_id=id, collection_name=knowledge.id),
                     user=user,
                     db=db,
                 )
+                if old_vector_ids:
+                    await ASYNC_VECTOR_DB_CLIENT.delete(collection_name=knowledge.id, ids=old_vector_ids)
             except Exception as e:
                 log.warning(f'Failed to update knowledge {knowledge.id} after content change for file {id}: {e}')
 
+        await publish_event(
+            request,
+            EVENTS.FILE_CONTENT_UPDATED,
+            actor=user,
+            subject_id=id,
+            data={'content_preview': form_data.content[:300]},
+        )
         return {'content': file.data.get('content', '')}
     else:
         raise HTTPException(
@@ -824,6 +955,7 @@ class FileRenameForm(BaseModel):
 
 @router.post('/{id}/rename')
 async def rename_file_by_id(
+    request: Request,
     id: str,
     form_data: FileRenameForm,
     user=Depends(get_verified_user),
@@ -840,6 +972,13 @@ async def rename_file_by_id(
     if file.user_id == user.id or user.role == 'admin' or await has_access_to_file(id, 'write', user, db=db):
         result = await Files.update_file_name_by_id(id, form_data.filename, db=db)
         if result:
+            await publish_event(
+                request,
+                EVENTS.FILE_RENAMED,
+                actor=user,
+                subject_id=id,
+                data={'filename': form_data.filename},
+            )
             return result
         else:
             raise HTTPException(
@@ -859,7 +998,9 @@ async def rename_file_by_id(
 
 
 @router.delete('/{id}')
-async def delete_file_by_id(id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)):
+async def delete_file_by_id(
+    request: Request, id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)
+):
     file = await Files.get_file_by_id(id, db=db)
 
     if not file:
@@ -894,6 +1035,13 @@ async def delete_file_by_id(id: str, user=Depends(get_verified_user), db: AsyncS
                     status_code=status.HTTP_400_BAD_REQUEST,
                     detail=ERROR_MESSAGES.DEFAULT('Error deleting files'),
                 )
+            await publish_event(
+                request,
+                EVENTS.FILE_DELETED,
+                actor=user,
+                subject_id=id,
+                data={'filename': file.filename},
+            )
             return {'message': 'File deleted successfully'}
         else:
             raise HTTPException(

backend/open_webui/routers/folders.py

@@ -6,11 +6,14 @@ import uuid
 from pathlib import Path
 from typing import Optional
 
-from fastapi import APIRouter, Depends, File, HTTPException, Request, UploadFile, status
+from fastapi import APIRouter, Depends, File, HTTPException, Query, Request, UploadFile, status
 from fastapi.responses import FileResponse, StreamingResponse
 from open_webui.config import UPLOAD_DIR
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.internal.db import get_async_session
+from open_webui.models.chat_messages import ChatMessages
+from open_webui.models.config import Config
 from open_webui.models.chats import Chats
 from open_webui.models.folders import (
     FolderForm,
@@ -19,9 +22,17 @@ from open_webui.models.folders import (
     Folders,
     FolderUpdateForm,
 )
+from open_webui.models.access_grants import AccessGrants
+from open_webui.models.automations import Automations
+from open_webui.models.groups import Groups
+from open_webui.models.users import Users
 from open_webui.utils.access_control import has_permission
-from open_webui.utils.access_control.files import get_accessible_folder_files
+from open_webui.utils.access_control import (
+    filter_allowed_access_grants,
+)
+from open_webui.utils.access_control.files import can_read_all_folder_files, get_accessible_folder_files
 from open_webui.utils.auth import get_admin_user, get_verified_user
+from open_webui.tasks import has_active_tasks
 from pydantic import BaseModel
 from sqlalchemy.ext.asyncio import AsyncSession
 
@@ -31,6 +42,47 @@ log = logging.getLogger(__name__)
 router = APIRouter()
 
 
+from open_webui.utils.access_control.folders import has_folder_access as _has_folder_access
+
+
+async def get_folder_unread_counts(user_id: str, db: AsyncSession | None = None) -> dict[str, int]:
+    folders = await Folders.get_folders_by_user_id(user_id, db=db)
+    parent_by_id = {folder.id: folder.parent_id for folder in folders}
+    unread_counts = dict.fromkeys(parent_by_id.keys(), 0)
+    direct_unread_counts = await Chats.count_unread_by_folder_ids(user_id, list(parent_by_id.keys()), db=db)
+
+    for unread_folder_id, unread_count in direct_unread_counts.items():
+        current_id = unread_folder_id
+        seen = set()
+        while current_id and current_id not in seen:
+            seen.add(current_id)
+            if current_id in unread_counts:
+                unread_counts[current_id] += unread_count
+            current_id = parent_by_id.get(current_id)
+
+    return unread_counts
+
+
+async def check_folders_permission(request: Request, user, db=None):
+    """Verify the folders feature is enabled and the user has permission."""
+    config = await Config.get_many('folders.enable', 'user.permissions')
+    if config.get('folders.enable') is False:
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+    if user.role != 'admin' and not await has_permission(
+        user.id,
+        'features.folders',
+        config.get('user.permissions'),
+        db=db,
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+
+
 ############################
 # Get Folders
 ############################
@@ -42,29 +94,15 @@ async def get_folders(
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    if request.app.state.config.ENABLE_FOLDERS is False:
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-    if user.role != 'admin' and not await has_permission(
-        user.id,
-        'features.folders',
-        request.app.state.config.USER_PERMISSIONS,
-        db=db,
-    ):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    await check_folders_permission(request, user, db=db)
 
     folders = await Folders.get_folders_by_user_id(user.id, db=db)
+    folder_ids = {folder.id for folder in folders}
 
     # Verify folder data integrity
     folder_list = []
     for folder in folders:
-        if folder.parent_id and not await Folders.get_folder_by_id_and_user_id(folder.parent_id, user.id, db=db):
+        if folder.parent_id and folder.parent_id not in folder_ids:
             folder = await Folders.update_folder_parent_id_by_id_and_user_id(folder.id, user.id, None, db=db)
 
         if folder.data and 'files' in folder.data:
@@ -75,9 +113,14 @@ async def get_folders(
                     folder.id, user.id, FolderUpdateForm(data=folder.data), db=db
                 )
 
-        folder_list.append(FolderNameIdResponse(**folder.model_dump()))
+        folder_list.append(folder)
 
-    return folder_list
+    unread_counts = await get_folder_unread_counts(user.id, db=db)
+
+    return [
+        FolderNameIdResponse(**folder.model_dump(), unread_count=unread_counts.get(folder.id, 0))
+        for folder in folder_list
+    ]
 
 
 ############################
@@ -87,10 +130,12 @@ async def get_folders(
 
 @router.post('/')
 async def create_folder(
+    request: Request,
     form_data: FolderForm,
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
+    await check_folders_permission(request, user, db=db)
     folder = await Folders.get_folder_by_parent_id_and_user_id_and_name(
         form_data.parent_id, user.id, form_data.name, db=db
     )
@@ -101,8 +146,65 @@ async def create_folder(
             detail=ERROR_MESSAGES.DEFAULT('Folder already exists'),
         )
 
+    # Check if creating a subfolder in a shared folder
+    if form_data.parent_id:
+        parent = await Folders.get_folder_by_id(form_data.parent_id, db=db)
+        if parent and parent.user_id != user.id:
+            # Creating subfolder in someone else's shared folder
+            if user.role != 'admin' and not await _has_folder_access(user.id, parent, 'write', db):
+                raise HTTPException(
+                    status_code=status.HTTP_403_FORBIDDEN,
+                    detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+                )
+            if form_data.data and 'files' in form_data.data:
+                owner = await Users.get_user_by_id(parent.user_id, db=db)
+                if not owner:
+                    raise HTTPException(
+                        status_code=status.HTTP_404_NOT_FOUND,
+                        detail=ERROR_MESSAGES.NOT_FOUND,
+                    )
+                if not await can_read_all_folder_files(form_data.data['files'], owner, db=db):
+                    raise HTTPException(
+                        status_code=status.HTTP_403_FORBIDDEN,
+                        detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+                    )
+            # Create as the folder owner's subfolder (keep tree consistent)
+            try:
+                folder = await Folders.insert_new_folder(parent.user_id, form_data, form_data.parent_id, db=db)
+                await publish_event(
+                    request,
+                    EVENTS.FOLDER_CREATED,
+                    actor=user,
+                    subject_id=folder.id,
+                    data={'name': folder.name, 'parent_id': folder.parent_id, 'owner_id': folder.user_id},
+                )
+                return folder
+            except Exception as e:
+                log.exception(e)
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=ERROR_MESSAGES.DEFAULT('Error creating folder'),
+                )
+
+    if (
+        form_data.data
+        and 'files' in form_data.data
+        and not await can_read_all_folder_files(form_data.data['files'], user, db=db)
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+
     try:
         folder = await Folders.insert_new_folder(user.id, form_data, form_data.parent_id, db=db)
+        await publish_event(
+            request,
+            EVENTS.FOLDER_CREATED,
+            actor=user,
+            subject_id=folder.id,
+            data={'name': folder.name, 'parent_id': folder.parent_id, 'owner_id': folder.user_id},
+        )
         return folder
     except Exception as e:
         log.exception(e)
@@ -113,21 +215,90 @@ async def create_folder(
         )
 
 
+############################
+# Get Shared Folders
+############################
+
+
+@router.get('/shared')
+async def get_shared_folders(
+    request: Request,
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
+):
+    """Get all folders shared with the current user (not owned by them)."""
+    await check_folders_permission(request, user, db=db)
+    groups = await Groups.get_groups_by_member_id(user.id, db=db)
+    group_ids = {g.id for g in groups}
+
+    folder_perms = await Folders.get_shared_folder_ids_for_user(user.id, group_ids, db=db)
+
+    # Filter out folders owned by the user
+    results = []
+    owner_cache = {}
+    for folder_id, permission in folder_perms.items():
+        folder = await Folders.get_folder_by_id(folder_id, db=db)
+        if not folder or folder.user_id == user.id:
+            continue
+
+        # Get owner name (cached)
+        if folder.user_id not in owner_cache:
+            owner = await Users.get_user_by_id(folder.user_id, db=db)
+            owner_cache[folder.user_id] = owner.name if owner else 'Unknown'
+
+        results.append(
+            {
+                **folder.model_dump(),
+                'owner_name': owner_cache[folder.user_id],
+                'permission': permission,
+            }
+        )
+
+    # Also include child folders of shared folders (inheritance)
+    shared_root_ids = {r['id'] for r in results}
+    for root_id in list(shared_root_ids):
+        root_folder = await Folders.get_folder_by_id(root_id, db=db)
+        if root_folder:
+            children = await Folders.get_children_folders_by_id_and_user_id(root_id, root_folder.user_id, db=db)
+            if children:
+                for child in children:
+                    if child.id not in {r['id'] for r in results}:
+                        results.append(
+                            {
+                                **child.model_dump(),
+                                'owner_name': owner_cache.get(child.user_id, 'Unknown'),
+                                'permission': folder_perms.get(root_id, 'read'),
+                            }
+                        )
+
+    return results
+
+
 ############################
 # Get Folders By Id
 ############################
 
 
-@router.get('/{id}', response_model=Optional[FolderModel])
-async def get_folder_by_id(id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)):
+@router.get('/{id}', response_model=None)
+async def get_folder_by_id(
+    request: Request, id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)
+):
+    await check_folders_permission(request, user, db=db)
     folder = await Folders.get_folder_by_id_and_user_id(id, user.id, db=db)
     if folder:
-        return folder
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
+        grants = await AccessGrants.get_grants_by_resource('folder', id, db=db)
+        return {**folder.model_dump(), 'access_grants': [g.model_dump() for g in grants]}
+
+    # Check shared access
+    folder = await Folders.get_folder_by_id(id, db=db)
+    if folder and (user.role == 'admin' or await _has_folder_access(user.id, folder, 'read', db)):
+        grants = await AccessGrants.get_grants_by_resource('folder', id, db=db)
+        return {**folder.model_dump(), 'access_grants': [g.model_dump() for g in grants]}
+
+    raise HTTPException(
+        status_code=status.HTTP_404_NOT_FOUND,
+        detail=ERROR_MESSAGES.NOT_FOUND,
+    )
 
 
 ############################
@@ -137,17 +308,28 @@ async def get_folder_by_id(id: str, user=Depends(get_verified_user), db: AsyncSe
 
 @router.post('/{id}/update')
 async def update_folder_name_by_id(
+    request: Request,
     id: str,
     form_data: FolderUpdateForm,
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
+    await check_folders_permission(request, user, db=db)
     folder = await Folders.get_folder_by_id_and_user_id(id, user.id, db=db)
+    if not folder:
+        # Check shared write access
+        folder = await Folders.get_folder_by_id(id, db=db)
+        if not folder or (user.role != 'admin' and not await _has_folder_access(user.id, folder, 'write', db)):
+            raise HTTPException(
+                status_code=status.HTTP_404_NOT_FOUND,
+                detail=ERROR_MESSAGES.NOT_FOUND,
+            )
+
     if folder:
         if form_data.name is not None:
             # Check if folder with same name exists
             existing_folder = await Folders.get_folder_by_parent_id_and_user_id_and_name(
-                folder.parent_id, user.id, form_data.name, db=db
+                folder.parent_id, folder.user_id, form_data.name, db=db
             )
             if existing_folder and existing_folder.id != id:
                 raise HTTPException(
@@ -155,18 +337,28 @@ async def update_folder_name_by_id(
                     detail=ERROR_MESSAGES.DEFAULT('Folder already exists'),
                 )
 
-        # Validate read access to every file/collection being attached.
-        # Folder files are consumed by chat middleware as RAG context.
-        if form_data.data and isinstance(form_data.data.get('files'), list):
-            accessible_files = await get_accessible_folder_files(form_data.data['files'], user, db=db)
-            if len(accessible_files) != len(form_data.data['files']):
+        if form_data.data and 'files' in form_data.data:
+            owner = user if folder.user_id == user.id else await Users.get_user_by_id(folder.user_id, db=db)
+            if not owner:
+                raise HTTPException(
+                    status_code=status.HTTP_404_NOT_FOUND,
+                    detail=ERROR_MESSAGES.NOT_FOUND,
+                )
+            if not await can_read_all_folder_files(form_data.data['files'], owner, db=db):
                 raise HTTPException(
                     status_code=status.HTTP_403_FORBIDDEN,
                     detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
                 )
 
         try:
-            folder = await Folders.update_folder_by_id_and_user_id(id, user.id, form_data, db=db)
+            folder = await Folders.update_folder_by_id_and_user_id(id, folder.user_id, form_data, db=db)
+            await publish_event(
+                request,
+                EVENTS.FOLDER_UPDATED,
+                actor=user,
+                subject_id=id,
+                data={'name': folder.name},
+            )
             return folder
         except Exception as e:
             log.exception(e)
@@ -175,11 +367,6 @@ async def update_folder_name_by_id(
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error updating folder'),
             )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
 
 
 ############################
@@ -193,11 +380,13 @@ class FolderParentIdForm(BaseModel):
 
 @router.post('/{id}/update/parent')
 async def update_folder_parent_id_by_id(
+    request: Request,
     id: str,
     form_data: FolderParentIdForm,
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
+    await check_folders_permission(request, user, db=db)
     folder = await Folders.get_folder_by_id_and_user_id(id, user.id, db=db)
     if folder:
         existing_folder = await Folders.get_folder_by_parent_id_and_user_id_and_name(
@@ -212,6 +401,13 @@ async def update_folder_parent_id_by_id(
 
         try:
             folder = await Folders.update_folder_parent_id_by_id_and_user_id(id, user.id, form_data.parent_id, db=db)
+            await publish_event(
+                request,
+                EVENTS.FOLDER_PARENT_UPDATED,
+                actor=user,
+                subject_id=id,
+                data={'parent_id': form_data.parent_id},
+            )
             return folder
         except Exception as e:
             log.exception(e)
@@ -238,12 +434,19 @@ class FolderIsExpandedForm(BaseModel):
 
 @router.post('/{id}/update/expanded')
 async def update_folder_is_expanded_by_id(
+    request: Request,
     id: str,
     form_data: FolderIsExpandedForm,
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
+    await check_folders_permission(request, user, db=db)
     folder = await Folders.get_folder_by_id_and_user_id(id, user.id, db=db)
+    if not folder:
+        folder = await Folders.get_folder_by_id(id, db=db)
+        if folder and (user.role == 'admin' or await _has_folder_access(user.id, folder, 'read', db)):
+            return folder
+
     if folder:
         try:
             folder = await Folders.update_folder_is_expanded_by_id_and_user_id(
@@ -264,6 +467,173 @@ async def update_folder_is_expanded_by_id(
         )
 
 
+############################
+# Update Folder Access By Id
+############################
+
+
+class FolderAccessGrantsForm(BaseModel):
+    access_grants: list[dict]
+
+
+@router.post('/{id}/access/update')
+async def update_folder_access_by_id(
+    request: Request,
+    id: str,
+    form_data: FolderAccessGrantsForm,
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
+):
+    await check_folders_permission(request, user, db=db)
+    folder = await Folders.get_folder_by_id(id, db=db)
+    if not folder:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+    # Only owner, admin, or write-granted user can update access
+    if user.role != 'admin' and user.id != folder.user_id:
+        if not await _has_folder_access(user.id, folder, 'write', db):
+            raise HTTPException(
+                status_code=status.HTTP_403_FORBIDDEN,
+                detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+            )
+
+    form_data.access_grants = await filter_allowed_access_grants(
+        await Config.get('user.permissions'),
+        user.id,
+        user.role,
+        form_data.access_grants,
+        None,
+        db=db,
+    )
+
+    await AccessGrants.set_access_grants('folder', id, form_data.access_grants, db=db)
+
+    grants = await AccessGrants.get_grants_by_resource('folder', id, db=db)
+    await publish_event(
+        request,
+        EVENTS.FOLDER_ACCESS_UPDATED,
+        actor=user,
+        subject_id=id,
+        data={'grant_count': len(grants)},
+    )
+    return {
+        **folder.model_dump(),
+        'access_grants': [g.model_dump() for g in grants],
+    }
+
+
+############################
+# Get Shared Folder Chats
+############################
+
+
+@router.get('/{id}/shared/chats')
+async def get_shared_folder_chats(
+    request: Request,
+    id: str,
+    page: int | None = Query(None, ge=1),
+    sort_by: str = Query('unread_updated_at'),
+    sort_dir: str = Query('desc'),
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
+):
+    """Get chats within a shared folder. Returns readonly flag based on permission."""
+    await check_folders_permission(request, user, db=db)
+    folder = await Folders.get_folder_by_id(id, db=db)
+    if not folder:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+    is_owner = user.id == folder.user_id
+    is_admin = user.role == 'admin'
+    has_write = is_owner or is_admin or await _has_folder_access(user.id, folder, 'write', db)
+    has_read = has_write or await _has_folder_access(user.id, folder, 'read', db)
+
+    if not has_read:
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+
+    limit = 10
+    skip = (page - 1) * limit if page is not None else 0
+    chats = await Chats.get_all_chats_by_folder_id(
+        id,
+        skip=skip,
+        limit=limit if page is not None else 60,
+        sort_by=sort_by,
+        sort_dir=sort_dir,
+        unread_for_user_id=user.id,
+        db=db,
+    )
+    total = await Chats.count_all_chats_by_folder_id(id, db=db) if page is not None else len(chats)
+
+    # Resolve owner names for display (avatar URLs are constructed client-side)
+    owner_cache: dict[str, str] = {}
+    for chat in chats:
+        uid = chat['user_id']
+        if uid not in owner_cache:
+            u = await Users.get_user_by_id(uid, db=db)
+            owner_cache[uid] = u.name if u else 'Unknown'
+        chat['owner_name'] = owner_cache[uid]
+        chat['active'] = False
+        if chat['user_id'] != user.id:
+            chat['last_read_at'] = chat['updated_at']
+        if await has_active_tasks(request.app.state.redis, chat['id']):
+            chat['active'] = await ChatMessages.has_unfinished_assistant_by_chat_id(chat['id'], db=db)
+
+    response = {
+        'chats': [{**chat, 'readonly': chat['user_id'] != user.id} for chat in chats],
+        'folder_permission': 'write' if has_write else 'read',
+    }
+    if page is not None:
+        response.update({'total': total, 'has_more': skip + limit < total})
+    return response
+
+
+@router.post('/{id}/read')
+async def mark_folder_chats_read_by_id(
+    request: Request,
+    id: str,
+    user=Depends(get_verified_user),
+    db: AsyncSession = Depends(get_async_session),
+):
+    await check_folders_permission(request, user, db=db)
+    folder = await Folders.get_folder_by_id(id, db=db)
+    if not folder:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+    is_owner = user.id == folder.user_id
+    is_admin = user.role == 'admin'
+    if not (is_owner or is_admin or await _has_folder_access(user.id, folder, 'read', db)):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+
+    folder_ids = (
+        await Folders.get_folder_ids_by_id_and_user_id_in_subtree(id, folder.user_id, db=db)
+        if is_owner or is_admin
+        else [id]
+    )
+    updated_count = await Chats.mark_chats_read_by_folder_ids(user.id, folder_ids, db=db)
+
+    return {
+        'folder_id': id,
+        'folder_ids': folder_ids,
+        'updated_count': updated_count,
+        'folder_unread_counts': await get_folder_unread_counts(user.id, db=db),
+    }
+
+
 ############################
 # Delete Folder By Id
 ############################
@@ -277,9 +647,29 @@ async def delete_folder_by_id(
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    if await Chats.count_chats_by_folder_id_and_user_id(id, user.id, db=db):
+    await check_folders_permission(request, user, db=db)
+    folder = await Folders.get_folder_by_id_and_user_id(id, user.id, db=db)
+
+    if not folder:
+        # Deletion cascades into the owner's data, so only the owner or an admin may delete
+        folder = await Folders.get_folder_by_id(id, db=db)
+        if not folder:
+            raise HTTPException(
+                status_code=status.HTTP_404_NOT_FOUND,
+                detail=ERROR_MESSAGES.NOT_FOUND,
+            )
+        if user.role != 'admin':
+            raise HTTPException(
+                status_code=status.HTTP_403_FORBIDDEN,
+                detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+            )
+
+    folder_owner_id = folder.user_id
+
+    folder_ids = await Folders.get_folder_ids_by_id_and_user_id_in_subtree(id, folder_owner_id, db=db)
+    if await Chats.count_chats_by_folder_ids_and_user_id(folder_ids, folder_owner_id, db=db):
         chat_delete_permission = await has_permission(
-            user.id, 'chat.delete', request.app.state.config.USER_PERMISSIONS, db=db
+            user.id, 'chat.delete', await Config.get('user.permissions'), db=db
         )
         if user.role != 'admin' and not chat_delete_permission:
             raise HTTPException(
@@ -288,19 +678,31 @@ async def delete_folder_by_id(
             )
 
     folders = []
-    folders.append(await Folders.get_folder_by_id_and_user_id(id, user.id, db=db))
+    folders.append(folder)
     while folders:
         folder = folders.pop()
         if folder:
             try:
-                folder_ids = await Folders.delete_folder_by_id_and_user_id(folder.id, user.id, db=db)
+                folder_ids = await Folders.delete_folder_by_id_and_user_id(folder.id, folder_owner_id, db=db)
 
                 for folder_id in folder_ids:
                     if delete_contents:
-                        await Chats.delete_chats_by_user_id_and_folder_id(user.id, folder_id, db=db)
+                        await Chats.delete_chats_by_user_id_and_folder_id(folder_owner_id, folder_id, db=db)
                     else:
-                        await Chats.move_chats_by_user_id_and_folder_id(user.id, folder_id, None, db=db)
+                        await Chats.move_chats_by_user_id_and_folder_id(folder_owner_id, folder_id, None, db=db)
 
+                    # Clean up access grants for this folder
+                    await AccessGrants.revoke_all_access('folder', folder_id, db=db)
+
+                await Automations.clear_folder_ids(folder_owner_id, folder_ids, db=db)
+
+                await publish_event(
+                    request,
+                    EVENTS.FOLDER_DELETED,
+                    actor=user,
+                    subject_id=id,
+                    data={'folder_ids': folder_ids, 'delete_contents': delete_contents},
+                )
                 return True
             except Exception as e:
                 log.exception(e)
@@ -311,7 +713,7 @@ async def delete_folder_by_id(
                 )
             finally:
                 # Get all subfolders
-                subfolders = await Folders.get_folders_by_parent_id_and_user_id(folder.id, user.id, db=db)
+                subfolders = await Folders.get_folders_by_parent_id_and_user_id(folder.id, folder_owner_id, db=db)
                 folders.extend(subfolders)
 
     else:

backend/open_webui/routers/functions.py

@@ -10,7 +10,8 @@ import aiohttp
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from open_webui.config import CACHE_DIR
 from open_webui.constants import ERROR_MESSAGES
-from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, AIOHTTP_CLIENT_TIMEOUT
+from open_webui.env import AIOHTTP_CLIENT_SESSION_SSL, AIOHTTP_CLIENT_TIMEOUT, ENABLE_PLUGINS
+from open_webui.events import EVENTS, build_event, dispatch_event_functions, publish_event, schedule_webhook_dispatch
 from open_webui.internal.db import get_async_session
 from open_webui.models.functions import (
     FunctionForm,
@@ -22,6 +23,7 @@ from open_webui.models.functions import (
 )
 from open_webui.utils.auth import get_admin_user, get_verified_user
 from open_webui.utils.plugin import (
+    get_functions_cache,
     get_function_module_from_cache,
     load_function_module_by_id,
     replace_imports,
@@ -44,11 +46,17 @@ router = APIRouter()
 
 @router.get('/', response_model=list[FunctionResponse])
 async def get_functions(user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)):
+    if not ENABLE_PLUGINS:
+        return []
+
     return await Functions.get_functions(db=db)
 
 
 @router.get('/list', response_model=list[FunctionUserResponse])
 async def get_function_list(user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+    if not ENABLE_PLUGINS:
+        return []
+
     return await Functions.get_function_list(db=db)
 
 
@@ -63,6 +71,9 @@ async def get_functions(
     user=Depends(get_admin_user),
     db: AsyncSession = Depends(get_async_session),
 ):
+    if not ENABLE_PLUGINS:
+        return []
+
     return await Functions.get_functions(include_valves=include_valves, db=db)
 
 
@@ -130,8 +141,13 @@ async def load_function_from_url(request: Request, form_data: LoadUrlForm, user=
             'name': function_name,
             'content': data,
         }
+    except HTTPException:
+        raise
     except Exception as e:
-        raise HTTPException(status_code=500, detail=ERROR_MESSAGES.DEFAULT(e))
+        raise HTTPException(
+            status_code=500,
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error fetching function'),
+        )
 
 
 ############################
@@ -171,7 +187,7 @@ async def sync_functions(
         log.exception(f'Failed to load a function: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error loading function'),
         )
 
 
@@ -205,7 +221,7 @@ async def create_new_function(
             )
             form_data.meta.manifest = frontmatter
 
-            FUNCTIONS = request.app.state.FUNCTIONS
+            FUNCTIONS = get_functions_cache(request)
             FUNCTIONS[form_data.id] = function_module
 
             function = await Functions.insert_new_function(user.id, function_type, form_data, db=db)
@@ -217,17 +233,26 @@ async def create_new_function(
                 await Functions.update_function_metadata_by_id(form_data.id, {'toggle': True}, db=db)
 
             if function:
+                await publish_event(
+                    request,
+                    EVENTS.FUNCTION_CREATED,
+                    actor=user,
+                    subject_id=function.id,
+                    data={'type': function.type, 'name': function.name},
+                )
                 return function
             else:
                 raise HTTPException(
                     status_code=status.HTTP_400_BAD_REQUEST,
                     detail=ERROR_MESSAGES.DEFAULT('Error creating function'),
                 )
+        except HTTPException:
+            raise
         except Exception as e:
             log.exception(f'Failed to create a new function: {e}')
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.DEFAULT(e),
+                detail=ERROR_MESSAGES.DEFAULT(e, 'Error creating function'),
             )
     else:
         raise HTTPException(
@@ -260,12 +285,41 @@ async def get_function_by_id(id: str, user=Depends(get_admin_user), db: AsyncSes
 
 
 @router.post('/id/{id}/toggle', response_model=FunctionModel | None)
-async def toggle_function_by_id(id: str, user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+async def toggle_function_by_id(
+    request: Request,
+    id: str,
+    user=Depends(get_admin_user),
+    db: AsyncSession = Depends(get_async_session),
+):
     function = await Functions.get_function_by_id(id, db=db)
     if function:
+        lifecycle_event = build_event(
+            request,
+            EVENTS.FUNCTION_DISABLE_STARTED if function.is_active else EVENTS.FUNCTION_ENABLE_STARTED,
+            actor=user,
+            subject_id=function.id,
+            subject_type='function',
+            data={'type': function.type, 'name': function.name},
+        )
+        await dispatch_event_functions(
+            request.app,
+            lifecycle_event,
+            request=request,
+            extra_function_ids=[function.id] if not function.is_active else None,
+        )
+        schedule_webhook_dispatch(request.app, lifecycle_event)
+
         function = await Functions.update_function_by_id(id, {'is_active': not function.is_active}, db=db)
 
         if function:
+            await publish_event(
+                request,
+                EVENTS.FUNCTION_ENABLED if function.is_active else EVENTS.FUNCTION_DISABLED,
+                actor=user,
+                subject_id=function.id,
+                subject_type='function',
+                data={'type': function.type, 'name': function.name},
+            )
             return function
         else:
             raise HTTPException(
@@ -285,12 +339,24 @@ async def toggle_function_by_id(id: str, user=Depends(get_admin_user), db: Async
 
 
 @router.post('/id/{id}/toggle/global', response_model=FunctionModel | None)
-async def toggle_global_by_id(id: str, user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+async def toggle_global_by_id(
+    request: Request,
+    id: str,
+    user=Depends(get_admin_user),
+    db: AsyncSession = Depends(get_async_session),
+):
     function = await Functions.get_function_by_id(id, db=db)
     if function:
         function = await Functions.update_function_by_id(id, {'is_global': not function.is_global}, db=db)
 
         if function:
+            await publish_event(
+                request,
+                EVENTS.FUNCTION_UPDATED,
+                actor=user,
+                subject_id=function.id,
+                data={'type': function.type, 'name': function.name, 'is_global': function.is_global},
+            )
             return function
         else:
             raise HTTPException(
@@ -322,7 +388,7 @@ async def update_function_by_id(
         function_module, function_type, frontmatter = await load_function_module_by_id(id, content=form_data.content)
         form_data.meta.manifest = frontmatter
 
-        FUNCTIONS = request.app.state.FUNCTIONS
+        FUNCTIONS = get_functions_cache(request)
         FUNCTIONS[id] = function_module
 
         updated = {**form_data.model_dump(exclude={'id'}), 'type': function_type}
@@ -334,6 +400,13 @@ async def update_function_by_id(
             await Functions.update_function_metadata_by_id(id, {'toggle': True}, db=db)
 
         if function:
+            await publish_event(
+                request,
+                EVENTS.FUNCTION_UPDATED,
+                actor=user,
+                subject_id=function.id,
+                data={'type': function.type, 'name': function.name},
+            )
             return function
         else:
             raise HTTPException(
@@ -341,10 +414,12 @@ async def update_function_by_id(
                 detail=ERROR_MESSAGES.DEFAULT('Error updating function'),
             )
 
+    except HTTPException:
+        raise
     except Exception as e:
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error updating function'),
         )
 
 
@@ -363,9 +438,14 @@ async def delete_function_by_id(
     result = await Functions.delete_function_by_id(id, db=db)
 
     if result:
-        FUNCTIONS = request.app.state.FUNCTIONS
-        if id in FUNCTIONS:
-            del FUNCTIONS[id]
+        FUNCTIONS = get_functions_cache(request)
+        FUNCTIONS.pop(id, None)
+        await publish_event(
+            request,
+            EVENTS.FUNCTION_DELETED,
+            actor=user,
+            subject_id=id,
+        )
 
     return result
 
@@ -387,7 +467,7 @@ async def get_function_valves_by_id(
         except Exception as e:
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.DEFAULT(e),
+                detail=ERROR_MESSAGES.DEFAULT(e, 'Error getting function valves'),
             )
     else:
         raise HTTPException(
@@ -452,12 +532,18 @@ async def update_function_valves_by_id(
 
                 valves_dict = valves.model_dump(exclude_unset=True)
                 await Functions.update_function_valves_by_id(id, valves_dict, db=db)
+                await publish_event(
+                    request,
+                    EVENTS.FUNCTION_VALVES_UPDATED,
+                    actor=user,
+                    subject_id=id,
+                )
                 return valves_dict
             except Exception as e:
                 log.exception(f'Error updating function values by id {id}: {e}')
                 raise HTTPException(
                     status_code=status.HTTP_400_BAD_REQUEST,
-                    detail=ERROR_MESSAGES.DEFAULT(e),
+                    detail=ERROR_MESSAGES.DEFAULT(e, 'Error updating function valves'),
                 )
         else:
             raise HTTPException(
@@ -489,7 +575,7 @@ async def get_function_user_valves_by_id(
         except Exception as e:
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.DEFAULT(e),
+                detail=ERROR_MESSAGES.DEFAULT(e, 'Error getting function user valves'),
             )
     else:
         raise HTTPException(
@@ -544,12 +630,19 @@ async def update_function_user_valves_by_id(
                 user_valves = UserValves(**form_data)
                 user_valves_dict = user_valves.model_dump(exclude_unset=True)
                 await Functions.update_user_valves_by_id_and_user_id(id, user.id, user_valves_dict, db=db)
+                await publish_event(
+                    request,
+                    EVENTS.FUNCTION_VALVES_UPDATED,
+                    actor=user,
+                    subject_id=id,
+                    data={'scope': 'user'},
+                )
                 return user_valves_dict
             except Exception as e:
                 log.exception(f'Error updating function user valves by id {id}: {e}')
                 raise HTTPException(
                     status_code=status.HTTP_400_BAD_REQUEST,
-                    detail=ERROR_MESSAGES.DEFAULT(e),
+                    detail=ERROR_MESSAGES.DEFAULT(e, 'Error updating function user valves'),
                 )
         else:
             raise HTTPException(

backend/open_webui/routers/groups.py

@@ -6,6 +6,7 @@ from typing import Optional
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from open_webui.config import CACHE_DIR
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.internal.db import get_async_session
 from open_webui.models.access_grants import AccessGrants
 from open_webui.models.groups import (
@@ -58,6 +59,7 @@ async def get_groups(
 
 @router.post('/create', response_model=Optional[GroupResponse])
 async def create_new_group(
+    request: Request,
     form_data: GroupForm,
     user=Depends(get_admin_user),
     db: AsyncSession = Depends(get_async_session),
@@ -65,6 +67,13 @@ async def create_new_group(
     try:
         group = await Groups.insert_new_group(user.id, form_data, db=db)
         if group:
+            await publish_event(
+                request,
+                EVENTS.GROUP_CREATED,
+                actor=user,
+                subject_id=group.id,
+                data={'name': group.name},
+            )
             return GroupResponse(
                 **group.model_dump(),
                 member_count=await Groups.get_group_member_count_by_id(group.id, db=db),
@@ -74,11 +83,13 @@ async def create_new_group(
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error creating group'),
             )
+    except HTTPException:
+        raise
     except Exception as e:
         log.exception(f'Error creating a new group: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error creating group'),
         )
 
 
@@ -157,7 +168,7 @@ async def get_users_in_group(id: str, user=Depends(get_admin_user), db: AsyncSes
         log.exception(f'Error adding users to group {id}: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error getting group members'),
         )
 
 
@@ -168,6 +179,7 @@ async def get_users_in_group(id: str, user=Depends(get_admin_user), db: AsyncSes
 
 @router.post('/id/{id}/update', response_model=Optional[GroupResponse])
 async def update_group_by_id(
+    request: Request,
     id: str,
     form_data: GroupUpdateForm,
     user=Depends(get_admin_user),
@@ -176,6 +188,13 @@ async def update_group_by_id(
     try:
         group = await Groups.update_group_by_id(id, form_data, db=db)
         if group:
+            await publish_event(
+                request,
+                EVENTS.GROUP_UPDATED,
+                actor=user,
+                subject_id=id,
+                data={'name': group.name},
+            )
             return GroupResponse(
                 **group.model_dump(),
                 member_count=await Groups.get_group_member_count_by_id(group.id, db=db),
@@ -185,11 +204,13 @@ async def update_group_by_id(
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error updating group'),
             )
+    except HTTPException:
+        raise
     except Exception as e:
         log.exception(f'Error updating group {id}: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error updating group'),
         )
 
 
@@ -200,6 +221,7 @@ async def update_group_by_id(
 
 @router.post('/id/{id}/users/add', response_model=Optional[GroupResponse])
 async def add_user_to_group(
+    request: Request,
     id: str,
     form_data: UserIdsForm,
     user=Depends(get_admin_user),
@@ -211,6 +233,13 @@ async def add_user_to_group(
 
         group = await Groups.add_users_to_group(id, form_data.user_ids, db=db)
         if group:
+            await publish_event(
+                request,
+                EVENTS.GROUP_MEMBER_ADDED,
+                actor=user,
+                subject_id=id,
+                data={'user_ids': form_data.user_ids},
+            )
             return GroupResponse(
                 **group.model_dump(),
                 member_count=await Groups.get_group_member_count_by_id(group.id, db=db),
@@ -220,16 +249,19 @@ async def add_user_to_group(
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error adding users to group'),
             )
+    except HTTPException:
+        raise
     except Exception as e:
         log.exception(f'Error adding users to group {id}: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error adding users to group'),
         )
 
 
 @router.post('/id/{id}/users/remove', response_model=Optional[GroupResponse])
 async def remove_users_from_group(
+    request: Request,
     id: str,
     form_data: UserIdsForm,
     user=Depends(get_admin_user),
@@ -238,6 +270,13 @@ async def remove_users_from_group(
     try:
         group = await Groups.remove_users_from_group(id, form_data.user_ids, db=db)
         if group:
+            await publish_event(
+                request,
+                EVENTS.GROUP_MEMBER_REMOVED,
+                actor=user,
+                subject_id=id,
+                data={'user_ids': form_data.user_ids},
+            )
             return GroupResponse(
                 **group.model_dump(),
                 member_count=await Groups.get_group_member_count_by_id(group.id, db=db),
@@ -247,11 +286,13 @@ async def remove_users_from_group(
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error removing users from group'),
             )
+    except HTTPException:
+        raise
     except Exception as e:
         log.exception(f'Error removing users from group {id}: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error removing users from group'),
         )
 
 
@@ -261,21 +302,31 @@ async def remove_users_from_group(
 
 
 @router.delete('/id/{id}/delete', response_model=bool)
-async def delete_group_by_id(id: str, user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+async def delete_group_by_id(
+    request: Request, id: str, user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)
+):
     try:
         result = await Groups.delete_group_by_id(id, db=db)
         if result:
+            await publish_event(
+                request,
+                EVENTS.GROUP_DELETED,
+                actor=user,
+                subject_id=id,
+            )
             return result
         else:
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
                 detail=ERROR_MESSAGES.DEFAULT('Error deleting group'),
             )
+    except HTTPException:
+        raise
     except Exception as e:
         log.exception(f'Error deleting group {id}: {e}')
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
+            detail=ERROR_MESSAGES.DEFAULT(e, 'Error deleting group'),
         )
 
 

backend/open_webui/routers/memories.py

@@ -2,16 +2,27 @@ from __future__ import annotations
 
 import asyncio
 import logging
-from typing import Optional
+from typing import Literal
 
 from fastapi import APIRouter, Depends, HTTPException, Request, status
+from open_webui.config import RAG_EMBEDDING_CONTENT_PREFIX, RAG_EMBEDDING_QUERY_PREFIX
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.internal.db import get_async_session
+from open_webui.models.config import Config
 from open_webui.models.memories import Memories, MemoryModel
 from open_webui.retrieval.vector.async_client import ASYNC_VECTOR_DB_CLIENT
-from open_webui.config import RAG_EMBEDDING_QUERY_PREFIX
 from open_webui.utils.access_control import has_permission
 from open_webui.utils.auth import get_verified_user
+from open_webui.utils.memory import (
+    clean_memory_content,
+    clean_memory_path,
+    list_memory_path_groups,
+    memory_vector_text,
+    read_memory_path_rows,
+    search_memory_rows,
+    validate_memory_operations,
+)
 from pydantic import BaseModel
 from sqlalchemy.ext.asyncio import AsyncSession
 
@@ -20,6 +31,21 @@ log = logging.getLogger(__name__)
 router = APIRouter()
 
 
+async def check_memories_permission(user):
+    config = await Config.get_many('memories.enable', 'user.permissions')
+    if not config.get('memories.enable'):
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=ERROR_MESSAGES.NOT_FOUND,
+        )
+
+    if user.role != 'admin' and not await has_permission(user.id, 'features.memories', config.get('user.permissions')):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
+        )
+
+
 ############################
 # GetMemories
 # Let what is remembered here spare someone the cost
@@ -33,17 +59,7 @@ async def get_memories(
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    await check_memories_permission(user)
 
     return await Memories.get_memories_by_user_id(user.id, db=db)
 
@@ -55,10 +71,57 @@ async def get_memories(
 
 class AddMemoryForm(BaseModel):
     content: str
+    type: Literal['user', 'context'] = 'context'
+    path: str | None = None
 
 
 class MemoryUpdateModel(BaseModel):
     content: str | None = None
+    type: Literal['user', 'context'] | None = None
+    path: str | None = None
+
+
+class MemoryOperationModel(BaseModel):
+    action: Literal['add', 'replace', 'remove', 'move']
+    id: str | None = None
+    content: str | None = None
+    type: Literal['user', 'context'] | None = None
+    path: str | None = None
+
+
+class UpdateMemoriesForm(BaseModel):
+    operations: list[MemoryOperationModel]
+    source: Literal['tool', 'background_review'] | None = None
+
+
+class SearchMemoriesForm(BaseModel):
+    query: str | None = None
+    type: Literal['user', 'context', 'all'] = 'all'
+    path: str | None = None
+    memory_id: str | None = None
+    limit: int = 20
+
+
+class ListMemoryPathsForm(BaseModel):
+    query: str | None = None
+    type: Literal['user', 'context', 'all'] = 'all'
+    limit: int = 100
+
+
+class ReadMemoryPathForm(BaseModel):
+    path: str
+    type: Literal['user', 'context', 'all'] = 'all'
+    include_children: bool = True
+    limit: int = 50
+
+
+def _memory_metadata(memory: MemoryModel) -> dict:
+    return {
+        'created_at': memory.created_at,
+        'updated_at': memory.updated_at,
+        'type': memory.type,
+        'path': memory.path,
+    }
 
 
 @router.post('/add', response_model=MemoryModel | None)
@@ -73,37 +136,131 @@ async def add_memory(
     own short-lived sessions so a connection is not held during the external
     embedding API call (``EMBEDDING_FUNCTION``), which can take 1-5+ seconds.
     """
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
+    await check_memories_permission(user)
 
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    content = clean_memory_content(form_data.content)
+    path = clean_memory_path(form_data.path)
+    memory = await Memories.insert_new_memory(
+        user.id,
+        content,
+        memory_type=form_data.type,
+        path=path,
+        meta={'created_by': 'manual'},
+    )
 
-    memory = await Memories.insert_new_memory(user.id, form_data.content)
-
-    vector = await request.app.state.EMBEDDING_FUNCTION(memory.content, user=user)
+    vector = await request.app.state.EMBEDDING_FUNCTION(
+        memory_vector_text(memory.content, memory.path), prefix=RAG_EMBEDDING_CONTENT_PREFIX, user=user
+    )
 
     await ASYNC_VECTOR_DB_CLIENT.upsert(
         collection_name=f'user-memory-{user.id}',
         items=[
             {
                 'id': memory.id,
-                'text': memory.content,
+                'text': memory_vector_text(memory.content, memory.path),
                 'vector': vector,
-                'metadata': {'created_at': memory.created_at},
+                'metadata': _memory_metadata(memory),
             }
         ],
     )
 
+    await publish_event(
+        request,
+        EVENTS.MEMORY_CREATED,
+        actor=user,
+        subject_id=memory.id,
+        data={'content_preview': memory.content[:300], 'type': memory.type, 'path': memory.path},
+    )
     return memory
 
 
+@router.post('/update', response_model=list[dict])
+async def update_memories(
+    request: Request,
+    form_data: UpdateMemoriesForm,
+    user=Depends(get_verified_user),
+):
+    await check_memories_permission(user)
+
+    operations = validate_memory_operations(form_data)
+    metadata = getattr(request.state, 'metadata', {}) or {}
+    source = form_data.source or 'tool'
+    for operation in operations:
+        if operation.get('action') in {'add', 'replace', 'move'}:
+            operation['meta'] = {
+                'created_by': source,
+                'chat_id': metadata.get('chat_id'),
+                'message_id': metadata.get('message_id'),
+                'model': metadata.get('model'),
+            }
+
+    try:
+        results = await Memories.apply_memory_operations(user.id, operations)
+    except ValueError as e:
+        raise HTTPException(status_code=404, detail=str(e))
+
+    upsert_items = []
+    delete_ids = []
+    response = []
+
+    for result in results:
+        memory = result.get('memory')
+        if isinstance(memory, MemoryModel):
+            result = {**result, 'memory': memory.model_dump()}
+            if result.get('status') in {'created', 'updated'}:
+                vector = await request.app.state.EMBEDDING_FUNCTION(
+                    memory_vector_text(memory.content, memory.path),
+                    prefix=RAG_EMBEDDING_CONTENT_PREFIX,
+                    user=user,
+                )
+                upsert_items.append(
+                    {
+                        'id': memory.id,
+                        'text': memory_vector_text(memory.content, memory.path),
+                        'vector': vector,
+                        'metadata': _memory_metadata(memory),
+                    }
+                )
+        if result.get('status') == 'deleted' and result.get('id'):
+            delete_ids.append(result['id'])
+        response.append(result)
+
+    if upsert_items:
+        await ASYNC_VECTOR_DB_CLIENT.upsert(collection_name=f'user-memory-{user.id}', items=upsert_items)
+
+    if delete_ids:
+        await ASYNC_VECTOR_DB_CLIENT.delete(collection_name=f'user-memory-{user.id}', ids=delete_ids)
+
+    for result in response:
+        status_value = result.get('status')
+        memory = result.get('memory') or {}
+        memory_id = memory.get('id') or result.get('id')
+
+        if status_value == 'created':
+            event = EVENTS.MEMORY_CREATED
+        elif status_value == 'updated':
+            event = EVENTS.MEMORY_UPDATED
+        elif status_value == 'deleted':
+            event = EVENTS.MEMORY_DELETED
+        else:
+            continue
+
+        await publish_event(
+            request,
+            event,
+            actor=user,
+            subject_id=memory_id,
+            data={
+                'content_preview': (memory.get('content') or '')[:300],
+                'type': memory.get('type'),
+                'path': memory.get('path'),
+                'operation': result.get('action'),
+            },
+        )
+
+    return response
+
+
 ############################
 # QueryMemory
 ############################
@@ -124,23 +281,13 @@ async def query_memory(
     # Database operations (get_memories_by_user_id) manage their own short-lived sessions.
     # This prevents holding a connection during EMBEDDING_FUNCTION()
     # which makes external embedding API calls (1-5+ seconds).
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    await check_memories_permission(user)
 
     memories = await Memories.get_memories_by_user_id(user.id)
     if not memories:
         raise HTTPException(status_code=404, detail='No memories found for user')
 
-    vector = await request.app.state.EMBEDDING_FUNCTION(form_data.content, RAG_EMBEDDING_QUERY_PREFIX, user=user)
+    vector = await request.app.state.EMBEDDING_FUNCTION(form_data.content, prefix=RAG_EMBEDDING_QUERY_PREFIX, user=user)
 
     results = await ASYNC_VECTOR_DB_CLIENT.search(
         collection_name=f'user-memory-{user.id}',
@@ -154,7 +301,7 @@ async def query_memory(
     # same RELEVANCE_THRESHOLD used by RAG ensures only genuinely matching
     # memories are surfaced (distances are normalised to 0→1, higher is
     # better).
-    relevance_threshold = getattr(request.app.state.config, 'RELEVANCE_THRESHOLD', 0.0)
+    relevance_threshold = await Config.get('rag.relevance_threshold', 0.0)
     if results and relevance_threshold > 0.0 and results.distances and results.distances[0]:
         from open_webui.retrieval.vector.main import SearchResult
 
@@ -183,6 +330,61 @@ async def query_memory(
     return results
 
 
+@router.post('/search', response_model=list[MemoryModel])
+async def search_memories(
+    form_data: SearchMemoriesForm,
+    user=Depends(get_verified_user),
+):
+    await check_memories_permission(user)
+
+    memories = await Memories.get_memories_by_user_id(user.id)
+    return search_memory_rows(
+        memories,
+        query=form_data.query,
+        path=form_data.path,
+        memory_id=form_data.memory_id,
+        memory_type=form_data.type,
+        limit=form_data.limit,
+    )
+
+
+@router.post('/paths')
+async def list_memory_paths(
+    form_data: ListMemoryPathsForm,
+    user=Depends(get_verified_user),
+):
+    await check_memories_permission(user)
+
+    memories = await Memories.get_memories_by_user_id(user.id)
+    return list_memory_path_groups(
+        memories,
+        query=form_data.query or '',
+        memory_type=form_data.type,
+        limit=form_data.limit,
+    )
+
+
+@router.post('/path')
+async def read_memory_path(
+    form_data: ReadMemoryPathForm,
+    user=Depends(get_verified_user),
+):
+    await check_memories_permission(user)
+
+    memories = await Memories.get_memories_by_user_id(user.id)
+    result = read_memory_path_rows(
+        memories,
+        path=form_data.path,
+        memory_type=form_data.type,
+        include_children=form_data.include_children,
+        limit=form_data.limit,
+    )
+    return {
+        **result,
+        'memories': [memory.model_dump() for memory in result['memories']],
+    }
+
+
 ############################
 # ResetMemoryFromVectorDB
 ############################
@@ -199,17 +401,7 @@ async def reset_memory_from_vector_db(
     calls simultaneously. With a session held, this could block a connection
     for MINUTES, completely exhausting the connection pool.
     """
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    await check_memories_permission(user)
 
     await ASYNC_VECTOR_DB_CLIENT.delete_collection(f'user-memory-{user.id}')
 
@@ -217,7 +409,12 @@ async def reset_memory_from_vector_db(
 
     # Generate vectors in parallel
     vectors = await asyncio.gather(
-        *[request.app.state.EMBEDDING_FUNCTION(memory.content, user=user) for memory in memories]
+        *[
+            request.app.state.EMBEDDING_FUNCTION(
+                memory_vector_text(memory.content, memory.path), prefix=RAG_EMBEDDING_CONTENT_PREFIX, user=user
+            )
+            for memory in memories
+        ]
     )
 
     await ASYNC_VECTOR_DB_CLIENT.upsert(
@@ -225,17 +422,22 @@ async def reset_memory_from_vector_db(
         items=[
             {
                 'id': memory.id,
-                'text': memory.content,
+                'text': memory_vector_text(memory.content, memory.path),
                 'vector': vectors[idx],
-                'metadata': {
-                    'created_at': memory.created_at,
-                    'updated_at': memory.updated_at,
-                },
+                'metadata': _memory_metadata(memory),
             }
             for idx, memory in enumerate(memories)
         ],
     )
 
+    await publish_event(
+        request,
+        EVENTS.MEMORY_RESET,
+        actor=user,
+        subject_id=user.id,
+        subject_type='user',
+        data={'count': len(memories)},
+    )
     return True
 
 
@@ -250,17 +452,7 @@ async def delete_memory_by_user_id(
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    await check_memories_permission(user)
 
     result = await Memories.delete_memories_by_user_id(user.id, db=db)
 
@@ -269,6 +461,13 @@ async def delete_memory_by_user_id(
             await ASYNC_VECTOR_DB_CLIENT.delete_collection(f'user-memory-{user.id}')
         except Exception as e:
             log.error(e)
+        await publish_event(
+            request,
+            EVENTS.MEMORY_DELETED,
+            actor=user,
+            subject_id=user.id,
+            subject_type='user',
+        )
         return True
 
     return False
@@ -290,40 +489,48 @@ async def update_memory_by_id(
     # Database operations (update_memory_by_id_and_user_id) manage their own
     # short-lived sessions. This prevents holding a connection during
     # EMBEDDING_FUNCTION() which makes external API calls (1-5+ seconds).
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
+    await check_memories_permission(user)
 
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-    memory = await Memories.update_memory_by_id_and_user_id(memory_id, user.id, form_data.content)
+    content = clean_memory_content(form_data.content) if form_data.content is not None else None
+    path = clean_memory_path(form_data.path)
+    if content is None and form_data.type is None and form_data.path is None:
+        raise HTTPException(status_code=400, detail='No memory update provided')
+    memory = await Memories.update_memory_by_id_and_user_id(
+        memory_id,
+        user.id,
+        content,
+        memory_type=form_data.type,
+        path=path,
+        update_path=form_data.path is not None,
+        meta={'created_by': 'manual'},
+    )
     if memory is None:
         raise HTTPException(status_code=404, detail=ERROR_MESSAGES.NOT_FOUND)
 
-    if form_data.content is not None:
-        vector = await request.app.state.EMBEDDING_FUNCTION(memory.content, user=user)
+    if form_data.content is not None or form_data.path is not None:
+        vector = await request.app.state.EMBEDDING_FUNCTION(
+            memory_vector_text(memory.content, memory.path), prefix=RAG_EMBEDDING_CONTENT_PREFIX, user=user
+        )
 
         await ASYNC_VECTOR_DB_CLIENT.upsert(
             collection_name=f'user-memory-{user.id}',
             items=[
                 {
                     'id': memory.id,
-                    'text': memory.content,
+                    'text': memory_vector_text(memory.content, memory.path),
                     'vector': vector,
-                    'metadata': {
-                        'created_at': memory.created_at,
-                        'updated_at': memory.updated_at,
-                    },
+                    'metadata': _memory_metadata(memory),
                 }
             ],
         )
 
+    await publish_event(
+        request,
+        EVENTS.MEMORY_UPDATED,
+        actor=user,
+        subject_id=memory.id,
+        data={'content_preview': memory.content[:300], 'type': memory.type, 'path': memory.path},
+    )
     return memory
 
 
@@ -339,22 +546,18 @@ async def delete_memory_by_id(
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    if not request.app.state.config.ENABLE_MEMORIES:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-    if not await has_permission(user.id, 'features.memories', request.app.state.config.USER_PERMISSIONS):
-        raise HTTPException(
-            status_code=status.HTTP_403_FORBIDDEN,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
+    await check_memories_permission(user)
 
     result = await Memories.delete_memory_by_id_and_user_id(memory_id, user.id, db=db)
 
     if result:
         await ASYNC_VECTOR_DB_CLIENT.delete(collection_name=f'user-memory-{user.id}', ids=[memory_id])
+        await publish_event(
+            request,
+            EVENTS.MEMORY_DELETED,
+            actor=user,
+            subject_id=memory_id,
+        )
         return True
 
     return False

backend/open_webui/routers/models.py

@@ -20,9 +20,11 @@ from fastapi import (
 from fastapi.responses import RedirectResponse, StreamingResponse
 from open_webui.config import BYPASS_ADMIN_ACCESS_CONTROL
 from open_webui.constants import ERROR_MESSAGES
+from open_webui.events import EVENTS, publish_event
 from open_webui.env import ENABLE_PROFILE_IMAGE_URL_FORWARDING, PROFILE_IMAGE_ALLOWED_MIME_TYPES
 from open_webui.internal.db import get_async_session
 from open_webui.models.access_grants import AccessGrants
+from open_webui.models.config import Config
 from open_webui.models.groups import Groups
 from open_webui.models.models import (
     ModelAccessListResponse,
@@ -38,6 +40,7 @@ from open_webui.models.models import (
 from open_webui.utils.access_control import filter_allowed_access_grants, has_permission
 from open_webui.utils.access_control.files import has_access_to_file
 from open_webui.utils.auth import get_admin_user, get_verified_user
+from open_webui.utils.chat_variables import get_chat_variables_schema
 from pydantic import BaseModel
 from sqlalchemy.ext.asyncio import AsyncSession
 
@@ -46,6 +49,14 @@ log = logging.getLogger(__name__)
 router = APIRouter()
 
 
+def add_chat_variables_schema(model_dict: dict) -> dict:
+    system = (model_dict.get('params') or {}).get('system') if isinstance(model_dict.get('params'), dict) else None
+    schema = get_chat_variables_schema(system)
+    if schema:
+        model_dict.setdefault('meta', {})['chat_variables_schema'] = schema
+    return model_dict
+
+
 def _safe_static_redirect_path(url: str) -> str | None:
     """
     If url is a same-origin static asset path, return a normalized path safe for
@@ -60,6 +71,9 @@ def _safe_static_redirect_path(url: str) -> str | None:
         if decoded == path:
             break
         path = decoded
+    # Fail closed: a value still encoded after the cap would be decoded further downstream.
+    if unquote(path) != path:
+        return None
     if '\x00' in path or '\\' in path:
         return None
     if not path.startswith('/'):
@@ -114,7 +128,11 @@ async def _verify_knowledge_file_access(
 PAGE_ITEM_COUNT = 30
 
 
-@router.get('/list', response_model=ModelAccessListResponse)  # do NOT use "/" as path, conflicts with main.py
+@router.get(
+    '/list',
+    response_model=ModelAccessListResponse,
+    response_model_exclude={'items': {'__all__': {'meta': {'profile_image_url'}}}},
+)  # do NOT use "/" as path, conflicts with main.py
 async def get_models(
     query: str | None = None,
     view_option: str | None = None,
@@ -168,19 +186,19 @@ async def get_models(
     # Strip profile_image_url from meta — images are served via /model/profile/image.
     items = []
     for model in result.items:
-        data = model.model_dump()
+        data = add_chat_variables_schema(model.model_dump())
         if data.get('meta'):
             data['meta'].pop('profile_image_url', None)
-        items.append(
-            ModelAccessResponse(
-                **data,
-                write_access=(
-                    (user.role == 'admin' and BYPASS_ADMIN_ACCESS_CONTROL)
-                    or user.id == model.user_id
-                    or model.id in writable_model_ids
-                ),
-            )
+        write_access = (
+            (user.role == 'admin' and BYPASS_ADMIN_ACCESS_CONTROL)
+            or user.id == model.user_id
+            or model.id in writable_model_ids
         )
+        # Strip params (system prompt and other curated config) for read-only
+        # callers, mirroring the per-id endpoint.
+        if not write_access:
+            data['params'] = {}
+        items.append(ModelAccessResponse(**data, write_access=write_access))
 
     return ModelAccessListResponse(
         items=items,
@@ -193,9 +211,19 @@ async def get_models(
 ###########################
 
 
+@router.get('/base/tags', response_model=list[str])
+async def get_base_model_tags(user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+    tags = await Models.get_all_tags(user_id=user.id, is_admin=True, is_base_model=True, db=db)
+    return sorted(tags)
+
+
 @router.get('/base', response_model=list[ModelResponse])
-async def get_base_models(user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
-    return await Models.get_base_models(db=db)
+async def get_base_models(
+    tag: str | None = None,
+    user=Depends(get_admin_user),
+    db: AsyncSession = Depends(get_async_session),
+):
+    return await Models.get_base_models(tag=tag, db=db)
 
 
 ###########################
@@ -227,7 +255,7 @@ async def create_new_model(
 ):
     """Create a new workspace model entry."""
     if user.role != 'admin' and not await has_permission(
-        user.id, 'workspace.models', request.app.state.config.USER_PERMISSIONS, db=db
+        user.id, 'workspace.models', await Config.get('user.permissions'), db=db
     ):
         raise HTTPException(
             status_code=status.HTTP_401_UNAUTHORIZED,
@@ -255,7 +283,7 @@ async def create_new_model(
         )
 
         form_data.access_grants = await filter_allowed_access_grants(
-            request.app.state.config.USER_PERMISSIONS,
+            await Config.get('user.permissions'),
             user.id,
             user.role,
             form_data.access_grants,
@@ -264,6 +292,13 @@ async def create_new_model(
 
         model = await Models.insert_new_model(form_data, user.id, db=db)
         if model:
+            await publish_event(
+                request,
+                EVENTS.MODEL_CREATED,
+                actor=user,
+                subject_id=model.id,
+                data={'name': model.name},
+            )
             return model
         else:
             raise HTTPException(
@@ -286,7 +321,7 @@ async def export_models(
     if user.role != 'admin' and not await has_permission(
         user.id,
         'workspace.models_export',
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         db=db,
     ):
         raise HTTPException(
@@ -319,7 +354,7 @@ async def import_models(
     if user.role != 'admin' and not await has_permission(
         user.id,
         'workspace.models_import',
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         db=db,
     ):
         raise HTTPException(
@@ -356,10 +391,12 @@ async def import_models(
             else:
                 writable_model_ids = set(existing_model_ids)
 
+            imported_ids = []
             for model_data in data:
                 model_id = model_data.get('id')
 
                 if model_id and is_valid_model_id(model_id):
+                    imported_ids.append(model_id)
                     # Defense-in-depth: skip models referencing inaccessible files
                     try:
                         await _verify_knowledge_file_access(
@@ -391,7 +428,10 @@ async def import_models(
                             continue
 
                         # Update existing model
-                        model_data['meta'] = model_data.get('meta', {})
+                        model_data['meta'] = {
+                            **existing_model.meta.model_dump(),
+                            **(model_data.get('meta') or {}),
+                        }
                         model_data['params'] = model_data.get('params', {})
 
                         updated_model = ModelForm(**{**existing_model.model_dump(), **model_data})
@@ -400,7 +440,7 @@ async def import_models(
                         # metadata-only imports.
                         if 'access_grants' in model_data:
                             updated_model.access_grants = await filter_allowed_access_grants(
-                                request.app.state.config.USER_PERMISSIONS,
+                                await Config.get('user.permissions'),
                                 user.id,
                                 user.role,
                                 updated_model.access_grants,
@@ -413,13 +453,20 @@ async def import_models(
                         model_data['params'] = model_data.get('params', {})
                         new_model = ModelForm(**model_data)
                         new_model.access_grants = await filter_allowed_access_grants(
-                            request.app.state.config.USER_PERMISSIONS,
+                            await Config.get('user.permissions'),
                             user.id,
                             user.role,
                             new_model.access_grants,
                             'sharing.public_models',
                         )
                         await Models.insert_new_model(user_id=user.id, form_data=new_model, db=db)
+            await publish_event(
+                request,
+                EVENTS.MODEL_IMPORTED,
+                actor=user,
+                subject_type='model',
+                data={'count': len(imported_ids), 'model_ids': imported_ids},
+            )
             return True
         else:
             raise HTTPException(status_code=400, detail='Invalid JSON format')
@@ -444,7 +491,15 @@ async def sync_models(
     user=Depends(get_admin_user),
     db: AsyncSession = Depends(get_async_session),
 ):
-    return await Models.sync_models(user.id, form_data.models, db=db)
+    models = await Models.sync_models(user.id, form_data.models, db=db)
+    await publish_event(
+        request,
+        EVENTS.MODEL_SYNCED,
+        actor=user,
+        subject_type='model',
+        data={'count': len(models), 'model_ids': [model.id for model in models]},
+    )
+    return models
 
 
 ###########################
@@ -481,6 +536,7 @@ async def get_model_by_id(id: str, user=Depends(get_verified_user), db: AsyncSes
             db=db,
         ):
             model_dict = model.model_dump()
+            model_dict = add_chat_variables_schema(model_dict)
             # Strip params (system prompt and other admin-curated config)
             # for read-only callers — matches the params strip already
             # enforced on /api/models in utils/models.py.  Owners, admins
@@ -528,11 +584,7 @@ async def get_model_profile_image(
 
     # Fallback: check arena models stored in config (not in the DB)
     if not profile_image_url:
-        arena_models = getattr(
-            getattr(request.app.state, 'config', None),
-            'EVALUATION_ARENA_MODELS',
-            [],
-        )
+        arena_models = await Config.get('evaluation.arena.models', []) or []
         for arena_model in arena_models:
             if arena_model.get('id') == id:
                 profile_image_url = arena_model.get('meta', {}).get('profile_image_url')
@@ -596,7 +648,9 @@ async def get_model_profile_image(
 
 
 @router.post('/model/toggle', response_model=ModelResponse | None)
-async def toggle_model_by_id(id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)):
+async def toggle_model_by_id(
+    request: Request, id: str, user=Depends(get_verified_user), db: AsyncSession = Depends(get_async_session)
+):
     model = await Models.get_model_by_id(id, db=db)
     if model:
         if (
@@ -613,6 +667,14 @@ async def toggle_model_by_id(id: str, user=Depends(get_verified_user), db: Async
             model = await Models.toggle_model_by_id(id, db=db)
 
             if model:
+                await publish_event(
+                    request,
+                    EVENTS.MODEL_ENABLED if model.is_active else EVENTS.MODEL_DISABLED,
+                    actor=user,
+                    subject_id=model.id,
+                    subject_type='model',
+                    data={'name': model.name},
+                )
                 return model
             else:
                 raise HTTPException(
@@ -673,8 +735,14 @@ async def update_model_by_id(
         db,
     )
 
+    if 'base_model_id' not in form_data.model_fields_set:
+        form_data.base_model_id = model.base_model_id
+
+    if 'profile_image_url' not in form_data.meta.model_fields_set:
+        form_data.meta.profile_image_url = model.meta.profile_image_url
+
     form_data.access_grants = await filter_allowed_access_grants(
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         user.id,
         user.role,
         form_data.access_grants,
@@ -682,6 +750,14 @@ async def update_model_by_id(
     )
 
     model = await Models.update_model_by_id(form_data.id, ModelForm(**form_data.model_dump()), db=db)
+    if model:
+        await publish_event(
+            request,
+            EVENTS.MODEL_UPDATED,
+            actor=user,
+            subject_id=model.id,
+            data={'name': model.name},
+        )
     return model
 
 
@@ -746,7 +822,7 @@ async def update_model_access_by_id(
         )
 
     form_data.access_grants = await filter_allowed_access_grants(
-        request.app.state.config.USER_PERMISSIONS,
+        await Config.get('user.permissions'),
         user.id,
         user.role,
         form_data.access_grants,
@@ -757,7 +833,14 @@ async def update_model_access_by_id(
 
     await Models.update_model_updated_at_by_id(form_data.id, db=db)
 
-    return await Models.get_model_by_id(form_data.id, db=db)
+    model = await Models.get_model_by_id(form_data.id, db=db)
+    await publish_event(
+        request,
+        EVENTS.MODEL_ACCESS_UPDATED,
+        actor=user,
+        subject_id=form_data.id,
+    )
+    return model
 
 
 ############################
@@ -767,6 +850,7 @@ async def update_model_access_by_id(
 
 @router.post('/model/delete', response_model=bool)
 async def delete_model_by_id(
+    request: Request,
     form_data: ModelIdForm,
     user=Depends(get_verified_user),
     db: AsyncSession = Depends(get_async_session),
@@ -795,10 +879,22 @@ async def delete_model_by_id(
         )
 
     result = await Models.delete_model_by_id(form_data.id, db=db)
+    if result:
+        await publish_event(
+            request,
+            EVENTS.MODEL_DELETED,
+            actor=user,
+            subject_id=form_data.id,
+            data={'name': model.name},
+        )
     return result
 
 
 @router.delete('/delete/all', response_model=bool)
-async def delete_all_models(user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)):
+async def delete_all_models(
+    request: Request, user=Depends(get_admin_user), db: AsyncSession = Depends(get_async_session)
+):
     result = await Models.delete_all_models(db=db)
+    if result:
+        await publish_event(request, EVENTS.MODEL_DELETED, actor=user, subject_type='model')
     return result