Merge pull request #21443 from open-webui/dev

0.8.2

.github/ISSUE_TEMPLATE/bug_report.yaml

@@ -11,9 +11,9 @@ body:
 
         ## Important Notes
 
-        - **Before submitting a bug report**: Please check the [Issues](https://github.com/open-webui/open-webui/issues) and [Discussions](https://github.com/open-webui/open-webui/discussions) sections to see if a similar issue has already been reported. If unsure, start a discussion first, as this helps us efficiently focus on improving the project. Duplicates may be closed without notice. **Please search for existing issues and discussions.**
+        - **Before submitting a bug report**: Please check the [Issues](https://github.com/open-webui/open-webui/issues) and [Discussions](https://github.com/open-webui/open-webui/discussions) sections to see if a similar issue has already been reported. If unsure, start a discussion first, as this helps us efficiently focus on improving the project. Duplicates may be closed without notice. **Please search for existing issues AND discussions. No matter open or closed.**
 
-        - Check for opened, **but also for (recently) CLOSED issues** as the issue you are trying to report **might already have been fixed!**
+        - Check for opened, **but also for (recently) CLOSED issues** as the issue you are trying to report **might already have been fixed on the dev branch!**
 
         - **Respectful collaboration**: Open WebUI is a volunteer-driven project with a single maintainer and contributors who also have full-time jobs. Please be constructive and respectful in your communication.
 
@@ -21,6 +21,8 @@ body:
 
         - **Bug Reproducibility**: If a bug cannot be reproduced using a `:main` or `:dev` Docker setup or with `pip install` on Python 3.11, community assistance may be required. In such cases, we will move it to the "[Issues](https://github.com/open-webui/open-webui/discussions/categories/issues)" Discussions section. Your help is appreciated!
 
+        - **Scope**: If you want to report a SECURITY VULNERABILITY, then do so through our [GitHub security page](https://github.com/open-webui/open-webui/security).
+
   - type: checkboxes
     id: issue-check
     attributes:
@@ -31,6 +33,8 @@ body:
           required: true
         - label: I have searched for any existing and/or related discussions.
           required: true
+        - label: I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
+          required: true
         - label: I am using the latest version of Open WebUI.
           required: true
 

.github/ISSUE_TEMPLATE/feature_request.yaml

@@ -8,10 +8,21 @@ body:
       value: |
         ## Important Notes
         ### Before submitting
-        Please check the [Issues](https://github.com/open-webui/open-webui/issues) or [Discussions](https://github.com/open-webui/open-webui/discussions) to see if a similar request has been posted.
+        
+        Please check the **open AND closed** [Issues](https://github.com/open-webui/open-webui/issues) AND [Discussions](https://github.com/open-webui/open-webui/discussions) to see if a similar request has been posted.
         It's likely we're already tracking it! If you’re unsure, start a discussion post first.
-        This will help us efficiently focus on improving the project.
 
+        #### Scope
+
+        If your feature request is likely to take more than a quick coding session to implement, test and verify, then open it in the **Ideas** section of the [Discussions](https://github.com/open-webui/open-webui/discussions) instead.
+        **We will close and force move your feature request to the Ideas section, if we believe your feature request is not trivial/quick to implement.**
+        This is to ensure the issues tab is used only for issues, quickly addressable feature requests and tracking tickets by the maintainers.
+        Other feature requests belong in the **Ideas** section of the [Discussions](https://github.com/open-webui/open-webui/discussions).
+        
+        If your feature request might impact others in the community, definitely open a discussion instead and evaluate whether and how to implement it.
+        
+        This will help us efficiently focus on improving the project.
+        
         ### Collaborate respectfully
         We value a **constructive attitude**, so please be mindful of your communication. If negativity is part of your approach, our capacity to engage may be limited. We're here to help if you're **open to learning** and **communicating positively**.
 
@@ -22,7 +33,6 @@ body:
 
         We appreciate your time and ask that you **respect ours**.
 
-
         ### Contributing
         If you encounter an issue, we highly encourage you to submit a pull request or fork the project. We actively work to prevent contributor burnout to maintain the quality and continuity of Open WebUI.
 
@@ -35,14 +45,22 @@ body:
       label: Check Existing Issues
       description: Please confirm that you've checked for existing similar requests
       options:
-        - label: I have searched the existing issues and discussions.
+        - label: I have searched for all existing **open AND closed** issues and discussions for similar requests. I have found none that is comparable to my request.
+          required: true
+  - type: checkboxes
+    id: feature-scope
+    attributes:
+      label: Verify Feature Scope
+      description: Please confirm the feature's scope is within the described scope
+      options:
+        - label: I have read through and understood the scope definition for feature requests in the Issues section. I believe my feature request meets the definition and belongs in the Issues section instead of the Discussions.
           required: true
   - type: textarea
     id: problem-description
     attributes:
       label: Problem Description
       description: Is your feature request related to a problem? Please provide a clear and concise description of what the problem is.
-      placeholder: "Ex. I'm always frustrated when..."
+      placeholder: "Ex. I'm always frustrated when... / Not related to a problem"
     validations:
       required: true
   - type: textarea

.github/pull_request_template.md

@@ -1,17 +1,27 @@
+<!--
+⚠️ CRITICAL CHECKS FOR CONTRIBUTORS (READ, DON'T DELETE) ⚠️
+1. Target the `dev` branch. PRs targeting `main` will be automatically closed.
+2. Do NOT delete the CLA section at the bottom. It is required for the bot to accept your PR.
+-->
+
 # Pull Request Checklist
 
-### Note to first-time contributors: Please open a discussion post in [Discussions](https://github.com/open-webui/open-webui/discussions) and describe your changes before submitting a pull request.
+### Note to first-time contributors: Please open a discussion post in [Discussions](https://github.com/open-webui/open-webui/discussions) to discuss your idea/fix with the community before creating a pull request, and describe your changes before submitting a pull request.
+
+This is to ensure large feature PRs are discussed with the community first, before starting work on it. If the community does not want this feature or it is not relevant for Open WebUI as a project, it can be identified in the discussion before working on the feature and submitting the PR.
 
 **Before submitting, make sure you've checked the following:**
 
-- [ ] **Target branch:** Verify that the pull request targets the `dev` branch. Not targeting the `dev` branch may lead to immediate closure of the PR.
-- [ ] **Description:** Provide a concise description of the changes made in this pull request.
+- [ ] **Target branch:** Verify that the pull request targets the `dev` branch. **PRs targeting `main` will be immediately closed.**
+- [ ] **Description:** Provide a concise description of the changes made in this pull request down below.
 - [ ] **Changelog:** Ensure a changelog entry following the format of [Keep a Changelog](https://keepachangelog.com/) is added at the bottom of the PR description.
-- [ ] **Documentation:** If necessary, update relevant documentation [Open WebUI Docs](https://github.com/open-webui/docs) like environment variables, the tutorials, or other documentation sources.
-- [ ] **Dependencies:** Are there any new dependencies? Have you updated the dependency versions in the documentation?
-- [ ] **Testing:** Perform manual tests to verify the implemented fix/feature works as intended AND does not break any other functionality. Take this as an opportunity to make screenshots of the feature/fix and include it in the PR description.
-- [ ] **Agentic AI Code:**: Confirm this Pull Request is **not written by any AI Agent** or has at least gone through additional human review **and** manual testing. If any AI Agent is the co-author of this PR, it may lead to immediate closure of the PR.
+- [ ] **Documentation:** Add docs in [Open WebUI Docs Repository](https://github.com/open-webui/docs). Document user-facing behavior, environment variables, public APIs/interfaces, or deployment steps.
+- [ ] **Dependencies:** Are there any new or upgraded dependencies? If so, explain why, update the changelog/docs, and include any compatibility notes. Actually run the code/function that uses updated library to ensure it doesn't crash.
+- [ ] **Testing:** Perform manual tests to **verify the implemented fix/feature works as intended AND does not break any other functionality**. Include reproducible steps to demonstrate the issue before the fix. Test edge cases (URL encoding, HTML entities, types). Take this as an opportunity to **make screenshots of the feature/fix and include them in the PR description**.
+- [ ] **Agentic AI Code:** Confirm this Pull Request is **not written by any AI Agent** or has at least **gone through additional human review AND manual testing**. If any AI Agent is the co-author of this PR, it may lead to immediate closure of the PR.
 - [ ] **Code review:** Have you performed a self-review of your code, addressing any coding standard issues and ensuring adherence to the project's coding standards?
+- [ ] **Design & Architecture:** Prefer smart defaults over adding new settings; use local state for ephemeral UI logic. Open a Discussion for major architectural or UX changes.
+- [ ] **Git Hygiene:** Keep PRs atomic (one logical change). Clean up commits and rebase on `dev` to ensure no unrelated commits (e.g. from `main`) are included. Push updates to the existing PR branch instead of closing and reopening.
 - [ ] **Title Prefix:** To clearly categorize this pull request, prefix the pull request title using one of the following:
   - **BREAKING CHANGE**: Significant changes that may affect compatibility
   - **build**: Changes that affect the build system or external dependencies
@@ -74,4 +84,13 @@
 
 ### Contributor License Agreement
 
+<!--
+🚨 DO NOT DELETE THE TEXT BELOW 🚨
+Keep the "Contributor License Agreement" confirmation text intact.
+Deleting it will trigger the CLA-Bot to INVALIDATE your PR.
+-->
+
 By submitting this pull request, I confirm that I have read and fully agree to the [Contributor License Agreement (CLA)](https://github.com/open-webui/open-webui/blob/main/CONTRIBUTOR_LICENSE_AGREEMENT), and I am providing my contributions under its terms.
+
+> [!NOTE]
+> Deleting the CLA section will lead to immediate closure of your PR and it will not be merged in.

.github/workflows/deploy-to-hf-spaces.yml

@@ -57,7 +57,8 @@ jobs:
           git lfs install
           git lfs track "*.ttf"
           git lfs track "*.jpg"
-          rm demo.gif
+          rm demo.png
+          rm banner.png
           git add .
           git commit -m "GitHub deploy: ${{ github.sha }}"
           git push --force https://open-webui:${HF_TOKEN}@huggingface.co/spaces/open-webui/open-webui main

.github/workflows/docker-build.yaml

@@ -141,6 +141,9 @@ jobs:
           platform=${{ matrix.platform }}
           echo "PLATFORM_PAIR=${platform//\//-}" >> $GITHUB_ENV
 
+      - name: Delete huge unnecessary tools folder
+        run: rm -rf /opt/hostedtoolcache
+
       - name: Checkout repository
         uses: actions/checkout@v5
 
@@ -243,6 +246,9 @@ jobs:
           platform=${{ matrix.platform }}
           echo "PLATFORM_PAIR=${platform//\//-}" >> $GITHUB_ENV
 
+      - name: Delete huge unnecessary tools folder
+        run: rm -rf /opt/hostedtoolcache
+
       - name: Checkout repository
         uses: actions/checkout@v5
 

.prettierignore

@@ -3,8 +3,6 @@ pnpm-lock.yaml
 package-lock.json
 yarn.lock
 
-kubernetes/
-
 # Copy of .gitignore
 .DS_Store
 node_modules

CHANGELOG.md

@@ -5,6 +5,789 @@ 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).
 
+## [0.8.2] - 2026-02-16
+
+### Added
+
+- 🧠 **Skill content handling.** User-selected skills now have their full content injected into the chat, while model-attached skills only display name and description in the available skills list. This allows users to override skill behavior while model-attached skills remain flexible. [Commit](https://github.com/open-webui/open-webui/commit/393c0071dc612c5ac982fb37dfc0288cb9911439)
+- ⚙️ **Chat toggles now control built-in tools.** Users can now disable web search, image generation, and code execution on a per-conversation basis, even when those tools are enabled as builtin tools on the model. [#20641](https://github.com/open-webui/open-webui/issues/20641), [#21318](https://github.com/open-webui/open-webui/discussions/21318), [Commit](https://github.com/open-webui/open-webui/commit/c46ef3b63bcc1e2e9adbdd18fab82c4bbe33ff6c), [Commit](https://github.com/open-webui/open-webui/commit/f1a1e64d2e9ad953b2bc2a9543e9a308b7c669c8)
+- 🖼️ **Image preview in file modal.** Images uploaded to chats can now be previewed directly in the file management modal, making it easier to identify and manage image files. [#21413](https://github.com/open-webui/open-webui/issues/21413), [Commit](https://github.com/open-webui/open-webui/commit/e1b3e7252c1896c04d498547908f0fce111434e1)
+- 🏷️ **Batch tag operations.** Tag creation, deletion, and orphan cleanup for chats now use batch database queries instead of per-tag loops, significantly reducing database round trips when updating, archiving, or deleting chats with multiple tags. [Commit](https://github.com/open-webui/open-webui/commit/c748c3ede)
+- 💨 **Faster group list loading.** Group lists and search results now load with a single database query that joins member counts, replacing the previous pattern of fetching groups first and then counting members in a separate batch query. [Commit](https://github.com/open-webui/open-webui/commit/33308022f)
+- 🔐 **Skills sharing permissions.** Administrators can now control skills sharing and public sharing permissions per-group, matching the existing capabilities for tools, knowledge, and prompts. [Commit](https://github.com/open-webui/open-webui/commit/88401e91c)
+- ⚡ **Long content truncation in preview modals.** Citation and file content modals now truncate markdown-rendered content at 10,000 characters with a "Show all" expansion button, preventing UI jank when previewing very large documents.
+- 🌐 **Translation updates.** Translations for Spanish and German were enhanced and expanded.
+
+### Fixed
+
+- 🔐 **OAuth session error handling.** Corrupted OAuth sessions are now gracefully handled and automatically cleaned up instead of causing errors. [Commit](https://github.com/open-webui/open-webui/commit/7e224e4a536b07ec008613f06592e34050e7067c)
+- 🐛 **Task model selector validation.** The task model selector in admin settings now correctly accepts models based on the new access grants system instead of rejecting all models with an incorrect error. [Commit](https://github.com/open-webui/open-webui/commit/9a2595f0706d0c9d809ae7746001cf799f98db1d)
+- 🔗 **Tool call message preservation.** Models no longer hallucinate tool outputs in multi-turn conversations because tool call history is now properly preserved instead of being merged into assistant messages. [#21098](https://github.com/open-webui/open-webui/discussions/21098), [#20600](https://github.com/open-webui/open-webui/issues/20600), [Commit](https://github.com/open-webui/open-webui/commit/f2aca781c87244cffc130aa2722e700c19a81d66)
+- 🔧 **Tool server startup initialization.** External tool servers configured via the "TOOL_SERVER_CONNECTIONS" environment variable now initialize automatically on startup, eliminating the need to manually visit the Admin Panel and save for tools to become available. This enables proper GitOps and containerized deployments. [#18140](https://github.com/open-webui/open-webui/issues/18140), [#20914](https://github.com/open-webui/open-webui/pull/20914), [Commit](https://github.com/open-webui/open-webui/commit/f20cc6d7e6da493eb75ca1618f5cbd068fa57684)
+- ♻️ **Resource handle cleanup.** File handles are now properly closed during audio transcription and pipeline uploads, preventing resource leaks that could cause system instability over time. [#21411](https://github.com/open-webui/open-webui/issues/21411)
+- ⌨️ **Strikethrough shortcut conflict fix.** Pressing Ctrl+Shift+S to toggle the sidebar no longer causes text to become struck through in the chat input, by disabling the TipTap Strike extension's default keyboard shortcut when rich text mode is off. [Commit](https://github.com/open-webui/open-webui/commit/38ae91ae2)
+- 🔧 **Tool call finish_reason fix.** API responses now correctly set finish_reason to "tool_calls" instead of "stop" when tool calls are present, fixing an issue where external API clients (such as OpenCode) would halt prematurely after tool execution when routing Ollama models through the Open WebUI API. [#20896](https://github.com/open-webui/open-webui/issues/20896)
+
+## [0.8.1] - 2026-02-14
+
+### Added
+
+- 🚀 **Channel user active status.** Checking user active status in channels is now faster thanks to optimized database queries. [Commit](https://github.com/open-webui/open-webui/commit/ca6b18ab5cb94153a9dae233f975d36bf6b19b76)
+- 🔗 **Responses API endpoint with model routing.** The OpenAI API proxy now supports a /responses endpoint that routes requests to the correct backend based on the model field in the request, instead of always using the first configured endpoint. This enables support for backends like vLLM that provide /skills and /v1/responses endpoints. [Commit](https://github.com/open-webui/open-webui/commit/abc9b63093d65f4d74342db85b7d5df1809aa0f0), [Commit](https://github.com/open-webui/open-webui/commit/79ecbfc757f0642740d0e44fab98263d84295490)
+- ⚡ **Model and prompt list optimization.** Improved performance when loading models and prompts by pre-fetching user group IDs once instead of making multiple database queries. [Commit](https://github.com/open-webui/open-webui/commit/20de5a87da0c12e4052b50887a42ddd7228c5ef5)
+- 🗄️ **Batch access control queries.** Improved performance when loading models, prompts, and knowledge bases by replacing multiple individual access checks with single batch queries, significantly reducing database load for large deployments. [Commit](https://github.com/open-webui/open-webui/commit/589c4e64c1b7bb7a7a5abc20382b92fb860e28c2)
+- 💨 **Faster user list loading.** User lists now load significantly faster by deferring profile image loading; images are fetched separately in parallel by the browser, improving caching and reducing database load. [Commit](https://github.com/open-webui/open-webui/commit/b7549d2f6ca2843661ec79a5a1e55da9e7553368)
+- 🔍 **Web search result count.** The built-in search_web tool now respects the admin-configured "Search Result Count" setting instead of always returning 5 results when using Native Function Calling mode. [#21373](https://github.com/open-webui/open-webui/pull/21373), [#21371](https://github.com/open-webui/open-webui/issues/21371)
+- 🔐 **SCIM externalId support.** SCIM-enabled deployments can now store and manage externalId for user provisioning, enabling better integration with identity providers like Microsoft Entra ID and Okta. [#21099](https://github.com/open-webui/open-webui/pull/21099), [#21280](https://github.com/open-webui/open-webui/issues/21280), [Commit](https://github.com/open-webui/open-webui/commit/d1d1efe212b16e0052359991d67fd813125077e8)
+- 🌐 **Translation updates.** Portuguese (Brazil) translations were updated.
+
+### Fixed
+
+- 🛡️ **Public sharing security fix.** Fixed a security issue where users with write access could see the Public sharing option regardless of their actual public sharing permission, and direct API calls could bypass frontend sharing restrictions. [#21358](https://github.com/open-webui/open-webui/pull/21358), [#21356](https://github.com/open-webui/open-webui/issues/21356)
+- 🔒 **Direct model access control fix.** Model access control changes now persist correctly for direct Ollama and OpenAI models that don't have database entries, and error messages display properly instead of showing "[object Object]". [Commit](https://github.com/open-webui/open-webui/commit/f027a01ab2ff3b6175af3dd13a4478c265c0544a), [#21377](https://github.com/open-webui/open-webui/issues/21377)
+- 💭 **Reasoning trace rendering performance.** Reasoning traces from models now render properly without being split into many fragments, preventing browser slowdowns during streaming responses. [#21348](https://github.com/open-webui/open-webui/issues/21348), [Commit](https://github.com/open-webui/open-webui/commit/3b61562c82448cf83710d8b6ed29b797991aa83a)
+- 🖥️ **ARM device compatibility fix.** Fixed an issue where upgrading to 0.8.0 would fail to start on ARM devices (like Raspberry Pi 4) due to torch 2.10.0 causing SIGILL errors; now pinned to torch<=2.9.1. [#21385](https://github.com/open-webui/open-webui/pull/21385), [#21349](https://github.com/open-webui/open-webui/issues/21349)
+- 🗄️ **Skills PostgreSQL compatibility fix.** Fixed a PostgreSQL compatibility issue where creating or listing skills would fail with a TypeError, while SQLite worked correctly. [#21372](https://github.com/open-webui/open-webui/pull/21372), [Commit](https://github.com/open-webui/open-webui/commit/b4c3f54f9648c4232a0fd6557703ffa66fcf4caa), [#21365](https://github.com/open-webui/open-webui/issues/21365)
+- 🗄️ **PostgreSQL analytics query fix.** Fixed an issue where retrieving chat IDs by model ID would fail on PostgreSQL due to incompatible DISTINCT ordering, while SQLite worked correctly. [#21347](https://github.com/open-webui/open-webui/issues/21347), [Commit](https://github.com/open-webui/open-webui/commit/7bda6bf767d5d5c4dc1111465096a88e10b5030e)
+- 🗃️ **SQLite cascade delete fix.** Deleting chats now properly removes all associated messages in SQLite, matching PostgreSQL behavior and preventing orphaned data. [#21362](https://github.com/open-webui/open-webui/pull/21362)
+- ☁️ **Ollama Cloud model naming fix.** Fixed an issue where using Ollama Cloud models would fail with "Model not found" errors because ":latest" was incorrectly appended to model names. [#21386](https://github.com/open-webui/open-webui/issues/21386)
+- 🛠️ **Knowledge selector tooltip z-index.** Fixed an issue where tooltips in the "Select Knowledge" dropdown were hidden behind the menu, making it difficult to read knowledge item names and descriptions. [#21375](https://github.com/open-webui/open-webui/pull/21375)
+- 🎯 **Model selector scroll position.** The model selector dropdown now correctly scrolls to and centers the currently selected model when opened, and resets scroll position when reopened. [Commit](https://github.com/open-webui/open-webui/commit/0b05b2fc7ed4c38af158707438ff404d1beb7c91)
+- 🐛 **Sync modal unexpected appearance.** Fixed an issue where the Sync Modal would appear unexpectedly after enabling the "Community Sharing" feature if the user had previously visited the app with the sync parameter. [#21376](https://github.com/open-webui/open-webui/pull/21376)
+- 🎨 **Knowledge collection layout fix.** Fixed a layout issue in the Knowledge integration menu where long collection names caused indentation artifacts and now properly truncate with ellipsis. [#21374](https://github.com/open-webui/open-webui/pull/21374)
+- 📝 **Metadata processing crash fix.** Fixed a latent bug where processing document metadata containing certain keys (content, pages, tables, paragraphs, sections, figures) would cause a RuntimeError due to dictionary mutation during iteration. [#21105](https://github.com/open-webui/open-webui/pull/21105)
+- 🔑 **Password validation regex fix.** Fixed the password validation regex by adding the raw string prefix, ensuring escape sequences like \d and \w are interpreted correctly. [#21400](https://github.com/open-webui/open-webui/pull/21400), [#21399](https://github.com/open-webui/open-webui/issues/21399)
+
+### 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.
+
+## [0.8.0] - 2026-02-12
+
+### Added
+
+- 📊 **Analytics dashboard.** Administrators now have access to an Analytics dashboard showing model usage statistics, token consumption by model and user, user activity rankings, and time-series charts with hourly or daily granularity; clicking any model opens a detail view with feedback history, associated tags, and chat browser, and results can be filtered by user group. [#21106](https://github.com/open-webui/open-webui/pull/21106), [Commit](https://github.com/open-webui/open-webui/commit/68a1e87b66a7ec8831d5ed52940c4ef110e3e264), [Commit](https://github.com/open-webui/open-webui/commit/e62649f94044abfed4d7d60647a2050383a67e3d)
+- 🎯 **Experimental support for Skills.** Open WebUI now supports the Skill standard — allowing users to create and manage reusable AI skills with detailed instructions, reference them in chats using the "$" command, or attach them to specific models for automatic context in conversations. [#21312](https://github.com/open-webui/open-webui/pull/21312)
+- 🧪 **Experimental support for Open Responses protocol.** Connections can now be configured to use the experimental Open Responses protocol instead of Chat Completions, enabling native support for extended thinking, streaming reasoning tokens, and richer tool call handling for compatible providers. [Commit](https://github.com/open-webui/open-webui/commit/d2c695eb11ddca9fc93499bb0c3fcafcff7099b5), [Commit](https://github.com/open-webui/open-webui/commit/90a057f4005c000bda6ff8703e13e529190af73a), [Commit](https://github.com/open-webui/open-webui/commit/0dc74a8a2e7adb76fb503ef0cd3c02daddd2f4bb), [Commit](https://github.com/open-webui/open-webui/commit/ea9c58ea80646cef05e06d0beaf5e81cc2f78cb1), [Commit](https://github.com/open-webui/open-webui/commit/6ffce4bccdc13b8b61a8b286e34094c981932eda), [Commit](https://github.com/open-webui/open-webui/commit/6719558150920f570d8febe021da65903e53c976), [Commit](https://github.com/open-webui/open-webui/commit/117c091b95a1b1a76a31c31b97304bac289d6f18), [Commit](https://github.com/open-webui/open-webui/commit/aa8c2959ca8476f269786e1317fb6d2938abd3f9), [Commit](https://github.com/open-webui/open-webui/commit/e2d09ac36174de48a7d85bafc8d3291c9ffe44cd)
+- 👥 **Redesigned access control UI.** The access control UI was redesigned with a more intuitive interface that makes it easier to add multiple groups at once. [#21277](https://github.com/open-webui/open-webui/pull/21277)
+- 👤 **Per-user resource sharing.** Resources including knowledge bases, prompts, models, tools, channels, and base models can now be shared directly to individual users alongside the existing per-group sharing capability. [#21277](https://github.com/open-webui/open-webui/pull/21277)
+- 📨 **Message queuing.** Messages can now be queued while a response is generating rather than being blocked, allowing you to continue your train of thought; queued messages are automatically combined and sent when generation completes, and can be edited, deleted, or sent immediately from the input area. [Commit](https://github.com/open-webui/open-webui/commit/62750b8980ef0a3f2da7bc64b5416706a7495686), [Commit](https://github.com/open-webui/open-webui/commit/d3f2cf74748db42311ca04a56ccd1ea15399eca0)
+- 💡 **Active task sidebar indicator.** Users can now see which chats have active tasks running directly in the sidebar. [Commit](https://github.com/open-webui/open-webui/commit/48522271586a5bf24b649610f03b4ffd8afb2782)
+- 📝 **Prompt version control.** Prompts now include version control with full history tracking, allowing users to commit changes with messages, view past versions, compare differences between versions, and roll back to previous versions when needed. [#20945](https://github.com/open-webui/open-webui/pull/20945)
+- 🏷️ **Prompt tags.** Prompts can now be organized with tags, and users can filter the prompt workspace by tag to quickly find related prompts across large collections. [#20945](https://github.com/open-webui/open-webui/pull/20945)
+- 🐍 **Native function calling code execution.** Code execution now works with Native function calling mode, allowing models to autonomously run Python code for calculations, data analysis, and visualizations without requiring Default mode. [#20592](https://github.com/open-webui/open-webui/pull/20592), [Docs:#998](https://github.com/open-webui/docs/pull/998)
+- 🚀 **Async web search.** Web search operations now run asynchronously in the background, allowing users to continue interacting with the application while searches complete. [#20630](https://github.com/open-webui/open-webui/pull/20630)
+- ⚡ **Search debouncing.** Search operations across the application now respond more efficiently with debouncing that reduces unnecessary server requests while typing, improving responsiveness when searching users, groups, functions, tools, prompts, knowledge bases, notes, and when using the knowledge and prompts commands in chat. [#20982](https://github.com/open-webui/open-webui/pull/20982), [Commit](https://github.com/open-webui/open-webui/commit/36766f157d46102fd76c526b42579400ca70de50), [Commit](https://github.com/open-webui/open-webui/commit/fa859de460376782bd0fa35512c8426c9cd0462c), [Commit](https://github.com/open-webui/open-webui/commit/57ec2aa088ffd5a8c3553c53d39799497ff70479)
+- 🤝 **Shared chats management.** Users can now view and manage all their shared chats from Settings, with options to copy share links or unshare conversations they no longer want public. [Commit](https://github.com/open-webui/open-webui/commit/a10ac774ab5d47b505e840b029c0c0340002508b)
+- 📁 **User file management.** Users can now view, search, and delete all their uploaded files from Settings, providing centralized file management in one place. [Commit](https://github.com/open-webui/open-webui/commit/93ed4ae2cda2f4311143e51f586aaa73b83a37a7), [#21047](https://github.com/open-webui/open-webui/pull/21047)
+- 🗑️ **Shift-click quick delete.** Files in the File Manager can now be quickly deleted by holding Shift and clicking the delete button, bypassing the confirmation dialog for faster bulk cleanup. [#21044](https://github.com/open-webui/open-webui/pull/21044)
+- ⌨️ **Model selector shortcut.** The model selector can now be opened with Ctrl+Shift+M keyboard shortcut. [#21130](https://github.com/open-webui/open-webui/pull/21130)
+- 🧠 **Smarter knowledge vs web search.** Models now choose more intelligently between knowledge base search and web search rather than always trying knowledge first. [#21115](https://github.com/open-webui/open-webui/pull/21115)
+- 🌍 **Community model reviews.** Users can now access community reviews for models directly from the model selector menu and are prompted to leave reviews after rating responses, with administrators able to disable this via the "Community Sharing" setting. [Commit](https://github.com/open-webui/open-webui/commit/bc90463ea60c9a66accb1fd242cf1853910ca838)
+- 📄 **Prompts workspace pagination.** The prompts workspace now includes pagination for large prompt collections, loading 30 prompts at a time with search, filtering, and sorting capabilities for improved performance and navigation. [Commit](https://github.com/open-webui/open-webui/commit/36766f157d46102fd76c526b42579400ca70de50)
+- 🎨 **Action function HTML rendering.** Action functions can now render rich HTML content directly in chat as embedded iframes, matching the capabilities that tools already had and eliminating the need for action authors to inject codeblocks. [#21294](https://github.com/open-webui/open-webui/pull/21294), [Commit](https://github.com/open-webui/open-webui/commit/60ada21c152ed642971429fdbe88dcbf478cf83a)
+- 🔒 **Password-masked valve fields.** Tool and function developers can now mark sensitive fields as passwords, which are automatically masked in the settings UI to prevent shoulder surfing and accidental exposure. [#20852](https://github.com/open-webui/open-webui/issues/20852), [Commit](https://github.com/open-webui/open-webui/commit/8c70453b2e3a6958437d951751e84acbbaafd9aa)
+- 📋 **Prompt quick copy.** Prompts in the workspace now include a quick copy button for easily copying prompt content to the clipboard. [Commit](https://github.com/open-webui/open-webui/commit/78f856e2049991441a3469230ae52799cb86954e)
+- 🔔 **Dismissible notification toasts.** Notification toasts for new messages and other events now include a close button that appears on hover, allowing users to dismiss them immediately instead of waiting for auto-dismissal. [#21056](https://github.com/open-webui/open-webui/issues/21056), [Commit](https://github.com/open-webui/open-webui/commit/73bb600034c8532e30726129743a5ffe9002c5fb)
+- 🔔 **Temporary chat notification privacy.** Notifications from temporary chats now only appear on the device where the chat is running, preventing privacy leaks across logged-in sessions. [#21292](https://github.com/open-webui/open-webui/pull/21292)
+- 💡 **Null chat title fallback.** Notifications without chat titles now display "New Chat" instead of showing null. [#21292](https://github.com/open-webui/open-webui/pull/21292)
+- 🖼️ **Concurrent image editing.** Image editing operations with multiple images now complete faster by loading all images concurrently instead of sequentially. [#20911](https://github.com/open-webui/open-webui/pull/20911)
+- 📧 **USER_EMAIL template variable.** Users can now reference their email address in prompts and system messages using the "{{USER_EMAIL}}" template variable. [#20881](https://github.com/open-webui/open-webui/pull/20881)
+- 🔤 **Alphabetical tool ordering.** Tools and Functions in the Chat Controls sidebar now appear in alphabetical order, making it easier to locate specific tools when working with multiple integrations. [#20871](https://github.com/open-webui/open-webui/pull/20871)
+- 👁️ **Model list status filtering.** Administrators can now filter the model list by status (enabled, disabled, visible, hidden) and bulk enable or disable all filtered models at once. [#20553](https://github.com/open-webui/open-webui/issues/20553), [#20774](https://github.com/open-webui/open-webui/issues/20774), [Commit](https://github.com/open-webui/open-webui/commit/96a9696383d450dad2cbb230f3756ebfa258e029)
+- ⚙️ **Per-model built-in tool toggles.** Administrators can now enable or disable individual built-in tools for each model, including time utilities, memory, chat history, notes, knowledge base, and channels. [#20641](https://github.com/open-webui/open-webui/issues/20641), [Commit](https://github.com/open-webui/open-webui/commit/c46ef3b63bcc1e2e9adbdd18fab82c4bbe33ff6c)
+- 📑 **PDF loading modes.** Administrators can now choose between "page" and "single" PDF loading modes, allowing documents to be processed as individual pages or as complete documents for better chunking across page boundaries. [Commit](https://github.com/open-webui/open-webui/commit/ecbdef732bc71a07c21bbb679edb420f26eac181)
+- 📑 **Model Settings pagination.** Administrators can now navigate large model lists more efficiently in Model Settings, with pagination displaying 30 models per page for smoother navigation. [Commit](https://github.com/open-webui/open-webui/commit/2f584c9f88aeb34ece07b10d05794020d1d656b8)
+- 📌 **Pin read-only models.** Users can now pin read-only models from the workspace. [#21308](https://github.com/open-webui/open-webui/issues/21308), [Commit](https://github.com/open-webui/open-webui/commit/97331bf11d41ca54e47f86777fb8dbd73988c631)
+- 🔍 **Yandex search provider.** Administrators can now configure Yandex as a web search provider, expanding search engine options for retrieval-augmented generation. [#20922](https://github.com/open-webui/open-webui/pull/20922)
+- 🔐 **Custom password hints.** Administrators can now provide custom password requirement hints to users via the "PASSWORD_VALIDATION_HINT" environment variable, making it clearer what password criteria must be met during signup or password changes. [#20647](https://github.com/open-webui/open-webui/issues/20647), [#20650](https://github.com/open-webui/open-webui/pull/20650)
+- 🔑 **OAuth token exchange.** Administrators can now enable OAuth token exchange via "ENABLE_OAUTH_TOKEN_EXCHANGE", allowing external applications to authenticate users by exchanging OAuth provider tokens for Open WebUI session tokens. [Commit](https://github.com/open-webui/open-webui/commit/655420fd25ed0ea872954baa485030079c00c10e)
+- 🗄️ **Weaviate custom endpoints.** Administrators can now connect to self-hosted Weaviate deployments with separate HTTP and gRPC endpoints via new environment variables. [#20620](https://github.com/open-webui/open-webui/pull/20620)
+- 🛡️ **MCP custom SSL certificates.** Administrators can now connect to MCP servers with self-signed or custom SSL certificates via the "AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL" environment variable. [#20875](https://github.com/open-webui/open-webui/issues/20875), [Commit](https://github.com/open-webui/open-webui/commit/c7f996d593e4bb48103b91316204fe7e50e25b35)
+- 🗃️ **Redis Sentinel reconnection delay.** Administrators using Redis Sentinel can now configure a reconnection delay via "REDIS_RECONNECT_DELAY" to prevent retry exhaustion during failover elections. [#21021](https://github.com/open-webui/open-webui/pull/21021)
+- 📡 **Custom user info headers.** Administrators can now customize the header names used when forwarding user information to external services, enabling compatibility with services like AWS Bedrock AgentCore that require specific header prefixes. [Commit](https://github.com/open-webui/open-webui/commit/6c0f886cdf4b4249dca29e9340b3b998a7262d61)
+- 🔗 **Forward user info to tool servers.** User identity and chat context can now be forwarded to MCP servers and external tool servers when "ENABLE_FORWARD_USER_INFO_HEADERS" is enabled, allowing tool providers to implement per-user authorization, auditing, and rate limiting. [#21092](https://github.com/open-webui/open-webui/pull/21092), [Commit](https://github.com/open-webui/open-webui/commit/2c37daef86a058e370151ecead17f10078102307)
+- 📬 **External tool event emitters.** External tools (OpenAPI/MCP) can now send tool events back to Open WebUI using the event emitter endpoint, as message ID is now forwarded alongside chat ID when "ENABLE_FORWARD_USER_INFO_HEADERS" is enabled. [#21214](https://github.com/open-webui/open-webui/pull/21214)
+- 📥 **Playground chat export.** Administrators can now export playground chats as JSON or plain text files, allowing them to save their conversations for backup or sharing outside the platform. [Commit](https://github.com/open-webui/open-webui/commit/8e2b0b6fd2ac99c833a110e2bc6aa655f1682669)
+- 🖼️ **Images playground.** Administrators can now test image generation and editing directly in a new Images playground, with support for uploading source images for edits and downloading results. [Commit](https://github.com/open-webui/open-webui/commit/94302de49b27bdf1df86b5c26f2cafb98f964e52)
+- 🛠️ **Dynamic dropdown valve fields.** Tool and function developers can now create dropdown fields with dynamically-generated options that update based on runtime context, such as available models or user permissions. [Commit](https://github.com/open-webui/open-webui/commit/474427c67e953bb9f7d122757a756a639214e0b2)
+- 🏎️ **Faster profile updates.** User profile updates and role changes are now faster by eliminating redundant database queries. [#21011](https://github.com/open-webui/open-webui/pull/21011)
+- 🔑 **Faster authentication.** User authentication is now 34% faster by combining database lookups into a single query. [#21010](https://github.com/open-webui/open-webui/pull/21010)
+- 🔋 **Faster chat completions.** Chat completions and embeddings now respond much faster by checking the model cache before fetching model lists, reducing Time To First Token from several seconds to subsecond for most requests. [#20886](https://github.com/open-webui/open-webui/pull/20886), [#20069](https://github.com/open-webui/open-webui/discussions/20069)
+- 🏎️ **Faster Redis model list loading.** Model list loading is now significantly faster when using Redis with many models, reducing API response latency by caching configuration values locally instead of making repeated Redis lookups on every model iteration. [#21306](https://github.com/open-webui/open-webui/pull/21306)
+- 💨 **Faster knowledge base file batch-add.** Batch-adding files to knowledge bases is now faster with a single database query instead of one query per file. [#21006](https://github.com/open-webui/open-webui/pull/21006)
+- ⚡ **Smoother model selector dropdown.** The model selector dropdown now renders smoothly even with hundreds of models, eliminating the lag and freezing that occurred when opening the dropdown with large model lists. [Commit](https://github.com/open-webui/open-webui/commit/4331029926245b7b74fa8e254610c91400b239b0)
+- 🚗 **Faster model visibility toggling.** Toggling model visibility in the admin panel is now faster with optimized database access. [#21009](https://github.com/open-webui/open-webui/pull/21009)
+- 💾 **Faster model access control checks.** Model access control checks are now faster by batch-fetching model info and group memberships upfront instead of querying for each model. [#21008](https://github.com/open-webui/open-webui/pull/21008)
+- ⚙️ **Faster model list and imports.** Model list loading and model imports are now faster by eliminating redundant database queries. [#21004](https://github.com/open-webui/open-webui/pull/21004)
+- 🏃 **Faster SCIM group member lookups.** SCIM group member lookups are now up to 13x faster by batching user queries instead of fetching each member individually. [#21005](https://github.com/open-webui/open-webui/pull/21005)
+- 💨 **Batched group member counts.** Group member counts are now fetched in a single batch query when loading group lists, eliminating redundant database lookups. [Commit](https://github.com/open-webui/open-webui/commit/96c07f44a8f5e6346b2ea6ac529ff4ec3c47e90a)
+- 💨 **Faster bulk operations.** Bulk feedback deletion and group member removal are now 4-5x faster with optimized batch operations. [#21019](https://github.com/open-webui/open-webui/pull/21019)
+- 🧠 **Faster memory updates.** Memory updates are now up to 39% faster by eliminating redundant database queries. [#21013](https://github.com/open-webui/open-webui/pull/21013)
+- ⚙️ **Faster filter function loading.** Filter function loading is now faster by batching database queries instead of fetching each function individually. [#21018](https://github.com/open-webui/open-webui/pull/21018)
+- 🖼️ **Image model regex configuration.** Administrators can now configure which image generation models support auto-sizing and URL responses via new regex environment variables, improving compatibility with LiteLLM and other proxies that use prefixed model names. [#21126](https://github.com/open-webui/open-webui/pull/21126), [Commit](https://github.com/open-webui/open-webui/commit/ecf3fa2feb28e74ff6c17ca97d94581f316da56a)
+- 🎁 **Easter eggs toggle.** Administrators can now control the visibility of easter egg features via the "ENABLE_EASTER_EGGS" environment variable. [Commit](https://github.com/open-webui/open-webui/commit/907dba4517903e5646e40223a0edca26a7107bc8)
+- 🔌 **Independent access control updates.** API endpoints now support independent access control updates for models, tools, knowledge bases, and notes, enabling finer-grained permission management. [Commit](https://github.com/open-webui/open-webui/commit/0044902c082f8475336cc7d5c57fe3f35ab0555d), [Commit](https://github.com/open-webui/open-webui/commit/c259c878060af1b03b702c943e8813d7b4fc3199), [Commit](https://github.com/open-webui/open-webui/commit/e3a825769063cee486650cc2eb9a032676e630c5)
+- ♿ **Screen reader accessibility.** Screen reader users now hear the password field label only once on the login page, improving form navigation for assistive technology users. [Commit](https://github.com/open-webui/open-webui/commit/1441d0d735c7a1470070b33327e1dd4dc5ca1131)
+- 🔄 **General improvements.** Various improvements were implemented across the application to enhance performance, stability, and security.
+- 🌐 **Translation updates.** Translations for Catalan, Finnish, Irish, French, German, Japanese, Latvian, Polish, Portuguese (Brazil), Simplified Chinese, Slovak, Spanish, and Traditional Chinese were enhanced and expanded.
+
+### Fixed
+
+- ⚡ **Connection pool exhaustion fix.** Database connection pool exhaustion and timeout errors during concurrent usage have been resolved by releasing connections before chat completion requests and embedding operations for memory and knowledge base processing. [#20569](https://github.com/open-webui/open-webui/pull/20569), [#20570](https://github.com/open-webui/open-webui/pull/20570), [#20571](https://github.com/open-webui/open-webui/pull/20571), [#20572](https://github.com/open-webui/open-webui/pull/20572), [#20573](https://github.com/open-webui/open-webui/pull/20573), [#20574](https://github.com/open-webui/open-webui/pull/20574), [#20575](https://github.com/open-webui/open-webui/pull/20575), [#20576](https://github.com/open-webui/open-webui/pull/20576), [#20577](https://github.com/open-webui/open-webui/pull/20577), [#20578](https://github.com/open-webui/open-webui/pull/20578), [#20579](https://github.com/open-webui/open-webui/pull/20579), [#20580](https://github.com/open-webui/open-webui/pull/20580), [#20581](https://github.com/open-webui/open-webui/pull/20581), [Commit](https://github.com/open-webui/open-webui/commit/7da37b4f66b9b2e821796b06b75e03cb0237e0a9), [Commit](https://github.com/open-webui/open-webui/commit/9af40624c5f0f8f7f640a11356e167543b07b2bb)
+- 🚫 **LDAP authentication hang fix.** LDAP authentication no longer freezes the entire service when logging in with non-existent accounts, preventing application hangs. [Commit](https://github.com/open-webui/open-webui/commit/a4281f6a7fbc9764b57830e4ef81bb780aa34af9), [#21300](https://github.com/open-webui/open-webui/issues/21300)
+- ✅ **Trusted Header auto-registration fix.** Trusted Header Authentication now properly auto-registers new users after the first login, assigning the configured default role instead of failing for users not yet in the database. [Commit](https://github.com/open-webui/open-webui/commit/9b30e8f6894c8c8bad0a9ce4693eab810962adc9)
+- 🛡️ **SSRF protection for image loading.** External image loading now validates URLs before fetching to prevent SSRF attacks against local and private network addresses. [Commit](https://github.com/open-webui/open-webui/commit/ce50d9bac4f30b054b09a2fbda52569b73ea591c)
+- 🛡️ **Malformed Authorization header fix.** Malformed Authorization headers no longer cause server crashes; requests are now handled gracefully instead of returning HTTP 500 errors. [#20938](https://github.com/open-webui/open-webui/issues/20938), [Commit](https://github.com/open-webui/open-webui/commit/7e79f8d1c6b5a02f1a46e792540c6bbf7bed8edc)
+- 🚪 **Channel notification access control.** Users without channel permissions can no longer access channels through notifications, properly enforcing access controls across all channel entry points. [#20883](https://github.com/open-webui/open-webui/pull/20883), [#20789](https://github.com/open-webui/open-webui/discussions/20789)
+- 🐛 **Ollama model name suffix fix.** Ollama-compatible providers that do not use ":latest" in model names can now successfully chat, fixing errors where model names were incorrectly appended with ":latest" suffixes. [#21331](https://github.com/open-webui/open-webui/issues/21331), [Commit](https://github.com/open-webui/open-webui/commit/05ae44b98dc279ee12cc8eab17278ccbfec60301)
+- ♻️ **Streaming connection cleanup.** Streaming responses now properly clean up network connections when interrupted, preventing "Unclosed client session" errors from accumulating over time. [#20889](https://github.com/open-webui/open-webui/pull/20889), [#17058](https://github.com/open-webui/open-webui/issues/17058)
+- 💾 **Inline image context exhaustion fix.** Inline images no longer exhaust the model's context window by including their full base64 data in chat metadata, preventing premature context exhaustion with image-heavy conversations. [#20916](https://github.com/open-webui/open-webui/pull/20916)
+- 🚀 **Status indicator GPU usage fix.** High GPU usage caused by the user online status indicator animation has been resolved, reducing consumption from 35-40% to near-zero in browsers with hardware acceleration. [#21062](https://github.com/open-webui/open-webui/issues/21062), [Commit](https://github.com/open-webui/open-webui/commit/938d1b0743c64f0ce513d68e57dfbb86987cb06b)
+- 🔧 **Async pipeline operations.** Pipeline operations now run asynchronously instead of blocking the FastAPI event loop, allowing the server to handle other requests while waiting for external pipeline API calls. [#20910](https://github.com/open-webui/open-webui/pull/20910)
+- 🔌 **MCP tools regression fix.** MCP tools now work reliably again after a regression in v0.7.2 that caused "cannot pickle '\_asyncio.Future' object" errors when attempting to use MCP servers in chat. [#20629](https://github.com/open-webui/open-webui/issues/20629), [#20500](https://github.com/open-webui/open-webui/issues/20500), [Commit](https://github.com/open-webui/open-webui/commit/886c12c5664bc2dd73313330f61c2257169da6d1)
+- 🔗 **Function chat ID propagation fix.** Functions now reliably receive the chat identifier during internal task invocations like web search query generation, RAG query generation, and image prompt generation, enabling stateful functions to maintain consistent per-chat state without fragmentation. [#20563](https://github.com/open-webui/open-webui/issues/20563), [#20585](https://github.com/open-webui/open-webui/pull/20585)
+- 💻 **Markdown fence code execution fix.** Code execution now works reliably when models wrap code in markdown fences, automatically stripping the backticks before execution to prevent syntax errors that affected most non-GPT models. [#20941](https://github.com/open-webui/open-webui/issues/20941), [Commit](https://github.com/open-webui/open-webui/commit/4a5516775927aaf002212f2e09c55a17c699bc46), [Commit](https://github.com/open-webui/open-webui/commit/683438b418fb3b453a8ad88c1ba1a9944eac3593)
+- 💻 **ANSI code execution fix.** Code execution is now reliable when LLMs include ANSI terminal color codes in their output, preventing random failures that previously caused syntax errors. [#21091](https://github.com/open-webui/open-webui/issues/21091), [Commit](https://github.com/open-webui/open-webui/commit/b1737040a7d3bb5efcfe0f1432e89d7e82e51d2d)
+- 🗨️ **Incomplete model metadata crash fix.** Starting chats with models that have incomplete metadata information no longer crashes the application. [#20565](https://github.com/open-webui/open-webui/issues/20565), [Commit](https://github.com/open-webui/open-webui/commit/14f6747dfc66fb7e942b930650286012121e5262)
+- 💬 **Unavailable model crash fix.** Adding message pairs with Ctrl+Shift+Enter no longer crashes when the chat's model is unavailable, showing a helpful error message instead. [#20663](https://github.com/open-webui/open-webui/pull/20663)
+- 📚 **Knowledge base file upload fix.** Uploading files to knowledge bases now works correctly, fixing database mapping errors that prevented file uploads. [#20925](https://github.com/open-webui/open-webui/issues/20925), [#20931](https://github.com/open-webui/open-webui/pull/20931)
+- 🧠 **Knowledge base query type fix.** Knowledge base queries no longer fail intermittently when models send tool call parameters as strings instead of their expected types. [#20705](https://github.com/open-webui/open-webui/pull/20705)
+- 📚 **Knowledge base reindex fix.** Reindexing knowledge base files now works correctly instead of failing with duplicate content errors. [#20854](https://github.com/open-webui/open-webui/issues/20854), [#20857](https://github.com/open-webui/open-webui/pull/20857)
+- 🔧 **Multi-worker knowledge base timeout fix.** In multi-worker deployments, uploading very large documents to knowledge bases no longer causes workers to be killed by health check timeouts, and administrators can now configure a custom embedding timeout via "RAG_EMBEDDING_TIMEOUT". [#21158](https://github.com/open-webui/open-webui/pull/21158), [Discussion](https://github.com/open-webui/open-webui/discussions/21151), [Commit](https://github.com/open-webui/open-webui/commit/c653e4ec54d070aee5e9568d016daebb61f06632)
+- 🌅 **Dark mode icon inversion fix.** Icons in chat and action menus are now displayed correctly in dark mode, fixing an issue where PNG icons with "svg" in their base64 encoding were randomly inverted. [#21272](https://github.com/open-webui/open-webui/pull/21272), [Commit](https://github.com/open-webui/open-webui/commit/0a44d80252afae73de4098ab1c3eb6cf54157fd6)
+- 🛠️ **Admin model write permission fix.** Fixed the admin panel allowing models to be assigned write permissions, since users with write permission are not admins and cannot write. [Commit](https://github.com/open-webui/open-webui/commit/4aedfdc5471a1f13c1084b34b48ea3ed6311cd42)
+- 🛠️ **Prompt access control save fix.** Prompt access control settings are now saved correctly when modifying resource permissions. [Commit](https://github.com/open-webui/open-webui/commit/30f72672fac2579c267a076e6ba89dfe1812137b)
+- ✏️ **Knowledge base file edit fix.** Editing files within knowledge bases now saves correctly and can be used for retrieval, fixing a silent failure where the save appeared successful but the file could not be searched. [Commit](https://github.com/open-webui/open-webui/commit/f9ab66f51a52388a4eb084c8f69044e79bf5cb04)
+- 🖼️ **Reasoning section artifact rendering fix.** Code blocks within model reasoning sections no longer incorrectly render as interactive artifacts, ensuring only intended output displays as previews. [#20801](https://github.com/open-webui/open-webui/issues/20801), [#20877](https://github.com/open-webui/open-webui/pull/20877), [Commit](https://github.com/open-webui/open-webui/commit/4c6f100b5fe2145a3d676b70b5f7c0e7f07cee20)
+- 🔐 **Group resource sharing fix.** Sharing resources with groups now works correctly, fixing database errors and an issue where models shared with read-only access were not visible to group members. [#20666](https://github.com/open-webui/open-webui/issues/20666), [#21043](https://github.com/open-webui/open-webui/issues/21043), [Commit](https://github.com/open-webui/open-webui/commit/5a075a2c836e46b83f8710285f09aff1f6125072)
+- 🔑 **Docling API key fix.** Docling API key authentication now works correctly by using the proper "X-Api-Key" header format instead of the incorrect "Bearer" authorization prefix. [#20652](https://github.com/open-webui/open-webui/pull/20652)
+- 🔌 **MCP OAuth 2.1 fix.** MCP OAuth 2.1 authentication now works correctly, resolving connection verification failures and 401 errors during the authorization callback. [#20808](https://github.com/open-webui/open-webui/issues/20808), [#20828](https://github.com/open-webui/open-webui/issues/20828), [Commit](https://github.com/open-webui/open-webui/commit/8eebc2aea63b7045e61c9689a65a2dfa9c797bcb)
+- 💻 **MATLAB syntax highlighting.** MATLAB code blocks now display with proper syntax highlighting in chat messages. [#20719](https://github.com/open-webui/open-webui/issues/20719), [#20773](https://github.com/open-webui/open-webui/pull/20773)
+- 📊 **CSV export HTML entity decoding.** Exporting tables to CSV now properly decodes HTML entities, ensuring special characters display correctly in the exported file. [#20688](https://github.com/open-webui/open-webui/pull/20688)
+- 📄 **Markdown Header Text Splitter persistence.** The "Markdown Header Text Splitter" document setting now persists correctly when disabled, preventing it from reverting to enabled after page refresh. [#20929](https://github.com/open-webui/open-webui/issues/20929), [#20930](https://github.com/open-webui/open-webui/pull/20930)
+- 🔌 **Audio service timeout handling.** Audio transcription and text-to-speech requests now have proper timeouts, preventing the UI from freezing when external services don't respond. [#21055](https://github.com/open-webui/open-webui/pull/21055)
+- 💬 **Reference Chats visibility fix.** The "Reference Chats" option now appears in the message input menu even when the sidebar is collapsed, fixing the issue where it was hidden on mobile devices and at first load. [#20827](https://github.com/open-webui/open-webui/issues/20827), [Commit](https://github.com/open-webui/open-webui/commit/a3600e8b219fc4c019b95258d16bd3e2827490c6)
+- 🔍 **Chat search self-exclusion.** The "search_chats" builtin tool now excludes the current conversation from search results, preventing redundant matches. [#20718](https://github.com/open-webui/open-webui/issues/20718), [Commit](https://github.com/open-webui/open-webui/commit/1a4bdd2b30017d901b9cac1e2e10684ec1edd062)
+- 📚 **Knowledge base pagination fix.** Paginating through knowledge base files no longer shows duplicates or skips files when multiple documents share the same update timestamp. [#20846](https://github.com/open-webui/open-webui/issues/20846), [Commit](https://github.com/open-webui/open-webui/commit/a9a0ce6beaa286cc18eff24b518a6f3d7a560e2f)
+- 📋 **Batch file error reporting.** Batch file processing operations now return properly structured error information when failures occur, making it clearer what went wrong during multi-file operations. [#20795](https://github.com/open-webui/open-webui/issues/20795), [Commit](https://github.com/open-webui/open-webui/commit/68b2872ed645cffb641fa5a21a784d6e9ea0d72b)
+- ⚙️ **Persistent config with Redis fix.** Configuration values now respect the "ENABLE_PERSISTENT_CONFIG" setting when Redis is used, ensuring environment variables are reloaded on restart when persistent config is disabled. [#20830](https://github.com/open-webui/open-webui/issues/20830), [Commit](https://github.com/open-webui/open-webui/commit/5d48e48e15b003874cc821d896998a01e87580a0)
+- 🔧 **Engine.IO logging fix.** The "WEBSOCKET_SERVER_ENGINEIO_LOGGING" environment variable now works correctly, allowing administrators to configure Engine.IO logging independently from general websocket logging. [#20727](https://github.com/open-webui/open-webui/pull/20727), [Commit](https://github.com/open-webui/open-webui/commit/5cfb7a08cbde5d39aaf4097b849a80da87c30d66)
+- 🌐 **French language default fix.** Browsers requesting French language now default to French (France) instead of French (Canada), matching standard language preference expectations. [#20603](https://github.com/open-webui/open-webui/pull/20603), [Commit](https://github.com/open-webui/open-webui/commit/4d9a7cc6c0adea54b58046c576250a0c3ae7b512)
+- 🔘 **Firefox delete button fix.** Pressing Enter after clicking delete buttons no longer incorrectly retriggers confirmation modals in Firefox. [Commit](https://github.com/open-webui/open-webui/commit/57a2024c58b9c674f2ae08eeb552994ef1796888)
+- 🌍 **RTL table rendering fix.** Chat markdown tables now correctly display right-to-left when containing RTL language content (Arabic, Hebrew, Farsi, etc.), matching the "Auto" direction setting behavior. [#21160](https://github.com/open-webui/open-webui/issues/21160), [Commit](https://github.com/open-webui/open-webui/commit/284b97bd84c824013ad00ea07621192ec69a5e93)
+- 🔒 **Write permission enforcement for tools.** Users without write permissions are now properly prevented from editing tools, with a clear error message displayed when attempting unauthorized edits. [Commit](https://github.com/open-webui/open-webui/commit/85e92fe3b062ae669985c09495f6ff1baf8176ab), [Commit](https://github.com/open-webui/open-webui/commit/91faa9fd5a1cfc5d3ab531d2d91d28db52bcc702)
+- 🛡️ **Chat Valves permission enforcement.** The "Allow Chat Valves" permission is now properly enforced in the integrations menu, preventing users from bypassing access restrictions. [#20691](https://github.com/open-webui/open-webui/pull/20691)
+- 📝 **Audit log browser session fix.** Audit logs now properly capture all user activity including browser-based sessions, not just API key requests. [#20651](https://github.com/open-webui/open-webui/issues/20651), [Commit](https://github.com/open-webui/open-webui/commit/86e6b2b68b85e958188881785495030de1a30402), [Commit](https://github.com/open-webui/open-webui/commit/ee5fd1246cb3f8f16ca5cbb24feeea43b7800dcb)
+- 🎨 **Long model name truncation.** Long model names and IDs in the admin panel now truncate properly to prevent visual overflow, with full names visible on hover. [#20696](https://github.com/open-webui/open-webui/pull/20696)
+- 👥 **Admin user filter pagination fix.** Filtering users in the admin panel now automatically resets to page 1, preventing empty results when searching from pages beyond the first. [#20723](https://github.com/open-webui/open-webui/pull/20723), [Commit](https://github.com/open-webui/open-webui/commit/be75bc506adb048ef11b1612c0e3662511c920d0)
+- 🔎 **Username search on workspace pages.** Searching for users by username now works correctly on Models, Knowledge, and Functions workspace pages, making it easier to find resources owned by specific users. [#20780](https://github.com/open-webui/open-webui/pull/20780)
+- 🗑️ **File deletion orphaned embeddings fix.** Deleting files now properly removes associated knowledge base embeddings, preventing orphaned data from accumulating. [Commit](https://github.com/open-webui/open-webui/commit/93ed4ae2cda2f4311143e51f586aaa73b83a37a7)
+- 🧹 **Event listener memory leak fix.** Memory leaks caused by event listeners not being cleaned up during navigation have been resolved. [#20913](https://github.com/open-webui/open-webui/pull/20913)
+- 🐳 **Docker Ollama update fix.** Ollama can now be updated within Docker containers after adding a missing zstd dependency. [#20994](https://github.com/open-webui/open-webui/issues/20994), [#21052](https://github.com/open-webui/open-webui/pull/21052)
+- 📝 **Workspace duplicate API request fix.** The prompts, knowledge, and models workspaces no longer make duplicate API requests when loading. [Commit](https://github.com/open-webui/open-webui/commit/ab5dfbda54664c9278b0d807ba06cad94edd798f), [Commit](https://github.com/open-webui/open-webui/commit/e5dbfc420dd3e7f6ba047a3e11584449ff0742b4)
+- 📡 **OpenTelemetry Redis cluster fix.** OpenTelemetry instrumentation now works correctly with Redis cluster mode deployments. [#21129](https://github.com/open-webui/open-webui/pull/21129)
+- 🐳 **Airgapped NLTK tokenizer fix.** Document extraction now works reliably in airgapped environments after container restarts by bundling NLTK tokenizer data in the Docker image. [#21165](https://github.com/open-webui/open-webui/pull/21165), [#21150](https://github.com/open-webui/open-webui/issues/21150)
+- 💬 **Channel model mention crash fix.** Mentioning a model in channels no longer crashes when older thread messages have missing data. [#21112](https://github.com/open-webui/open-webui/pull/21112)
+- 🔧 **OpenAPI tool import fix.** Importing OpenAPI tool specifications no longer crashes when parameters lack explicit name fields, fixing compatibility with complex request body definitions. [#21121](https://github.com/open-webui/open-webui/pull/21121), [Commit](https://github.com/open-webui/open-webui/commit/8e79b3d0bc4903f30e747b663ac818976618c83c)
+- 🌐 **Webpage attachment content fix.** Attaching webpages to chats now retrieves full content instead of only metadata, fixing an unawaited coroutine in SSL certificate verification. [#21166](https://github.com/open-webui/open-webui/issues/21166), [Commit](https://github.com/open-webui/open-webui/commit/a214ec40ea00eebcba49570647ca6ab8f61765d5)
+- 💾 **File upload settings persistence.** File upload settings (Max Upload Size, Max File Count, Image Compression dimensions) now persist correctly and are no longer erased when updating other RAG configuration settings. [#21057](https://github.com/open-webui/open-webui/issues/21057), [Commit](https://github.com/open-webui/open-webui/commit/258454276e1ef8ded24968515f7bf5e1833ca011)
+- 📦 **Tool call expand/collapse fix.** Tool call results in chat can now be expanded and collapsed again after a recent refactor disabled this behavior. [#21205](https://github.com/open-webui/open-webui/pull/21205)
+- 🪛 **Disabled API endpoint bypass fix.** Fixed Ollama/OpenAI API endpoints bypassing 'ENABLE_OLLAMA_API' and 'ENABLE_OPENAI_API' flags when the 'url_idx' parameter was provided. Endpoints now properly return a 503 error with a clear "API is disabled" message instead of attempting to connect and logging confusing connection errors.
+- 🛠️ **OpenSearch 3.0 compatibility fix.** Document uploads to knowledge bases now work correctly when using OpenSearch backend with opensearch-py >= 3.0.0, fixing a TypeError that previously caused failures. [#21248](https://github.com/open-webui/open-webui/pull/21248), [#20649](https://github.com/open-webui/open-webui/issues/20649)
+- 📱 **Gboard multi-line paste fix.** Multi-line text pasted from Gboard on Android now inserts correctly instead of being replaced with a single newline, fixing a bug where the keyboard's clipboard suggestion strip sent text via 'insertText' events instead of standard paste events. [#21265](https://github.com/open-webui/open-webui/pull/21265)
+- 🔧 **Batch embeddings endpoint fix.** The '/api/embeddings' endpoint now correctly returns separate embeddings for each input string when processing batch requests to Ollama providers. [Commit](https://github.com/open-webui/open-webui/commit/8fd5c06e5bf7e0ccbda15d83338912ea17f66783), [#21279](https://github.com/open-webui/open-webui/issues/21279)
+- 🗝️ **SSL verification for embeddings.** SSL certificate verification now respects the "AIOHTTP_CLIENT_SESSION_SSL" setting for OpenAI and Azure OpenAI embedding requests, allowing connections to self-signed certificate endpoints when disabled. [Commit](https://github.com/open-webui/open-webui/commit/cd31b8301b38bfa86872608cfbd022ff74e3ae52)
+- 🔧 **Tool call HTML entity fix.** Models now receive properly formatted tool call results in multi-turn conversations, fixing an issue where HTML entities caused malformed content that was hard to parse. [#20755](https://github.com/open-webui/open-webui/pull/20755)
+- 💾 **Duplicate inline image context fix.** Inline images no longer exhaust the model's context window by including their full base64 data in chat metadata, preventing premature context exhaustion with image-heavy conversations. [#20916](https://github.com/open-webui/open-webui/pull/20916)
+- 🐛 **OpenAI model cache lookup fix.** The OpenAI API router model lookup was corrected to use the proper model identifier when checking the cache, ensuring consistent and correct model retrieval during chat completions. [#21327](https://github.com/open-webui/open-webui/pull/21327)
+- 🐛 **Ollama latest suffix fix.** Ollama-compatible providers that don't use ":latest" in model names can now successfully chat, fixing errors where model names were incorrectly appended with ":latest" suffixes. [#21331](https://github.com/open-webui/open-webui/issues/21331), [Commit](https://github.com/open-webui/open-webui/commit/05ae44b98dc279ee12cc8eab17278ccbfec60301)
+- ⛔ **OpenAI endpoint detection fix.** OpenAI API endpoint detection was corrected to use exact hostname matching instead of substring matching, preventing third-party providers with similar URL patterns from being incorrectly filtered. [Commit](https://github.com/open-webui/open-webui/commit/423d8b18170a0b92b582aba6ef7bb9ba173e876e)
+- 🛠️ **RedisCluster task stopping fix.** Task stopping now works correctly in RedisCluster deployments, fixing an issue where tasks would remain active after cancellation attempts. [#20803](https://github.com/open-webui/open-webui/pull/20803), [Commit](https://github.com/open-webui/open-webui/commit/0dcbd05e2436929ae9d2c559a204844ae0239b57)
+- 📎 **Citation parsing error fix.** Citation parsing no longer crashes when builtin tools return error responses, fixing AttributeError issues when tools like search_web fail. [#21071](https://github.com/open-webui/open-webui/pull/21071)
+
+### Changed
+
+- ‼️ **Database Migration Required** — This release includes database schema changes; multi-worker, multi-server, or load-balanced deployments must update all instances simultaneously rather than performing rolling updates, as running mixed versions will cause application failures due to schema incompatibility between old and new instances.
+- ⚠️ **Chat Message Table Migration** — This release includes a new chat message table migration that can take a significant amount of time to complete in larger deployments with extensive chat histories. Administrators should plan for adequate maintenance windows and allow the migration to complete fully without interruption. Running the migration with insufficient time or resources may result in data integrity issues.
+- 🔗 **Prompt ID-based URLs.** Prompts now use unique ID-based URLs instead of command-based URLs, allowing more flexible command renaming without breaking saved links or integrations. [#20945](https://github.com/open-webui/open-webui/pull/20945)
+
+## [0.7.2] - 2026-01-10
+
+### Fixed
+
+- ⚡ Users no longer experience database connection timeouts under high concurrency due to connections being held during LLM calls, telemetry collection, and file status streaming. [#20545](https://github.com/open-webui/open-webui/pull/20545), [#20542](https://github.com/open-webui/open-webui/pull/20542), [#20547](https://github.com/open-webui/open-webui/pull/20547)
+- 📝 Users can now create and save prompts in the workspace prompts editor without encountering errors. [Commit](https://github.com/open-webui/open-webui/commit/ab99d3b1129cffbc13cf7de5aa897692e3f8662e)
+- 🎙️ Users can now use local Whisper for speech-to-text when STT_ENGINE is left empty (the default for local mode). [#20534](https://github.com/open-webui/open-webui/pull/20534)
+- 📊 The Evaluations page now loads faster by eliminating duplicate API calls to the leaderboard and feedbacks endpoints. [Commit](https://github.com/open-webui/open-webui/commit/2dd09223f2aac301a4d5c17fb667d974c34f3ff1)
+- 🌐 Fixed missing Settings tab i18n label keys. [#20526](https://github.com/open-webui/open-webui/pull/20526)
+
+## [0.7.1] - 2026-01-09
+
+### Fixed
+
+- ⚡ **Improved reliability for low-spec and SQLite deployments.** Fixed page timeouts by disabling database session sharing by default, improving stability for resource-constrained environments. Users can re-enable via 'DATABASE_ENABLE_SESSION_SHARING=true' if needed. [#20520](https://github.com/open-webui/open-webui/issues/20520)
+
+## [0.7.0] - 2026-01-09
+
+### Added
+
+- 🤖 **Native Function Calling with Built-in Tools.** Users can now ask models to perform multi-step tasks that combine web research, knowledge base queries, note-taking, and image generation in a single conversation—for example, "research the latest on X, save key findings to a note, and generate an infographic." Requires models with native function calling support and function calling mode set to "Native" in Chat Controls. [#19397](https://github.com/open-webui/open-webui/issues/19397), [Commit](https://github.com/open-webui/open-webui/commit/5c1d52231a3997a17381c48639bd7e339262cf7c)
+- 🧠 Users can now ask the model to find relevant context from their notes, past chats, and channel messages—for example, "what did I discuss about project X last week?" or "find the conversation where I brainstormed ideas for Y." [Commit](https://github.com/open-webui/open-webui/commit/646835d76744ad9b2e67ede0407a61d62e969aab)
+- 📚 Users can now ask the model to search their knowledge bases and retrieve documents without manually attaching files—for example, "find the section about authentication in our API docs" or "what do our internal guidelines say about X?" [Commit](https://github.com/open-webui/open-webui/commit/c8622adcb01f3091b17ca50f8c8e2f20c7b9cd2a)
+- 💭 Users with models that support interleaved thinking now get more refined results from multi-step workflows, as the model can analyze each tool's output before deciding what to do next.
+- 🔍 When models invoke web search, search results appear as clickable citations in real-time for full source verification. [Commit](https://github.com/open-webui/open-webui/commit/2789f6a24d8405c30cd48ae460071f6a4f2c35f9)
+- 🎚️ Users can selectively disable specific built-in tools (timestamps, memory, chat history, notes, web search, knowledge bases) per model via the model editor's capabilities settings. [Commit](https://github.com/open-webui/open-webui/commit/60e916d6c0c5f7db9e6d670e12be2d1d4abc2dd6)
+- 👁️ Pending tool calls are now displayed during response generation, so users know which tools are being invoked. [Commit](https://github.com/open-webui/open-webui/commit/1d08376860e775049abd1dd5f568ac0c6466c944)
+- 📁 Administrators can now limit the number of files that can be uploaded to folders using the "FOLDER_MAX_FILE_COUNT" setting, preventing resource exhaustion from bulk uploads. [#19810](https://github.com/open-webui/open-webui/issues/19810), [Commit](https://github.com/open-webui/open-webui/commit/a1036e544d573e3d35e05c1c2472ba762c32431b), [Commit](https://github.com/open-webui/open-webui/commit/d3ee3fd23e762c9d83fe1da5636d03259e186e57)
+- ⚡ Users experience transformative speed improvements across the entire application through completely reengineered database connection handling, delivering noticeably faster page loads, butter-smooth interactions, and rock-solid stability during intensive operations like user management and bulk data processing. [Commit](https://github.com/open-webui/open-webui/commit/2041ab483e21b3a757baa25c47dc2fa29018674f), [Commit](https://github.com/open-webui/open-webui/commit/145c7516f227ce56fd52373cee86217fadf16181), [Commit](https://github.com/open-webui/open-webui/commit/475dd91ed798f2efcdf27799d5c5cae3f0e6e847), [Commit](https://github.com/open-webui/open-webui/commit/5d1459df166cce8445eb1556cc26abbd65a3f9f4), [Commit](https://github.com/open-webui/open-webui/commit/2453b75ff0fb2dc75b929d96f417d84e332922d9), [Commit](https://github.com/open-webui/open-webui/commit/5649a668fad15393a52c27a2f188841af8b66989)
+- 🚀 Users experience significantly faster initial page load times through dynamic loading of document processing libraries, reducing the initial bundle size. [#20200](https://github.com/open-webui/open-webui/pull/20200), [#20202](https://github.com/open-webui/open-webui/pull/20202), [#20203](https://github.com/open-webui/open-webui/pull/20203), [#20204](https://github.com/open-webui/open-webui/pull/20204)
+- 💨 Administrators experience dramatically faster user list loading through optimized database queries that eliminate N+1 query patterns, reducing query count from 1+N to just 2 total queries regardless of user count. [#20427](https://github.com/open-webui/open-webui/pull/20427)
+- 📋 Notes now load faster through optimized database queries that batch user lookups instead of fetching each note's author individually. [Commit](https://github.com/open-webui/open-webui/commit/084f0ef6a5491e186bf6b71c6386973ba18ef2fa)
+- 💬 Channel messages, pinned messages, and thread replies now load faster through batched user lookups instead of individual queries per message. [#20458](https://github.com/open-webui/open-webui/pull/20458), [#20459](https://github.com/open-webui/open-webui/pull/20459), [#20460](https://github.com/open-webui/open-webui/pull/20460)
+- 🔗 Users can now click citation content links to jump directly to the relevant portion of source documents with automatic text highlighting, making it easier to verify AI responses against their original sources. [#20116](https://github.com/open-webui/open-webui/pull/20116), [Commit](https://github.com/open-webui/open-webui/commit/40c45ffe1f9b45538d32c8ecba8cac62c6eca503)
+- 📌 Users can now pin or hide models directly from the Workspace Models page and Admin Settings Models page, making it easier to manage which models appear in the sidebar without switching to the chat interface. [#20176](https://github.com/open-webui/open-webui/pull/20176)
+- 🔎 Administrators can now quickly find settings using the new search bar in the Admin Settings sidebar, which supports fuzzy filtering by category names and related keywords like "whisper" for Audio or "rag" for Documents. [#20434](https://github.com/open-webui/open-webui/pull/20434)
+- 🎛️ Users can now view read-only models in the workspace models list, with clear "Read Only" badges indicating when editing is restricted. [#20243](https://github.com/open-webui/open-webui/issues/20243), [#20369](https://github.com/open-webui/open-webui/pull/20369)
+- 📝 Users can now view read-only prompts in the workspace prompts list, with clear "Read Only" badges indicating when editing is restricted. [#20368](https://github.com/open-webui/open-webui/pull/20368)
+- 🔧 Users can now view read-only tools in the workspace tools list, with clear "Read Only" badges indicating when editing is restricted. [#20243](https://github.com/open-webui/open-webui/issues/20243), [#20370](https://github.com/open-webui/open-webui/pull/20370)
+- 📂 Searching for files is now significantly faster, especially for users with large file collections. [Commit](https://github.com/open-webui/open-webui/commit/a9a979fb3db1743553ca0705f571c0b9c252841f)
+- 🏆 The Evaluations leaderboard now calculates Elo ratings on the backend instead of in the browser, improving performance and enabling topic-based model ranking through semantic search. [#15392](https://github.com/open-webui/open-webui/pull/15392), [#20476](https://github.com/open-webui/open-webui/issues/20476), [Commit](https://github.com/open-webui/open-webui/commit/10838b3654bf6fdef02d57311f7f1c01df4cd033)
+- 📊 The Evaluations leaderboard now includes a per-model activity chart displaying daily wins and losses as a diverging bar chart, with 30-day, 1-year, and all-time views using weekly aggregation for longer timeframes.
+- 🎞️ Users can now upload animated GIF and WebP formats as model profile images, with animation preserved by skipping resize processing for these file types. [Commit](https://github.com/open-webui/open-webui/commit/00af37bb4ed1ea0957c7c84a6a8def3a7998b8ca)
+- 📸 Users uploading profile images for users, models, and arena models now benefit from WebP compression at 80% quality instead of JPEG, resulting in significantly smaller file sizes and faster uploads while maintaining visual quality. [Commit](https://github.com/open-webui/open-webui/commit/b1d30673b69571e081abb881d34944cf33cdc67e)
+- ⭐ Action Function developers can now update message favorite status using the new "chat:message:favorite" event, enabling the development of pin/unpin message actions without race conditions from frontend auto-save. [#20375](https://github.com/open-webui/open-webui/pull/20375)
+- 🌐 Users with OpenAI-compatible models that have web search capabilities now see URL citations displayed as sources in the interface. [#20172](https://github.com/open-webui/open-webui/pull/20172), [Commit](https://github.com/open-webui/open-webui/commit/fe84afd09a2bc8a89311f30186ec2608a4edda3a)
+- 📰 Users can now dismiss the "What's New" changelog modal permanently using the X button, matching the behavior of the "Okay, Let's Go!" button. [#20258](https://github.com/open-webui/open-webui/pull/20258)
+- 📧 Administrators can now configure the admin contact email displayed in the Account Pending overlay directly from the Admin Panel instead of only through environment variables. [#12500](https://github.com/open-webui/open-webui/issues/12500), [#20260](https://github.com/open-webui/open-webui/pull/20260)
+- 📄 Administrators can now enable markdown header text splitting as a preprocessing step that works with either character or token splitting, through the new "ENABLE_MARKDOWN_HEADER_TEXT_SPLITTER" setting. [Commit](https://github.com/open-webui/open-webui/commit/d3a682759f375c6cb0bc8c170f62863a070f712e), [Commit](https://github.com/open-webui/open-webui/commit/18a33a079bf07487edffc421a721c86194cc90c9), [Commit](https://github.com/open-webui/open-webui/commit/08bf4670ec862018f9dc57296cb19fd5eab14ef6)
+- 🧩 Administrators can now set a minimum chunk size target using the "CHUNK_MIN_SIZE_TARGET" setting to merge small markdown header chunks with neighbors, which improves retrieval quality by eliminating tiny meaningless fragments, significantly speeds up document processing and embedding, reduces storage costs, and lowers embedding API costs or local compute requirements. [#19595](https://github.com/open-webui/open-webui/issues/19595), [#20314](https://github.com/open-webui/open-webui/pull/20314), [Commit](https://github.com/open-webui/open-webui/commit/c32435958073cf002d87e78544baa88bc4e15d7f)
+- 💨 Administrators can now enable KV prefix caching optimization by setting "RAG_SYSTEM_CONTEXT" to true, which injects RAG context into the system message instead of user messages, enabling models to reuse cached tokens for follow-up questions instead of reprocessing the entire context on each turn, significantly improving response times and reducing costs for cloud-based models. [#20301](https://github.com/open-webui/open-webui/discussions/20301), [#20317](https://github.com/open-webui/open-webui/pull/20317)
+- 🖼️ Administrators and Action developers can now control image generation denoising steps per-request using a steps parameter, allowing Actions and API calls to override the global IMAGE_STEPS configuration for both ComfyUI and Automatic1111 engines. [#20337](https://github.com/open-webui/open-webui/pull/20337)
+- 🗄️ Administrators running multi-pod deployments can now designate a master pod to handle database migrations using the "ENABLE_DB_MIGRATIONS" environment variable. [Commit](https://github.com/open-webui/open-webui/commit/9824f0e33359a917ac07b60bf1f972074d5c8203)
+- 🎙️ Administrators can now configure Whisper's compute type using the "WHISPER_COMPUTE_TYPE" environment variable to fix compatibility issues with CUDA/GPU deployments. [Commit](https://github.com/open-webui/open-webui/commit/26af1f92e21ddfd08348570bf54a6f345ac69648)
+- 🔍 Administrators can now control sigmoid normalization for CrossEncoder reranking models using the "SENTENCE_TRANSFORMERS_CROSS_ENCODER_SIGMOID_ACTIVATION_FUNCTION" environment variable, enabled by default for proper relevance threshold behavior with MS MARCO models. [#20228](https://github.com/open-webui/open-webui/pull/20228)
+- 🔒 Administrators can now disable SSL certificate verification for external tools using the "REQUESTS_VERIFY" environment variable, enabling integration with self-signed certificates for Tika, Ollama embeddings, and external rerankers. [#19968](https://github.com/open-webui/open-webui/issues/19968), [Commit](https://github.com/open-webui/open-webui/commit/dfc5dad63167eabb7fb027e63c324675b23f2e9d)
+- 📈 Administrators can now control audit log output destinations using "ENABLE_AUDIT_STDOUT" and "ENABLE_AUDIT_LOGS_FILE" environment variables, allowing audit logs to be sent to container logs for centralized logging systems. [#20114](https://github.com/open-webui/open-webui/pull/20114), [Commit](https://github.com/open-webui/open-webui/commit/fdae5644e36972384b3e2513e3074f95f9f7381f)
+- 🛡️ Administrators can now restrict non-admin user access to Interface Settings through per-user or per-group permissions. [#20424](https://github.com/open-webui/open-webui/pull/20424)
+- 🧠 Administrators can now globally enable or disable the Memories feature and control access through per-user or per-group permissions, with the Personalization tab automatically hidden when the feature is disabled. [#20462](https://github.com/open-webui/open-webui/pull/20462)
+- 🟢 Administrators can now globally enable or disable user status visibility through the "ENABLE_USER_STATUS" setting in Admin Settings. [#20488](https://github.com/open-webui/open-webui/pull/20488)
+- 🪝 Channel managers can now create webhooks to allow external services to post messages to channels without authentication. [Commit](https://github.com/open-webui/open-webui/commit/cd296fcf0d79cecd1a6a3ee4e492c6b5246ca7ae)
+- 📄 In the model editor users can now disable the "File Context" capability to skip automatic file content extraction and injection, forwarding raw messages with file attachment metadata instead for use with custom tools or future built-in file access tools. [Commit](https://github.com/open-webui/open-webui/commit/daccf0713e3ecd6d24f003a87b5f8b3c61852958), [Docs:Commit](https://github.com/open-webui/docs/commit/18ec6eaefc071a278ec57d4d1b8d66d686af0870)
+- 🔊 In the model editor users can now configure a specific TTS voice for each model, overriding user preferences and global defaults to give different AI personas distinct voices. [#3097](https://github.com/open-webui/open-webui/issues/3097), [Commit](https://github.com/open-webui/open-webui/commit/bb6188abf04302f79d80b0d6cc42c232624b5757)
+- 👥 Administrators now have three granular group sharing permission options instead of a simple on/off toggle, allowing them to choose between "No one", "Members", or "Anyone" for who can share content to each group. [Commit](https://github.com/open-webui/open-webui/commit/ca514cd3eda2524b8da472ef17c0ccb216bac2e8)
+- 📦 Administrators can now export knowledge bases as zip files containing text files for backup and archival purposes. [#20120](https://github.com/open-webui/open-webui/issues/20120), [Commit](https://github.com/open-webui/open-webui/commit/c1147578c073a8c7fa7e7f836149e1cdfec8f18d)
+- 🚀 Administrators can now create an admin account automatically at startup via "WEBUI_ADMIN_EMAIL", "WEBUI_ADMIN_PASSWORD", and "WEBUI_ADMIN_NAME" environment variables, enabling headless and automated deployments without exposing the signup page. [#17654](https://github.com/open-webui/open-webui/issues/17654), [Commit](https://github.com/open-webui/open-webui/commit/1138929f4d083931305f1f925899971b190562ae)
+- 🦆 Administrators can now select a specific search backend for DDGS instead of random selection, with options including Bing, Brave, DuckDuckGo, Google, Wikipedia, Yahoo, and others. [#20330](https://github.com/open-webui/open-webui/issues/20330), [#20366](https://github.com/open-webui/open-webui/pull/20366)
+- 🧭 Administrators can now configure custom Jina Search API endpoints using the "JINA_API_BASE_URL" environment variable, enabling region-specific deployments such as EU data processing. [#19718](https://github.com/open-webui/open-webui/pull/19718), [Commit](https://github.com/open-webui/open-webui/commit/f7f8a263b92289df8d4f8dbc3bae09bd009a5699)
+- 🔥 Administrators can now configure Firecrawl timeout values using the "FIRECRAWL_TIMEOUT" environment variable to control web scraping wait times. [#19973](https://github.com/open-webui/open-webui/pull/19973), [Commit](https://github.com/open-webui/open-webui/commit/89ad1c68d1aadf849960b5e202aa4651096b05f5)
+- 💾 Administrators can now use openGauss as the vector database backend for knowledge base document storage and retrieval. [#20179](https://github.com/open-webui/open-webui/pull/20179)
+- 🔄 Various improvements were implemented across the application to enhance performance, stability, and security.
+- 📊 Users can now sync their anonymous usage statistics to the Open WebUI Community platform to power community leaderboards, drive model evaluations, and contribute to open-source AI research that benefits everyone, all while keeping conversations completely private (only metadata like model names, message counts, and ratings are shared). By sharing your stats, you're helping the community identify which models perform best, contributing to transparent AI benchmarking, and supporting the collective effort to make AI better for all. You can also download your stats as JSON for personal analysis.
+- 🌐 Translations for German, Portuguese (Brazil), Spanish, Simplified Chinese, Traditional Chinese, and Polish were enhanced and expanded.
+
+### Fixed
+
+- 🔊 Text-to-speech now correctly splits on newlines in addition to punctuation, so markdown bullet points and lists are spoken as separate sentences instead of being merged together. [#5924](https://github.com/open-webui/open-webui/issues/5924), [Commit](https://github.com/open-webui/open-webui/commit/869108a3e1ce2b8110084113c1b392072e98fd5f)
+- 🔒 Users are now protected from stored XSS vulnerabilities in iFrame embeds for citations and response messages through configurable same-origin sandbox settings instead of hardcoded values. [#20209](https://github.com/open-webui/open-webui/pull/20209), [#20210](https://github.com/open-webui/open-webui/pull/20210)
+- 🔑 Image Generation, Web Search, and Audio (TTS/STT) API endpoints now enforce permission checks on the backend, closing a security gap where disabled features could previously be accessed via direct API calls. [#20471](https://github.com/open-webui/open-webui/pull/20471)
+- 🛠️ Tools and Tool Servers (MCP and OpenAPI) now enforce access control checks on the backend, ensuring users can only access tools they have permission to use even via direct API calls. [#20443](https://github.com/open-webui/open-webui/issues/20443), [Commit](https://github.com/open-webui/open-webui/commit/9b06fdc8fe1c933071610336be05f11e77e6c8eb)
+- 🔁 System prompts are no longer duplicated when using native function calling, fixing an issue where the prompt would be applied twice during tool-calling workflows. [Commit](https://github.com/open-webui/open-webui/commit/9223efaff0db6e56bfa157ef214d9590005156d2)
+- 🗂️ Knowledge base uploads to folders no longer fail when "FOLDER_MAX_FILE_COUNT" is unset, fixing an issue where the default null value caused all uploads to error. [Commit](https://github.com/open-webui/open-webui/commit/ef9cd0e0ad6e45b8a3efec6f3858b3d69d42f619)
+- 📝 The "Create Note" button in the chat input now correctly hides for users without Notes permissions instead of showing and returning a 401 error when clicked. [#20486](https://github.com/open-webui/open-webui/issues/20486), [Commit](https://github.com/open-webui/open-webui/commit/9e9616b670c1c4389193b18500a7d80d86d7e280)
+- 📊 The Evaluations page no longer crashes when administrators have large amounts of feedback data, as the leaderboard now fetches only the minimal required fields instead of loading entire conversation snapshots. [#20476](https://github.com/open-webui/open-webui/issues/20476), [#20489](https://github.com/open-webui/open-webui/pull/20489), [Commit](https://github.com/open-webui/open-webui/commit/b2a1f71d920e55b143f1c02e61104938d2588762)
+- 💬 Users can now export chats, use the Ask/Explain popup, and view chat lists correctly again after these features were broken by recent refactoring changes that caused 500 and 400 server errors. [#20146](https://github.com/open-webui/open-webui/issues/20146), [#20205](https://github.com/open-webui/open-webui/issues/20205), [#20206](https://github.com/open-webui/open-webui/issues/20206), [#20212](https://github.com/open-webui/open-webui/pull/20212)
+- 💭 Users no longer experience data corruption when switching between chats during background operations like image generation, where messages from one chat would appear in another chat's history. [#20266](https://github.com/open-webui/open-webui/pull/20266)
+- 🛡️ Users no longer encounter critical chat stability errors, including duplicate key errors from circular message dependencies, null message access during chat loading, and errors in the chat overview visualization. [#20268](https://github.com/open-webui/open-webui/pull/20268)
+- 📡 Users with Channels no longer experience infinite recursion and connection pool exhaustion when fetching threaded replies, preventing RecursionError crashes during chat history loading. [#20299](https://github.com/open-webui/open-webui/pull/20299), [Commit](https://github.com/open-webui/open-webui/commit/c144122f608759c2b79472e1f6948a7c1600a3d1)
+- 📎 Users no longer encounter TypeError crashes when viewing messages with file attachments that have undefined URL properties. [#20343](https://github.com/open-webui/open-webui/pull/20343)
+- 🔐 Users with MCP integrations now experience reliable OAuth 2.1 token refresh after access token expiration through proper Protected Resource discovery, preventing integration failures that caused sessions to be deleted. [#19794](https://github.com/open-webui/open-webui/issues/19794), [#20138](https://github.com/open-webui/open-webui/pull/20138), [#20291](https://github.com/open-webui/open-webui/issues/20291), [Commit](https://github.com/open-webui/open-webui/commit/bf2b2962399e341926bdbf9e0a82101f31a90b23), [Commit](https://github.com/open-webui/open-webui/commit/89565c58c6ae6b5b129559ef68b5a0c18c110765)
+- 📚 Users who belong to multiple groups can now see Knowledge Bases shared with those groups, fixing an issue where they would disappear when shared with more than one group. [#20124](https://github.com/open-webui/open-webui/issues/20124), [#20229](https://github.com/open-webui/open-webui/issues/20229), [Commit](https://github.com/open-webui/open-webui/commit/61e25dc2dce9c12dcb5b88a6b814060c4338e67b)
+- 📂 Users now see the correct Knowledge Base name when hovering over # file references in chat input instead of "undefined". [#20329](https://github.com/open-webui/open-webui/issues/20329), [#20333](https://github.com/open-webui/open-webui/pull/20333)
+- 📋 Users now see notes displayed in correct chronological order within their time range groupings, fixing an issue where insertion order was not preserved. [Commit](https://github.com/open-webui/open-webui/commit/3f577c0c3fbfd9f09c02940e4ae474f987149277)
+- 📑 Users collaborating on notes now experience proper content sync when initializing from both HTML and JSON formats, fixing sync failures in collaborative editing sessions. [Commit](https://github.com/open-webui/open-webui/commit/e27fb3e291a735c715a089a80e7a49d2c2209096)
+- 🔎 Users searching notes can now find hyphenated words and variations with spaces, so searching "todo" now finds "to-do" and "to do". [Commit](https://github.com/open-webui/open-webui/commit/a3270648d8b8535443d8ce2ea719f8e678e4e358)
+- 📥 Users no longer experience false duplicate file warnings when reuploading files after initial processing failed, as the file hash is now only stored after successful processing completion. [#19264](https://github.com/open-webui/open-webui/issues/19264), [#20282](https://github.com/open-webui/open-webui/pull/20282), [Commit](https://github.com/open-webui/open-webui/commit/d3ab9f4b96eee7f91c9b1355cee055fdabca9730)
+- 💾 Users experience significantly improved page load performance as model profile images now cache properly in browsers, avoiding unnecessary image refetches. [Commit](https://github.com/open-webui/open-webui/commit/bb821ab654e93908a3b4632c359753eeff053264)
+- 🎨 Users can now successfully edit uploaded images instead of having new images generated, fixing an issue introduced by the file storage refactor where images with type "file" and content_type starting with "image/" weren't being recognized as editable images. [#20237](https://github.com/open-webui/open-webui/issues/20237), [#20169](https://github.com/open-webui/open-webui/pull/20169), [#20239](https://github.com/open-webui/open-webui/pull/20239), [Commit](https://github.com/open-webui/open-webui/commit/1148d1c927d096e14917b6d762789fca3188f281)
+- 🌐 Users writing in Persian and Arabic now see properly displayed right-to-left text in the notes section through automatic text direction detection. [#19743](https://github.com/open-webui/open-webui/issues/19743), [#20102](https://github.com/open-webui/open-webui/pull/20102), [Commit](https://github.com/open-webui/open-webui/commit/b619a157bc54c5bc44d223d2ae3acb9ce4ac6a6c)
+- 🤖 Users can now successfully @ mention models in Channels instead of experiencing silent failures. [Commit](https://github.com/open-webui/open-webui/commit/59957715836acb635f4b1c4ddbfb4ba7b82b3281)
+- 📋 Users on Windows now see correctly preserved line breaks when using the {{CLIPBOARD}} variable through CRLF to LF normalization. [#19370](https://github.com/open-webui/open-webui/issues/19370), [#20283](https://github.com/open-webui/open-webui/pull/20283)
+- 📁 Users now see the Knowledge Selector dropdown correctly displayed above the Create Folder modal instead of being hidden behind it. [#20219](https://github.com/open-webui/open-webui/issues/20219), [#20213](https://github.com/open-webui/open-webui/pull/20213)
+- 🌅 Users now see profile images in non-PNG formats like SVG, JPEG, and GIF displayed correctly instead of appearing broken. [#20171](https://github.com/open-webui/open-webui/pull/20171)
+- 🆕 Non-admin users with disabled temporary chat permissions can now successfully create new chats and use pinned models from the sidebar. [#20336](https://github.com/open-webui/open-webui/issues/20336), [#20367](https://github.com/open-webui/open-webui/pull/20367), [Commit](https://github.com/open-webui/open-webui/commit/e754940c031f9689fb4f6edb3625aa06aeb53377)
+- 🎛️ Users can now successfully use workspace models in chat, fixing "Model not found" errors that occurred when using custom model presets. [#20340](https://github.com/open-webui/open-webui/issues/20340), [#20344](https://github.com/open-webui/open-webui/pull/20344), [Commit](https://github.com/open-webui/open-webui/commit/b55a46ae99c32068ed306a5ecdaafa9f75504cd7), [Commit](https://github.com/open-webui/open-webui/commit/2bb13d5dbc6e233856e8aa26143222ceda8f6c11)
+- 🔁 Users can now regenerate messages without crashes when the parent message is missing or corrupted in the chat history. [#20264](https://github.com/open-webui/open-webui/pull/20264)
+- ✏️ Users no longer experience TipTap rich text editor crashes with "editor view is not available" errors when plugins or async methods try to access the editor after it has been destroyed. [#20266](https://github.com/open-webui/open-webui/pull/20266)
+- 📗 Administrators with bypass access control enabled now correctly have write access to all knowledge bases. [#20371](https://github.com/open-webui/open-webui/pull/20371)
+- 🔍 Administrators using local CrossEncoder reranking models now see proper relevance threshold behavior through MS MARCO model score normalization to the 0-1 range via sigmoid activation. [#19999](https://github.com/open-webui/open-webui/issues/19999), [#20228](https://github.com/open-webui/open-webui/pull/20228)
+- 🎯 Administrators using local SentenceTransformers embedding engine now benefit from proper batch size settings, preventing excessive memory usage from the default batch size of 32. [#20053](https://github.com/open-webui/open-webui/issues/20053), [#20054](https://github.com/open-webui/open-webui/pull/20054), [Commit](https://github.com/open-webui/open-webui/commit/e4a5b06ca68303512678b4d2dc296bc78b9f983f)
+- 🔧 Administrators and users in offline mode or restricted environments like uv, poetry, and NixOS no longer experience crashes when Tools and Functions have frontmatter requirements, as pip installation is now skipped when offline mode is enabled. [#20320](https://github.com/open-webui/open-webui/issues/20320), [#20321](https://github.com/open-webui/open-webui/pull/20321), [Commit](https://github.com/open-webui/open-webui/commit/bd07ef8)
+- 📄 Administrators can now properly configure the MinerU document parsing service as the MinerU Cloud API key field is now available in the Admin Panel Documents settings. [#20319](https://github.com/open-webui/open-webui/issues/20319), [#20328](https://github.com/open-webui/open-webui/pull/20328)
+- ⚠️ Administrators no longer see SyntaxWarnings for invalid escape sequences in password validation regex patterns. [#20298](https://github.com/open-webui/open-webui/pull/20298), [Commit](https://github.com/open-webui/open-webui/commit/e55bf2c2ac391caed871d41f0484820091081908)
+- 🎨 Users with ComfyUI workflows now see only the intended final output images in chat instead of duplicate images from intermediate processing nodes like masks, crops, or segmentation previews. [#20158](https://github.com/open-webui/open-webui/issues/20158), [#20182](https://github.com/open-webui/open-webui/pull/20182)
+- 🖼️ Users with image generation enabled no longer see false vision capability warnings, allowing them to send follow-up messages after generating images and to send images to non-vision models for image editing. [#20129](https://github.com/open-webui/open-webui/issues/20129), [#20256](https://github.com/open-webui/open-webui/pull/20256)
+- 🔌 Administrators no longer experience infinite loading screens when invalid or MCP-style configurations are used with OpenAPI connection types for external tools. [#20207](https://github.com/open-webui/open-webui/issues/20207), [#20257](https://github.com/open-webui/open-webui/pull/20257)
+- 📥 Administrators no longer encounter TypeError crashes during SHA256 verification when uploading GGUF models via URL, fixing 500 Internal Server Error crashes. [#20263](https://github.com/open-webui/open-webui/issues/20263)
+- 🚦 Users with Brave Search now experience automatic retry with a 1-second delay when hitting rate limits, preventing failures when sequential requests exceed the 1 request per second limit, though this only works reliably when web search concurrency is set to a maximum of 1. [#15134](https://github.com/open-webui/open-webui/issues/15134), [#20255](https://github.com/open-webui/open-webui/pull/20255)
+- 🗄️ Administrators with Redis Sentinel deployments no longer experience crashes during websocket disconnections due to improper async-generator handling in the YDocManager. [#20142](https://github.com/open-webui/open-webui/issues/20142), [#20145](https://github.com/open-webui/open-webui/pull/20145)
+- 🔐 Administrators using SCIM group management no longer encounter 500 errors when working with groups that have no members. [#20187](https://github.com/open-webui/open-webui/pull/20187)
+- 🔗 Users now experience more reliable citations from AI models, especially when using smaller or weaker models that may not format citation references perfectly. [Commit](https://github.com/open-webui/open-webui/commit/c0ec04935b4eea3d334bfdec2fc41278f1085a49)
+- 🕸️ Administrators can now successfully save WebSearch settings without encountering validation errors for domain filter lists, YouTube language settings, or timeout values. [#20422](https://github.com/open-webui/open-webui/pull/20422)
+- 📦 Administrators installing with the uv package manager now experience successful installation after deprecated dependencies that were causing conflicts were removed. [#20177](https://github.com/open-webui/open-webui/issues/20177), [#20192](https://github.com/open-webui/open-webui/pull/20192)
+- ⏱️ Administrators using custom "AIOHTTP_CLIENT_TIMEOUT" settings now see the configured timeout correctly applied to embedding generation, OAuth discovery, webhook calls, and tool/function loading instead of falling back to the default 300-second timeout. [Commit](https://github.com/open-webui/open-webui/commit/e67891a374625d9888ec391da561f0b4ed79ed5d)
+
+### Changed
+
+- ⚠️ This release includes a major overhaul of database connection handling in the backend that requires all instances in multi-worker, multi-server, or load-balanced deployments to be updated simultaneously; running mixed versions will cause failures due to incompatible database connection management between old and new instances.
+- 📝 Administrators who previously used the standalone "Markdown (Header)" text splitter must now switch to "character" or "token" mode with the new "ENABLE_MARKDOWN_HEADER_TEXT_SPLITTER" toggle enabled, as document chunking now applies markdown header splitting as a preprocessing step before character or token splitting. [Commit](https://github.com/open-webui/open-webui/commit/d3a682759f375c6cb0bc8c170f62863a070f712e), [Commit](https://github.com/open-webui/open-webui/commit/18a33a079bf07487edffc421a721c86194cc90c9), [Commit](https://github.com/open-webui/open-webui/commit/08bf4670ec862018f9dc57296cb19fd5eab14ef6)
+- 🖼️ Users no longer see the "Generate Image" action button in chat message interfaces; custom function should be used. [Commit](https://github.com/open-webui/open-webui/commit/f0829ba6e6fd200702fb76efc43dd785cf87fec9)
+- 🔗 Administrators will find the Admin Evaluations page at the new URL "/admin/evaluations/feedback" instead of "/admin/evaluations/feedbacks" to use the correct uncountable form of the word. [#20296](https://github.com/open-webui/open-webui/pull/20296)
+- 🔐 Scripts or integrations that directly called Image Generation, Web Search, or Audio APIs while those features were disabled in the Admin UI will now receive 403 Forbidden errors, as backend permission enforcement has been added to match frontend restrictions. [#20471](https://github.com/open-webui/open-webui/pull/20471)
+- 👥 The default group sharing permission changed from "Members" to "Anyone", meaning users can now share content to any group configured with "Anyone" permission regardless of their membership in that group. [Commit](https://github.com/open-webui/open-webui/commit/ca514cd3eda2524b8da472ef17c0ccb216bac2e8)
+
+## [0.6.43] - 2025-12-22
+
+### Fixed
+
+- 🐍 **Python dependency installation issues** were resolved by correcting pip dependency handling, preventing installation failures in certain environments and improving setup reliability. [Commit](https://github.com/open-webui/open-webui/commit/5c5f87a)
+- 🎙️ **Speech-to-Text default content type handling** was fixed and refactored to ensure correct MIME type usage, improving compatibility across STT providers and preventing transcription errors caused by incorrect defaults. [Commit](https://github.com/open-webui/open-webui/commit/4ab917c)
+- 🖼️ **Temporary chat image handling** was fixed and refactored, ensuring images generated or edited in temporary chats are correctly processed, stored, and displayed without inconsistencies or missing references. [Commit](https://github.com/open-webui/open-webui/commit/423983f)
+- 🎨 **Image action button fixed**, restoring the ability to trigger image generation, editing, and related image actions from the chat UI. [Commit](https://github.com/open-webui/open-webui/commit/def8a00)
+
+## [0.6.42] - 2025-12-21
+
+### Added
+
+- 📚 Knowledge base file management was overhauled with server-side pagination loading 30 files at a time instead of loading entire collections at once, dramatically improving performance and responsiveness for large knowledge bases with hundreds or thousands of files, reducing initial load times and memory usage while adding server-side search and filtering, view options for files added by the user versus shared files, customizable sorting by name or date, and file authorship tracking with upload timestamps. [Commit](https://github.com/open-webui/open-webui/commit/94a8439105f30203ea9d729787c9c5978f5c22a2)
+- ✨ Knowledge base file management was enhanced with automatic list refresh after file operations ensuring immediate UI updates, improved permission validation at the model layer, and automatic channel-file association for files uploaded with channel metadata. [Commit](https://github.com/open-webui/open-webui/commit/c15201620d03a9b60b800a34d8dc3426722c5b8b)
+- 🔎 Knowledge command in chat input now uses server-side search for massive performance increases when selecting knowledge bases and files. [Commit](https://github.com/open-webui/open-webui/commit/0addc1ea461d7b4eee8fe0ca2fedd615b3988b0e)
+- 🗂️ Knowledge workspace listing now uses server-side pagination loading 30 collections at a time with new search endpoints supporting query filtering and view options for created versus shared collections. [Commit](https://github.com/open-webui/open-webui/commit/ceae3d48e603f53313d5483abe94099e20e914e8)
+- 📖 Knowledge workspace now displays all collections with read access including shared read-only collections, enabling users to discover and explore knowledge bases they don't own while maintaining proper access controls through visual "Read Only" badges and automatically disabled editing controls for name, description, file uploads, content editing, and deletion operations. [Commit](https://github.com/open-webui/open-webui/commit/693636d971d0e8398fa0c9ec3897686750007af5)
+- 📁 Bulk website and YouTube video attachment now supports adding multiple URLs at once (newline-separated) with automatic YouTube detection and transcript retrieval, processed sequentially to prevent resource strain, and both websites and videos can now be added directly to knowledge bases through the workspace UI. [Commit](https://github.com/open-webui/open-webui/commit/7746e9f4b831f09953ad2b659b96e0fd52911031), [#6202](https://github.com/open-webui/open-webui/issues/6202), [#19587](https://github.com/open-webui/open-webui/pull/19587), [#8231](https://github.com/open-webui/open-webui/pull/8231)
+- 🪟 Sidebar width is now resizable on desktop devices with persistent storage in localStorage, enforcing minimum and maximum width constraints (220px to 480px) while all layout components now reference the dynamic sidebar width via CSS variables for consistent responsive behavior. [Commit](https://github.com/open-webui/open-webui/commit/b364cf43d3e8fd3557f65f17bc285bfaca5ed368)
+- 📝 Notes feature now supports server-side search and filtering with view options for notes created by the user versus notes shared with them, customizable sorting by name or date in both list and grid view modes within a redesigned interface featuring consolidated note management controls in a unified header, group-based permission sharing with read, write, and read-only access control displaying note authorship and sharing status for better collaboration, and paginated infinite scroll for improved performance with large note collections. [Commit](https://github.com/open-webui/open-webui/commit/9b24cddef6c4862bd899eb8d6332cafff54e871d)
+- 👁️ Notes now support read-only access permissions, allowing users to share notes for viewing without granting edit rights, with the editor automatically becoming non-editable and appropriate UI indicators when read-only access is detected. [Commit](https://github.com/open-webui/open-webui/commit/4363df175d50e0f9729381ac2ba9b37a3c3a966d)
+- 📄 Notes can now be created directly from the chat input field, allowing users to save drafted messages or content as notes without navigation or retyping. [Commit](https://github.com/open-webui/open-webui/commit/00c2b6ca405d617e3d7520953a00a36c19c790ec)
+- 🪟 Sidebar folders, channels, and pinned models sections now automatically expand when creating new items or pinning models, providing immediate visual feedback for user actions. [Commit](https://github.com/open-webui/open-webui/commit/f826d3ed75213a0a1b31b50d030bfb1d5e91d199), [#19929](https://github.com/open-webui/open-webui/pull/19929)
+- 📋 Chat file associations are now properly tracked in the database through a new "chat_file" table, enabling accurate file management across chats and ensuring proper cleanup of files when chats are deleted, while improving database consistency in multi-node deployments. [Commit](https://github.com/open-webui/open-webui/commit/f1bf4f20c53e6493f0eb6fa2f12cb84c2d22da52)
+- 🖼️ User-uploaded images are now automatically converted from base64 to actual file storage on the server, eliminating large inline base64 strings from being stored in chat history and reducing message payload sizes while enabling better image management and sharing across multiple chats. [Commit](https://github.com/open-webui/open-webui/commit/f1bf4f20c53e6493f0eb6fa2f12cb84c2d22da52)
+- 📸 Shared chats with generated or edited images now correctly display images when accessed by other users by properly linking generated images to their chat and message through the chat_file table, ensuring images remain accessible in shared chat links. [Commit](https://github.com/open-webui/open-webui/commit/446cc0ac6063402a743e949f50612376ed5a8437), [#19393](https://github.com/open-webui/open-webui/issues/19393)
+- 📊 File viewer modal was significantly enhanced with native-like viewers for Excel/CSV spreadsheets rendering as interactive scrollable tables with multi-sheet navigation support, Markdown documents displaying with full typography including headers, lists, links, and tables, and source code files showing syntax highlighting, all accessible through a tabbed interface defaulting to raw text view. [#20035](https://github.com/open-webui/open-webui/pull/20035), [#2867](https://github.com/open-webui/open-webui/issues/2867)
+- 📏 Chat input now displays an expand button in the top-right corner when messages exceed two lines, providing optional access to a full-screen editor for composing longer messages with enhanced workspace and visibility while temporarily disabling the main input to prevent editing conflicts. [Commit](https://github.com/open-webui/open-webui/commit/205c7111200c22da42e9b5fe1e676aec9cca6daa)
+- 💬 Channel message data lazy loading was implemented, deferring attachment and file metadata retrieval until needed to improve initial message list load performance. [Commit](https://github.com/open-webui/open-webui/commit/54b7ec56d6bcd2d79addc1694b757dab18cf18c5)
+- 🖼️ Channel image upload handling was optimized to process and store compressed images directly as files rather than inline data, improving memory efficiency and message load times. [Commit](https://github.com/open-webui/open-webui/commit/22f1b764a7ea1add0a896906a9ef00b4b6743adc)
+- 🎥 Video file playback support was added to channel messages, enabling inline video viewing with native player controls. [Commit](https://github.com/open-webui/open-webui/commit/7b126b23d50a0bd36a350fe09dc1dbe3df105318)
+- 🔐 LDAP authentication now supports user entries with multiple username attributes, correctly handling cases where the username field contains a list of values. [Commit](https://github.com/open-webui/open-webui/commit/379f888c9dc6dce21c3ef7a1fc455258aff993dc), [#19878](https://github.com/open-webui/open-webui/issues/19878)
+- 👨‍👩‍👧‍👦 The "ENABLE_PUBLIC_ACTIVE_USERS_COUNT" environment variable now allows restricting active user count visibility to administrators, reducing backend load and addressing privacy concerns in large deployments. [#20027](https://github.com/open-webui/open-webui/pull/20027), [#13026](https://github.com/open-webui/open-webui/issues/13026)
+- 🚀 Models page search input performance was optimized with a 300ms debounce to reduce server load and improve responsiveness. [#19832](https://github.com/open-webui/open-webui/pull/19832)
+- 💨 Frontend performance was optimized by preventing unnecessary API calls for API Keys and Channels features when they are disabled in admin settings, reducing backend noise and improving overall system efficiency. [#20043](https://github.com/open-webui/open-webui/pull/20043), [#19967](https://github.com/open-webui/open-webui/issues/19967)
+- 📎 Channel file association tracking was implemented, automatically linking uploaded files to their respective channels with a dedicated association table enabling better organization and future file management features within channels. [Commit](https://github.com/open-webui/open-webui/commit/2bccf8350d0915f69b8020934bb179c52e81b7b5)
+- 👥 User profile previews now display group membership information for easier identification of user roles and permissions. [Commit](https://github.com/open-webui/open-webui/commit/2b1a29d44bde9fbc20ff9f0a5ded1ce8ded9d90d)
+- 🌍 The "SEARXNG_LANGUAGE" environment variable now allows configuring search language for SearXNG queries, replacing the hardcoded "en-US" default with a configurable setting that defaults to "all". [#19909](https://github.com/open-webui/open-webui/pull/19909)
+- ⏳ The "MINERU_API_TIMEOUT" environment variable now allows configuring request timeouts for MinerU document processing operations. [#20016](https://github.com/open-webui/open-webui/pull/20016), [#18495](https://github.com/open-webui/open-webui/issues/18495)
+- 🔧 The "RAG_EXTERNAL_RERANKER_TIMEOUT" environment variable now allows configuring request timeouts for external reranker operations. [#20049](https://github.com/open-webui/open-webui/pull/20049), [#19900](https://github.com/open-webui/open-webui/issues/19900)
+- 🎨 OpenAI GPT-IMAGE 1.5 model support was added for image generation and editing with automatic image size capabilities. [Commit](https://github.com/open-webui/open-webui/commit/4c2e5c93e9287479f56f780708656136849ccaee)
+- 🔑 The "OAUTH_AUDIENCE" environment variable now allows OAuth providers to specify audience parameters for JWT access token generation. [#19768](https://github.com/open-webui/open-webui/pull/19768)
+- ⏰ The "REDIS_SOCKET_CONNECT_TIMEOUT" environment variable now allows configuring socket connection timeouts for Redis and Sentinel connections, addressing potential failover and responsiveness issues in distributed deployments. [#19799](https://github.com/open-webui/open-webui/pull/19799), [Docs:#882](https://github.com/open-webui/docs/pull/882)
+- ⏱️ The "WEB_LOADER_TIMEOUT" environment variable now allows configuring request timeouts for SafeWebBaseLoader operations. [#19804](https://github.com/open-webui/open-webui/pull/19804), [#19734](https://github.com/open-webui/open-webui/issues/19734)
+- 🚀 Models API endpoint performance was optimized through batched model loading, eliminating N+1 queries and significantly reducing response times when filtering models by user permissions. [Commit](https://github.com/open-webui/open-webui/commit/0dd2cfe1f273fbacdbe90300a97c021f2e678656)
+- 🔀 Custom model fallback handling was added, allowing workspace-created custom models to automatically fall back to the default chat model when their configured base model is not found; set "ENABLE_CUSTOM_MODEL_FALLBACK" to true to enable, preventing workflow disruption when base models are removed or renamed, while ensuring other requests remain unaffected. [Commit](https://github.com/open-webui/open-webui/commit/b35aeb8f46e0e278c6f4538382c2b6838e24cc5a), [#19985](https://github.com/open-webui/open-webui/pull/19985)
+- 📡 A new /feedbacks/all/ids API endpoint was added to return only feedback IDs without metadata, significantly improving performance for external integrations working with large feedback collections. [Commit](https://github.com/open-webui/open-webui/commit/53c1ca64b7205d85f6de06bd69e3e265d15546b8)
+- 📈 An experimental chat usage statistics endpoint (GET /api/v1/chats/stats/usage) was added with pagination support (50 chats per page) and comprehensive per-chat analytics including model usage counts, user and assistant message breakdowns, average response times calculated from message timestamps, average content lengths, and last activity timestamps; this endpoint remains experimental and not suitable for production use as it performs intensive calculations by processing entire message histories for each chat without caching. [Commit](https://github.com/open-webui/open-webui/commit/a7993f6f4e4591cd2aaa4718ece9e5623557d019)
+- 🔄 Various improvements were implemented across the frontend and backend to enhance performance, stability, and security.
+- 🌐 Translations for German, Danish, Finnish, Korean, Portuguese (Brazil), Simplified Chinese, Traditional Chinese, Catalan, and Spanish were enhanced and expanded.
+
+### Fixed
+
+- ⚡ External reranker operations were optimized to prevent event loop blocking by offloading synchronous HTTP requests to a thread pool using asyncio.to_thread(), eliminating application freezes during RAG reranking queries. [#20049](https://github.com/open-webui/open-webui/pull/20049), [#19900](https://github.com/open-webui/open-webui/issues/19900)
+- 💭 Text loss in the explanation feature when using the "CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE" environment variable was resolved by correcting newline handling in streaming responses. [#19829](https://github.com/open-webui/open-webui/pull/19829)
+- 📚 Knowledge base batch file addition failures caused by Pydantic validation errors are now prevented by making the meta field optional in file metadata responses, allowing files without metadata to be processed correctly. [#20022](https://github.com/open-webui/open-webui/pull/20022), [#14220](https://github.com/open-webui/open-webui/issues/14220)
+- 🗄️ PostgreSQL null byte insertion failures when attaching web pages or processing embedded content are now prevented by consolidating text sanitization logic across chat messages, web search results, and knowledge base documents, removing null bytes and invalid UTF-8 surrogates before database insertion. [#20072](https://github.com/open-webui/open-webui/pull/20072), [#19867](https://github.com/open-webui/open-webui/issues/19867), [#18201](https://github.com/open-webui/open-webui/issues/18201), [#15616](https://github.com/open-webui/open-webui/issues/15616)
+- 🎫 MCP OAuth 2.1 token exchange failures are now fixed by removing duplicate credential passing that caused "ID1,ID1" concatenation and 401 errors from the token endpoint. [#20076](https://github.com/open-webui/open-webui/pull/20076), [#19823](https://github.com/open-webui/open-webui/issues/19823)
+- 📝 Notes "Improve" action now works correctly after the streaming API change in v0.6.41 by ensuring uploaded files are fully retrieved with complete metadata before processing, restoring note improvement and summarization functionality. [Commit](https://github.com/open-webui/open-webui/commit/a3458f492c53a3b00405f59fbe1ea953fe364f18), [#20078](https://github.com/open-webui/open-webui/discussions/20078)
+- 🔑 MCP OAuth 2.1 tool servers now work correctly in multi-node deployments through lazy-loading of OAuth clients from Redis-synced configuration, eliminating 404 errors when load balancers route requests to nodes that didn't process the original config update. [#20076](https://github.com/open-webui/open-webui/pull/20076), [#19902](https://github.com/open-webui/open-webui/pull/19902), [#19901](https://github.com/open-webui/open-webui/issues/19901)
+- 🧩 Chat loading failures when channels permissions were disabled are now prevented through graceful error handling. [Commit](https://github.com/open-webui/open-webui/commit/5c2df97f04cce5cb7087d288f816f91a739688c1)
+- 🔍 Search bar freezing and crashing issues in Models, Chat, and Archived Chat pages caused by excessively long queries exceeding server URL limits were resolved by truncating queries to 500 characters, and knowledge base layout shifting with long names was fixed by adjusting flex container properties. [#19832](https://github.com/open-webui/open-webui/pull/19832)
+- 🎛️ Rate limiting errors (HTTP 429) with Brave Search free tier when generating multiple queries are now prevented through asyncio.Semaphore-based concurrency control applied globally to all search engines. [#20070](https://github.com/open-webui/open-webui/pull/20070), [#20003](https://github.com/open-webui/open-webui/issues/20003), [#14107](https://github.com/open-webui/open-webui/issues/14107), [#15134](https://github.com/open-webui/open-webui/issues/15134)
+- 💥 UI crashes and white screen errors caused by null chat lists during loading or network failures were prevented by adding null safety checks to chat iteration in folder placeholders and archived chat modals. [#19898](https://github.com/open-webui/open-webui/pull/19898)
+- 🧩 Chat overview tab crashes caused by undefined model references were resolved by adding proper null checks when accessing deleted or ejected models. [#19935](https://github.com/open-webui/open-webui/pull/19935)
+- 🔄 MultiResponseMessages component crashes when navigating chat history after removing or changing selected models are now prevented through proper component re-initialization. [Commit](https://github.com/open-webui/open-webui/commit/870e29e3738da968c396b70532f365a3c2f71995), [#18599](https://github.com/open-webui/open-webui/issues/18599)
+- 🚫 Channel API endpoint access is now correctly blocked when channels are globally disabled, preventing users with channel permissions from accessing channel data via API requests when the feature is turned off in admin settings. [#19957](https://github.com/open-webui/open-webui/pull/19957), [#19914](https://github.com/open-webui/open-webui/issues/19914)
+- 👤 User list popup display in the admin panel was fixed to correctly track user identity when sorting or filtering changes the list order, preventing popups from showing incorrect user information. [Commit](https://github.com/open-webui/open-webui/commit/ae47101dc6aef2c7d8ae0d843985341fff820057), [#20046](https://github.com/open-webui/open-webui/issues/20046)
+- 👥 User selection in the "Edit User Group" modal now preserves pagination position, allowing administrators to select multiple users across pages without resetting to page 1. [#19959](https://github.com/open-webui/open-webui/pull/19959)
+- 📸 Model avatar images now update immediately in the admin models list through proper Cache-Control headers, eliminating the need for manual cache clearing. [#19959](https://github.com/open-webui/open-webui/pull/19959)
+- 🔒 Temporary chat permission enforcement now correctly prevents users from enabling the feature through personal settings when disabled in default or group permissions. [#19785](https://github.com/open-webui/open-webui/issues/19785)
+- 🎨 Image editing with reference images now correctly uses both previously generated images and newly uploaded reference images. [Commit](https://github.com/open-webui/open-webui/commit/bcd50ed8f1b7387fd700538ae0d74fc72f3c53d0)
+- 🧠 Image generation and editing operations are now explicitly injected into system context, improving LLM comprehension even for weaker models so they reliably acknowledge operations instead of incorrectly claiming they cannot generate images. [Commit](https://github.com/open-webui/open-webui/commit/28b2fcab0cd036dbe646a66fe81890f288c77121)
+- 📑 Source citation rendering errors when citation syntax appeared in user messages or contexts without source data were resolved. [Commit](https://github.com/open-webui/open-webui/commit/3c8f1cf8e58d52e86375634b0381374298b1b4f3)
+- 📄 DOCX file parsing now works correctly in temporary chats through client-side text extraction, preventing raw data from being displayed. [Commit](https://github.com/open-webui/open-webui/commit/6993b0b40b10af8cdbe6626702cc94080fff9e22)
+- 🔧 Pipeline settings save failures when valve properties contain null values are now handled correctly. [#19791](https://github.com/open-webui/open-webui/pull/19791)
+- ⚙️ Model usage settings are now correctly preserved when switching between models instead of being unexpectedly cleared or reset. [#19868](https://github.com/open-webui/open-webui/pull/19868), [#19549](https://github.com/open-webui/open-webui/issues/19549)
+- 🛡️ Invalid PASSWORD_VALIDATION_REGEX_PATTERN configurations no longer cause startup warnings, with automatic fallback to the default pattern when regex compilation fails. [#20058](https://github.com/open-webui/open-webui/pull/20058)
+- 🎯 The DefaultFiltersSelector component in model settings now correctly displays when only global toggleable filters are present, enabling per-model default configuration. [#20066](https://github.com/open-webui/open-webui/pull/20066)
+- 🎤 Audio file upload failures caused by MIME type matching issues with spacing variations and codec parameters were resolved by implementing proper MIME type parsing. [#17771](https://github.com/open-webui/open-webui/pull/17771), [#17761](https://github.com/open-webui/open-webui/issues/17761)
+- ⌨️ Regenerate response keyboard shortcut now only activates when chat input is selected, preventing unintended regeneration when modals are open or other UI elements are focused. [#19875](https://github.com/open-webui/open-webui/pull/19875)
+- 📋 Log truncation issues in Docker deployments during application crashes were resolved by disabling Python stdio buffering, ensuring complete diagnostic output is captured. [#19844](https://github.com/open-webui/open-webui/issues/19844)
+- 🔴 Redis cluster compatibility issues with disabled KEYS command were resolved by replacing blocking KEYS operations with production-safe SCAN iterations. [#19871](https://github.com/open-webui/open-webui/pull/19871), [#15834](https://github.com/open-webui/open-webui/issues/15834)
+- 🔤 File attachment container layout issues when using RTL languages were resolved by applying chat direction settings to file containers across all message types. [#19891](https://github.com/open-webui/open-webui/pull/19891), [#19742](https://github.com/open-webui/open-webui/issues/19742)
+- 🔃 Ollama model list now automatically refreshes after model deletion, preventing deleted models from persisting in the UI and being inadvertently re-downloaded during subsequent pull operations. [#19912](https://github.com/open-webui/open-webui/pull/19912)
+- 🌐 Ollama Cloud web search now correctly applies domain filtering to search results. [Commit](https://github.com/open-webui/open-webui/commit/d4bd938a77c22409a1643c058b937a06e07baca9)
+- 📜 Tool specification serialization now preserves non-ASCII characters including Chinese text, improving LLM comprehension and tool selection accuracy by avoiding Unicode escape sequences. [#19942](https://github.com/open-webui/open-webui/pull/19942)
+- 🛟 Model editor stability was improved with null safety checks for tools, functions, and file input operations, preventing crashes when stores are undefined or file objects are invalid. [#19939](https://github.com/open-webui/open-webui/pull/19939)
+- 🗣️ MoA completion handling stability was improved with null safety checks for response objects, boolean casting for settings, and proper timeout type definitions. [#19921](https://github.com/open-webui/open-webui/pull/19921)
+- 🎛️ Chat functionality failures caused by empty logit_bias parameter values are now prevented by properly handling empty strings in the parameter parsing middleware. [#19982](https://github.com/open-webui/open-webui/issues/19982)
+- 🔏 Administrators can now delete read-only knowledge bases from deleted users, resolving permission issues that previously prevented cleanup of orphaned read-only content. [Commit](https://github.com/open-webui/open-webui/commit/59d6eb2badf46f9c2b1e879484ac33432915b575)
+- 💾 Cloned prompts and tools now correctly preserve their access control settings instead of being reset to null, preventing unintended visibility changes when duplicating private or restricted items. [#19960](https://github.com/open-webui/open-webui/pull/19960), [#19360](https://github.com/open-webui/open-webui/issues/19360)
+- 🎚️ Text scale adjustment buttons in Interface Settings were fixed to correctly increment and decrement the scale value. [#19699](https://github.com/open-webui/open-webui/pull/19699)
+- 🎭 Group channel invite button text visibility in light theme was corrected to display properly against dark backgrounds. [#19828](https://github.com/open-webui/open-webui/issues/19828)
+- 📁 The move button is now hidden when no folders exist, preventing display of non-functional controls. [#19705](https://github.com/open-webui/open-webui/pull/19705)
+- 📦 Qdrant client dependency was updated to resolve startup version incompatibility warnings. [#19757](https://github.com/open-webui/open-webui/pull/19757)
+- 🧮 The "ENABLE_ASYNC_EMBEDDING" environment variable is now correctly applied to embedding operations when configured exclusively via environment variables. [#19748](https://github.com/open-webui/open-webui/pull/19748)
+- 🌄 The "COMFYUI_WORKFLOW_NODES" and "IMAGES_EDIT_COMFYUI_WORKFLOW_NODES" environment variables are now correctly loaded and parsed as JSON lists, and the configuration key name was corrected from "COMFYUI_WORKFLOW" to "COMFYUI_WORKFLOW_NODES". [#19918](https://github.com/open-webui/open-webui/pull/19918), [#19886](https://github.com/open-webui/open-webui/issues/19886)
+- 💫 Channel name length is now limited to 128 characters with validation to prevent display issues caused by excessively long names. [Commit](https://github.com/open-webui/open-webui/commit/f509f5542dde384d34402f6df763f49a06bea109)
+- 🔐 Invalid PASSWORD_VALIDATION_REGEX_PATTERN configurations no longer cause startup warnings, with automatic fallback to the default pattern when regex compilation fails. [#20058](https://github.com/open-webui/open-webui/pull/20058)
+- 🔎 Bocha search with filter list functionality now works correctly by returning results as a list instead of a dictionary wrapper, ensuring compatibility with result filtering operations. [Commit](https://github.com/open-webui/open-webui/commit/b5bd8704fe1672da839bb3be6210d7cb494797ce), [#19733](https://github.com/open-webui/open-webui/issues/19733)
+
+### Changed
+
+- ⚠️ This release includes database schema changes; multi-worker, multi-server, or load-balanced deployments must update all instances simultaneously rather than performing rolling updates, as running mixed versions will cause application failures due to schema incompatibility between old and new instances.
+- 📡 WEB_SEARCH_CONCURRENT_REQUESTS default changed from 10 to 0 (unlimited) — This setting now applies to all search engines instead of only DuckDuckGo; previously users were implicitly limited to 10 concurrent queries, but now have unlimited parallel requests by default; set to 1 for sequential execution if using rate-limited APIs like Brave free tier. [#20070](https://github.com/open-webui/open-webui/pull/20070)
+- 💾 SQLCipher absolute path handling was fixed to properly support absolute database paths (e.g., "/app/data.db") instead of incorrectly stripping leading slashes and converting them to relative paths; this restores functionality for Docker volume mounts and explicit absolute path configurations while maintaining backward compatibility with relative paths. [#20074](https://github.com/open-webui/open-webui/pull/20074)
+- 🔌 Knowledge base file listing API was redesigned with paginated responses and new filtering parameters; the GET /knowledge/{id}/files endpoint now returns paginated results with user attribution instead of embedding all files in the knowledge object, which may require updates to custom integrations or scripts accessing knowledge base data programmatically. [Commit](https://github.com/open-webui/open-webui/commit/94a8439105f30203ea9d729787c9c5978f5c22a2)
+- 🗑️ Legacy knowledge base support for deprecated document collections and tag-based collections was removed; users with pre-knowledge base documents must migrate to the current knowledge base system as legacy items will no longer appear in selectors or command menus. [Commit](https://github.com/open-webui/open-webui/commit/a934dc997ed67a036dd7975e380f8036c447d3ed)
+- 🔨 Source-level log environment variables (AUDIO_LOG_LEVEL, CONFIG_LOG_LEVEL, MODELS_LOG_LEVEL, etc.) were removed as they provided limited configuration options and added significant complexity across 100+ files; the GLOBAL_LOG_LEVEL environment variable, which already took precedence over source-level settings, now serves as the exclusive logging configuration method. [#20045](https://github.com/open-webui/open-webui/pull/20045)
+- 🐍 LangChain was upgraded to version 1.2.0, representing a major dependency update and significant progress toward Python 3.13 compatibility while improving RAG pipeline functionality for document loading and retrieval operations. [#19991](https://github.com/open-webui/open-webui/pull/19991)
+
+## [0.6.41] - 2025-12-02
+
+### Added
+
+- 🚦 Sign-in rate limiting was implemented to protect against brute force attacks, limiting login attempts to 15 per 3-minute window per email address using Redis with automatic fallback to in-memory storage when Redis is unavailable. [Commit](https://github.com/open-webui/open-webui/commit/7b166370432414ce8f186747fb098e0c70fb2d6b)
+- 📂 Administrators can now globally disable the folders feature and control user-level folder permissions through the admin panel, enabling minimalist interface configurations for deployments that don't require workspace organization features. [#19529](https://github.com/open-webui/open-webui/pull/19529), [#19210](https://github.com/open-webui/open-webui/discussions/19210), [#18459](https://github.com/open-webui/open-webui/discussions/18459), [#18299](https://github.com/open-webui/open-webui/discussions/18299)
+- 👥 Group channels were introduced as a new channel type enabling membership-based collaboration spaces where users explicitly join as members rather than accessing through permissions, with support for public or private visibility, automatic member inclusion from specified user groups, member role tracking with invitation metadata, and post-creation member management allowing channel managers to add or remove members through the channel info modal. [Commit](https://github.com/open-webui/open-webui/commit/f589b7c1895a6a77166c047891acfa21bc0936c4), [Commit](https://github.com/open-webui/open-webui/commit/3f1d9ccbf8443a2fa5278f36202bad930a216680)
+- 💬 Direct Message channels were introduced with a dedicated channel type selector and multi-user member selection interface, enabling private conversations between specific users without requiring full channel visibility. [Commit](https://github.com/open-webui/open-webui/commit/64b4d5d9c280b926746584aaf92b447d09deb386)
+- 📨 Direct Message channels now support a complete user-to-user messaging system with member-based access control, automatic deduplication for one-on-one conversations, optional channel naming, and distinct visual presentation using participant avatars instead of channel icons. [Commit](https://github.com/open-webui/open-webui/commit/acccb9afdd557274d6296c70258bb897bbb6652f)
+- 🙈 Users can now hide Direct Message channels from their sidebar while preserving message history, with automatic reactivation when new messages arrive from other participants, providing a cleaner interface for managing active conversations. [Commit](https://github.com/open-webui/open-webui/commit/acccb9afdd557274d6296c70258bb897bbb6652f)
+- ☑️ A comprehensive user selection component was added to the channel creation modal, featuring search functionality, sortable user lists, pagination support, and multi-select checkboxes for building Direct Message participant lists. [Commit](https://github.com/open-webui/open-webui/commit/acccb9afdd557274d6296c70258bb897bbb6652f)
+- 🔴 Channel unread message count tracking was implemented with visual badge indicators in the sidebar, automatically updating counts in real-time and marking messages as read when users view channels, with join/leave functionality to manage membership status. [Commit](https://github.com/open-webui/open-webui/commit/64b4d5d9c280b926746584aaf92b447d09deb386)
+- 📌 Message pinning functionality was added to channels, allowing users to pin important messages for easy reference with visual highlighting, a dedicated pinned messages modal accessible from the navbar, and complete backend support for tracking pinned status, pin timestamp, and the user who pinned each message. [Commit](https://github.com/open-webui/open-webui/commit/64b4d5d9c280b926746584aaf92b447d09deb386), [Commit](https://github.com/open-webui/open-webui/commit/aae2fce17355419d9c29f8100409108037895201)
+- 🟢 Direct Message channels now display an active status indicator for one-on-one conversations, showing a green dot when the other participant is currently online or a gray dot when offline. [Commit](https://github.com/open-webui/open-webui/commit/4b6773885cd7527c5a56b963781dac5e95105eec), [Commit](https://github.com/open-webui/open-webui/commit/39645102d14f34e71b34e5ddce0625790be33f6f)
+- 🆔 Users can now start Direct Message conversations directly from user profile previews by clicking the "Message" button, enabling quick access to private messaging without navigating away from the current channel. [Commit](https://github.com/open-webui/open-webui/commit/a0826ec9fedb56320532616d568fa59dda831d4e)
+- ⚡ Channel messages now appear instantly when sent using optimistic UI rendering, displaying with a pending state while the server confirms delivery, providing a more responsive messaging experience. [Commit](https://github.com/open-webui/open-webui/commit/25994dd3da90600401f53596d4e4fb067c1b8eaa)
+- 👍 Channel message reactions now display the names of users who reacted when hovering over the emoji, showing up to three names with a count for additional reactors. [Commit](https://github.com/open-webui/open-webui/commit/05e79bdd0c7af70b631e958924e3656db1013b80)
+- 🛠️ Channel creators can now edit and delete their own group and DM channels without requiring administrator privileges, enabling users to manage the channels they create independently. [Commit](https://github.com/open-webui/open-webui/commit/f589b7c1895a6a77166c047891acfa21bc0936c4)
+- 🔌 A new API endpoint was added to directly get or create a Direct Message channel with a specific user by their ID, streamlining programmatic DM channel creation for integrations and frontend workflows. [Commit](https://github.com/open-webui/open-webui/commit/f589b7c1895a6a77166c047891acfa21bc0936c4)
+- 💭 Users can now set a custom status with an emoji and message that displays in profile previews, the sidebar user menu, and Direct Message channel items in the sidebar, with the ability to clear status at any time, providing visibility into availability or current focus similar to team communication platforms. [Commit](https://github.com/open-webui/open-webui/commit/51621ba91a982e52da168ce823abffd11ad3e4fa), [Commit](https://github.com/open-webui/open-webui/commit/f5e8d4d5a004115489c35725408b057e24dfe318)
+- 📤 A group export API endpoint was added, enabling administrators to export complete group data including member lists for backup and migration purposes. [Commit](https://github.com/open-webui/open-webui/commit/09b6ea38c579659f8ca43ae5ea3746df3ac561ad)
+- 📡 A new API endpoint was added to retrieve all users belonging to a specific group, enabling programmatic access to group membership information for administrative workflows. [Commit](https://github.com/open-webui/open-webui/commit/01868e856a10f474f74fbd1b4425dafdf949222f)
+- 👁️ The admin user list now displays an active status indicator next to each user, showing a visual green dot for users who have been active within the last three minutes. [Commit](https://github.com/open-webui/open-webui/commit/1b095d12ff2465b83afa94af89ded9593f8a8655)
+- 🔑 The admin user edit modal now displays OAuth identity information with a per-provider breakdown, showing each linked identity provider and its associated subject identifier separately. [#19573](https://github.com/open-webui/open-webui/pull/19573)
+- 🧩 OAuth role claim parsing now respects the "OAUTH_ROLES_SEPARATOR" configuration, enabling proper parsing of roles returned as comma-separated strings and providing consistent behavior with group claim handling. [#19514](https://github.com/open-webui/open-webui/pull/19514)
+- 🎛️ Channel feature access can now be controlled through both the "USER_PERMISSIONS_FEATURES_CHANNELS" environment variable and group permission toggles in the admin panel, allowing administrators to restrict channel functionality for specific users or groups while defaulting to enabled for all users. [Commit](https://github.com/open-webui/open-webui/commit/f589b7c1895a6a77166c047891acfa21bc0936c4)
+- 🎨 The model editor interface was refined with access control settings moved to a dedicated modal, group member counts now displayed when configuring permissions, reorganized layout with improved visual hierarchy, and redesigned prompt suggestions cards with tooltips for field guidance. [Commit](https://github.com/open-webui/open-webui/commit/e65d92fc6f49da5ca059e1c65a729e7973354b99), [Commit](https://github.com/open-webui/open-webui/commit/9d39b9b42c653ee2acf2674b2df343ecbceb4954)
+- 🏗️ Knowledge base file management was rebuilt with a dedicated database table replacing the previous JSON array storage, enabling pagination support for large knowledge bases, significantly faster file listing performance, and more reliable file-knowledge base relationship tracking. [Commit](https://github.com/open-webui/open-webui/commit/d19023288e2ca40f86e2dc3fd9f230540f3e70d7)
+- ☁️ Azure Document Intelligence model selection was added, allowing administrators to specify which model to use for document processing via the "DOCUMENT_INTELLIGENCE_MODEL" environment variable or admin UI setting, with "prebuilt-layout" as the default. [#19692](https://github.com/open-webui/open-webui/pull/19692), [Docs:#872](https://github.com/open-webui/docs/pull/872)
+- 🚀 Milvus multitenancy vector database performance was improved by removing manual flush calls after upsert operations, eliminating rate limit errors and reducing load on etcd and MinIO/S3 storage by allowing Milvus to manage segment persistence automatically via its WAL and auto-flush policies. [#19680](https://github.com/open-webui/open-webui/pull/19680)
+- ✨ Various improvements were implemented across the frontend and backend to enhance performance, stability, and security.
+- 🌍 Translations for German, French, Portuguese (Brazil), Catalan, Simplified Chinese, and Traditional Chinese were enhanced and expanded.
+
+### Fixed
+
+- 🔄 Tool call response token duplication was fixed by removing redundant message history additions in non-native function calling mode, resolving an issue where tool results were included twice in the context and causing 2x token consumption. [#19656](https://github.com/open-webui/open-webui/issues/19656), [Commit](https://github.com/open-webui/open-webui/commit/52ccab8)
+- 🛡️ Web search domain filtering was corrected to properly block results when any resolved hostname or IP address matches a blocked domain, preventing blocked sites from appearing in search results due to permissive hostname resolution logic that previously allowed results through if any single resolved address passed the filter. [#19670](https://github.com/open-webui/open-webui/pull/19670), [#19669](https://github.com/open-webui/open-webui/issues/19669)
+- 🧠 Custom models based on Ollama or OpenAI now properly inherit the connection type from their base model, ensuring they appear correctly in the "Local" or "External" model selection tabs instead of only appearing under "All". [#19183](https://github.com/open-webui/open-webui/issues/19183), [Commit](https://github.com/open-webui/open-webui/commit/39f7575)
+- 🐍 SentenceTransformers embedding initialization was fixed by updating the transformers dependency to version 4.57.3, resolving a regression in v0.6.40 where document ingestion failed with "'NoneType' object has no attribute 'encode'" errors due to a bug in transformers 4.57.2. [#19512](https://github.com/open-webui/open-webui/issues/19512), [#19513](https://github.com/open-webui/open-webui/pull/19513)
+- 📈 Active user count accuracy was significantly improved by replacing the socket-based USER_POOL tracking with a database-backed heartbeat mechanism, resolving long-standing issues where Redis deployments displayed inflated user counts due to stale sessions never being cleaned up on disconnect. [#16074](https://github.com/open-webui/open-webui/discussions/16074), [Commit](https://github.com/open-webui/open-webui/commit/70948f8803e417459d5203839f8077fdbfbbb213)
+- 👥 Default group assignment now applies consistently across all user registration methods including OAuth/SSO, LDAP, and admin-created users, fixing an issue where the "DEFAULT_GROUP_ID" setting was only being applied to users who signed up via the email/password signup form. [#19685](https://github.com/open-webui/open-webui/pull/19685)
+- 🔦 Model list filtering in workspaces was corrected to properly include models shared with user groups, ensuring members can view models they have write access to through group permissions. [#19461](https://github.com/open-webui/open-webui/issues/19461), [Commit](https://github.com/open-webui/open-webui/commit/69722ba973768a5f689f2e2351bf583a8db9bba8)
+- 🖼️ User profile image display in preview contexts was fixed by resolving a Pydantic validation error that prevented proper rendering. [Commit](https://github.com/open-webui/open-webui/commit/c7eb7136893b0ddfdc5d55ffc7a05bd84a00f5d6)
+- 🔒 Redis TLS connection failures were resolved by updating the python-socketio dependency to version 5.15.0, restoring support for the "rediss://" URL schema. [#19480](https://github.com/open-webui/open-webui/issues/19480), [#19488](https://github.com/open-webui/open-webui/pull/19488)
+- 📝 MCP tool server configuration was corrected to properly handle the "Function Name Filter List" as both string and list types, preventing AttributeError when the field is empty and ensuring backward compatibility. [#19486](https://github.com/open-webui/open-webui/issues/19486), [Commit](https://github.com/open-webui/open-webui/commit/c5b73d71843edc024325d4a6e625ec939a747279), [Commit](https://github.com/open-webui/open-webui/commit/477097c2e42985c14892301d0127314629d07df1)
+- 📎 Web page attachment failures causing TypeError on metadata checks were resolved by correcting async threadpool parameter passing in vector database operations. [#19493](https://github.com/open-webui/open-webui/issues/19493), [Commit](https://github.com/open-webui/open-webui/commit/4370dee79e19d77062c03fba81780cb3b779fca3)
+- 💾 Model allowlist persistence in multi-worker deployments was fixed by implementing Redis-based shared state for the internal models dictionary, ensuring configuration changes are consistently visible across all worker processes. [#19395](https://github.com/open-webui/open-webui/issues/19395), [Commit](https://github.com/open-webui/open-webui/commit/b5e5617d7f7ad3e4eec9f15f4cc7f07cb5afc2fa)
+- ⏳ Chat history infinite loading was prevented by enhancing message data structure to properly track parent message relationships, resolving issues where missing parentId fields caused perpetual loading states. [#19225](https://github.com/open-webui/open-webui/issues/19225), [Commit](https://github.com/open-webui/open-webui/commit/ff4b1b9862d15adfa15eac17d2ce066c3d8ae38f)
+- 🩹 Database migration robustness was improved by automatically detecting and correcting missing primary key constraints on the user table, ensuring successful schema upgrades for databases with non-standard configurations. [#19487](https://github.com/open-webui/open-webui/discussions/19487), [Commit](https://github.com/open-webui/open-webui/commit/453ea9b9a167c0b03d86c46e6efd086bf10056ce)
+- 🏷️ OAuth group assignment now updates correctly on first login when users transition from admin to user role, ensuring group memberships reflect immediately when group management is enabled. [#19475](https://github.com/open-webui/open-webui/issues/19475), [#19476](https://github.com/open-webui/open-webui/pull/19476)
+- 💡 Knowledge base file tooltips now properly display the parent collection name when referencing files with the hash symbol, preventing confusion between identically-named files in different collections. [#19491](https://github.com/open-webui/open-webui/issues/19491), [Commit](https://github.com/open-webui/open-webui/commit/3fe5a47b0ff84ac97f8e4ff56a19fa2ec065bf66)
+- 🔐 Knowledge base file access inconsistencies were resolved where authorized non-admin users received "Not found" or permission errors for certain files due to race conditions during upload causing mismatched collection_name values, with file access validation now properly checking against knowledge base file associations. [#18689](https://github.com/open-webui/open-webui/issues/18689), [#19523](https://github.com/open-webui/open-webui/pull/19523), [Commit](https://github.com/open-webui/open-webui/commit/e301d1962e45900ababd3eabb7e9a2ad275a5761)
+- 📦 Knowledge API batch file addition endpoint was corrected to properly handle async operations, resolving 500 Internal Server Error responses when adding multiple files simultaneously. [#19538](https://github.com/open-webui/open-webui/issues/19538), [Commit](https://github.com/open-webui/open-webui/commit/28659f60d94feb4f6a99bb1a5b54d7f45e5ea10f)
+- 🤖 Embedding model auto-update functionality was fixed to properly respect the "RAG_EMBEDDING_MODEL_AUTO_UPDATE" setting by correctly passing the flag to the model path resolver, ensuring models update as expected when the auto-update option is enabled. [#19687](https://github.com/open-webui/open-webui/pull/19687)
+- 📉 API response payload sizes were dramatically reduced by removing base64-encoded profile images from most endpoints, eliminating multi-megabyte responses caused by high-resolution avatars and enabling better browser caching. [#19519](https://github.com/open-webui/open-webui/issues/19519), [Commit](https://github.com/open-webui/open-webui/commit/384753c4c17f62a68d38af4bbcf55a21ee08e0f2)
+- 📞 Redundant API calls on the admin user overview page were eliminated by consolidating reactive statements, reducing four duplicate requests to a single efficient call and significantly improving page load performance. [#19509](https://github.com/open-webui/open-webui/issues/19509), [Commit](https://github.com/open-webui/open-webui/commit/9f89cc5e9f7e1c6c9e2bc91177e08df7c79f66f9)
+- 🧹 Duplicate API calls on the workspace models page were eliminated by removing redundant model list fetching, reducing two identical requests to a single call and improving page responsiveness. [#19517](https://github.com/open-webui/open-webui/issues/19517), [Commit](https://github.com/open-webui/open-webui/commit/d1bbf6be7a4d1d53fa8ad46ca4f62fc4b2e6a8cb)
+- 🔘 The model valves button was corrected to prevent unintended form submission by adding explicit button type attribute, ensuring it no longer triggers message sending when the input area contains text. [#19534](https://github.com/open-webui/open-webui/pull/19534)
+- 🗑️ Ollama model deletion was fixed by correcting the request payload format and ensuring the model selector properly displays the placeholder option. [Commit](https://github.com/open-webui/open-webui/commit/0f3156651c64bc5af188a65fc2908bdcecf30c74)
+- 🎨 Image generation in temporary chats was fixed by correctly handling local chat sessions that are not persisted to the database. [Commit](https://github.com/open-webui/open-webui/commit/a7c7993bbf3a21cb7ba416525b89233cf2ad877f)
+- 🕵️‍♂️ Audit logging was fixed by correctly awaiting the async user authentication call, resolving failures where coroutine objects were passed instead of user data. [#19658](https://github.com/open-webui/open-webui/pull/19658), [Commit](https://github.com/open-webui/open-webui/commit/dba86bc)
+- 🌙 Dark mode select dropdown styling was corrected to use proper background colors, fixing an issue where dropdown borders and hover states appeared white instead of matching the dark theme. [#19693](https://github.com/open-webui/open-webui/pull/19693), [#19442](https://github.com/open-webui/open-webui/issues/19442)
+- 🔍 Milvus vector database query filtering was fixed by correcting string quote handling in filter expressions and using the proper parameter name for queries, resolving false "duplicate content detected" errors that prevented uploading multiple files to knowledge bases. [#19602](https://github.com/open-webui/open-webui/pull/19602), [#18119](https://github.com/open-webui/open-webui/issues/18119), [#16345](https://github.com/open-webui/open-webui/issues/16345), [#17088](https://github.com/open-webui/open-webui/issues/17088), [#18485](https://github.com/open-webui/open-webui/issues/18485)
+- 🆙 Milvus multitenancy vector database was updated to use query_iterator() for improved robustness and consistency with the standard Milvus implementation, fixing the same false duplicate detection errors and improving handling of large result sets in multi-tenant deployments. [#19695](https://github.com/open-webui/open-webui/pull/19695)
+
+### Changed
+
+- ⚠️ **IMPORTANT for Multi-Instance Deployments** — This release includes database schema changes; multi-worker, multi-server, or load-balanced deployments must update all instances simultaneously rather than performing rolling updates, as running mixed versions will cause application failures due to schema incompatibility between old and new instances.
+- 👮 Channel creation is now restricted to administrators only, with the channel add button hidden for regular users to maintain organizational control over communication channels. [Commit](https://github.com/open-webui/open-webui/commit/421aba7cd7cd708168b1f2565026c74525a67905)
+- ➖ The active user count indicator was removed from the bottom-left user menu in the sidebar to streamline the interface. [Commit](https://github.com/open-webui/open-webui/commit/848f3fd4d86ca66656e0ff0335773945af8d7d8d)
+- 🗂️ The user table was restructured with API keys migrated to a dedicated table supporting future multi-key functionality, OAuth data storage converted to a JSON structure enabling multiple identity providers per user account, and internal column types optimized from TEXT to JSON for the "info" and "settings" fields, with automatic migration preserving all existing data and associations. [#19573](https://github.com/open-webui/open-webui/pull/19573)
+- 🔄 The knowledge base API was restructured to support the new file relationship model.
+
+## [0.6.40] - 2025-11-25
+
+### Fixed
+
+- 🗄️ A critical PostgreSQL user listing performance issue was resolved by removing a redundant count operation that caused severe database slowdowns and potential timeouts when viewing user lists in admin panels.
+
+## [0.6.39] - 2025-11-25
+
+### Added
+
+- 💬 A user list modal was added to channels, displaying all users with access and featuring search, sorting, and pagination capabilities. [Commit](https://github.com/open-webui/open-webui/commit/c0e120353824be00a2ef63cbde8be5d625bd6fd0)
+- 💬 Channel navigation now displays the total number of users with access to the channel. [Commit](https://github.com/open-webui/open-webui/commit/3b5710d0cd445cf86423187f5ee7c40472a0df0b)
+- 🔌 Tool servers and MCP connections now support function name filtering, allowing administrators to selectively enable or block specific functions using allow/block lists. [Commit](https://github.com/open-webui/open-webui/commit/743199f2d097ae1458381bce450d9025a0ab3f3d)
+- ⚡ A toggle to disable parallel embedding processing was added via "ENABLE_ASYNC_EMBEDDING", allowing sequential processing for rate-limited or resource-constrained local embedding setups. [#19444](https://github.com/open-webui/open-webui/pull/19444)
+- 🔄 Various improvements were implemented across the frontend and backend to enhance performance, stability, and security.
+- 🌐 Localization improvements were made for German (de-DE) and Portuguese (Brazil) translations.
+
+### Fixed
+
+- 📝 Inline citations now render correctly within markdown lists and nested elements instead of displaying as "undefined" values. [#19452](https://github.com/open-webui/open-webui/issues/19452)
+- 👥 Group member selection now works correctly without randomly selecting other users or causing the user list to jump around. [#19426](https://github.com/open-webui/open-webui/issues/19426)
+- 👥 Admin panel user list now displays the correct total user count and properly paginates 30 items per page after fixing database query issues with group member joins. [#19429](https://github.com/open-webui/open-webui/issues/19429)
+- 🔍 Knowledge base reindexing now works correctly after resolving async execution chain issues by implementing threadpool workers for embedding operations. [#19434](https://github.com/open-webui/open-webui/pull/19434)
+- 🖼️ OpenAI image generation now works correctly after fixing a connection adapter error caused by incorrect URL formatting. [#19435](https://github.com/open-webui/open-webui/pull/19435)
+
+### Changed
+
+- 🔧 BREAKING: Docling configuration has been consolidated from individual environment variables into a single "DOCLING_PARAMS" JSON configuration and now supports API key authentication via "DOCLING_API_KEY", requiring users to migrate existing Docling settings to the new format. [#16841](https://github.com/open-webui/open-webui/issues/16841), [#19427](https://github.com/open-webui/open-webui/pull/19427)
+- 🔧 The environment variable "REPLACE_IMAGE_URLS_IN_CHAT_RESPONSE" has been renamed to "ENABLE_CHAT_RESPONSE_BASE64_IMAGE_URL_CONVERSION" for naming consistency.
+
+## [0.6.38] - 2025-11-24
+
+### Fixed
+
+- 🔍 Hybrid search now works reliably after recent changes.
+- 🛠️ Tool server saving now handles errors gracefully, preventing failed saves from impacting the UI.
+- 🔐 SSO/OIDC code fixed to improve login reliability and better handle edge cases.
+
+## [0.6.37] - 2025-11-24
+
+### Added
+
+- 🔐 Granular sharing permissions are now available with two-tiered control separating group sharing from public sharing, allowing administrators to independently configure whether users can share workspace items with groups or make them publicly accessible, with separate permission toggles for models, knowledge bases, prompts, tools, and notes, configurable via "USER_PERMISSIONS_WORKSPACE_MODELS_ALLOW_SHARING", "USER_PERMISSIONS_WORKSPACE_MODELS_ALLOW_PUBLIC_SHARING", and corresponding environment variables for other workspace item types, while groups can now be configured to opt-out of sharing via the "Allow Group Sharing" setting. [Commit](https://github.com/open-webui/open-webui/commit/7be750bcbb40da91912a0a66b7ab791effdcc3b6), [Commit](https://github.com/open-webui/open-webui/commit/f69e37a8507d6d57382d6670641b367f3127f90a)
+- 🔐 Password policy enforcement is now available with configurable validation rules, allowing administrators to require specific password complexity requirements via "ENABLE_PASSWORD_VALIDATION" and "PASSWORD_VALIDATION_REGEX_PATTERN" environment variables, with default pattern requiring minimum 8 characters including uppercase, lowercase, digit, and special character. [#17794](https://github.com/open-webui/open-webui/pull/17794)
+- 🔐 Granular import and export permissions are now available for workspace items, introducing six separate permission toggles for models, prompts, and tools that are disabled by default for enhanced security. [#19242](https://github.com/open-webui/open-webui/pull/19242)
+- 👥 Default group assignment is now available for new users, allowing administrators to automatically assign newly registered users to a specified group for streamlined access control to models, prompts, and tools, particularly useful for organizations with group-based model access policies. [#19325](https://github.com/open-webui/open-webui/pull/19325), [#17842](https://github.com/open-webui/open-webui/issues/17842)
+- 🔒 Password-based authentication can now be fully disabled via "ENABLE_PASSWORD_AUTH" environment variable, enforcing SSO-only authentication and preventing password login fallback when SSO is configured. [#19113](https://github.com/open-webui/open-webui/pull/19113)
+- 🖼️ Large stream chunk handling was implemented to support models that generate images directly in their output responses, with configurable buffer size via "CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE" environment variable, resolving compatibility issues with models like Gemini 2.5 Flash Image. [#18884](https://github.com/open-webui/open-webui/pull/18884), [#17626](https://github.com/open-webui/open-webui/issues/17626)
+- 🖼️ Streaming response middleware now handles images in delta updates with automatic base64 conversion, enabling proper display of images from models using the "choices[0].delta.images.image_url" format such as Gemini 2.5 Flash Image Preview on OpenRouter. [#19073](https://github.com/open-webui/open-webui/pull/19073), [#19019](https://github.com/open-webui/open-webui/issues/19019)
+- 📈 Model list API performance was optimized by pre-fetching user group memberships and removing profile image URLs from response payloads, significantly reducing both database queries and payload size for instances with large model lists, with profile images now served dynamically via dedicated endpoints. [#19097](https://github.com/open-webui/open-webui/pull/19097), [#18950](https://github.com/open-webui/open-webui/issues/18950)
+- ⏩ Batch file processing performance was improved by reducing database queries by 67% while ensuring data consistency between vector and relational databases. [#18953](https://github.com/open-webui/open-webui/pull/18953)
+- 🚀 Chat import performance was dramatically improved by replacing individual per-chat API requests with a bulk import endpoint, reducing import time by up to 95% for large chat collections and providing user feedback via toast notifications displaying the number of successfully imported chats. [#17861](https://github.com/open-webui/open-webui/pull/17861)
+- ⚡ Socket event broadcasting performance was optimized by implementing user-specific rooms, significantly reducing server overhead particularly for users with multiple concurrent sessions. [#18996](https://github.com/open-webui/open-webui/pull/18996)
+- 🗄️ Weaviate is now supported as a vector database option, providing an additional choice for RAG document storage alongside existing ChromaDB, Milvus, Qdrant, and OpenSearch integrations. [#14747](https://github.com/open-webui/open-webui/pull/14747)
+- 🗄️ PostgreSQL pgvector now supports HNSW index types and large dimensional embeddings exceeding 2000 dimensions through automatic halfvec type selection, with configurable index methods via "PGVECTOR_INDEX_METHOD", "PGVECTOR_HNSW_M", "PGVECTOR_HNSW_EF_CONSTRUCTION", and "PGVECTOR_IVFFLAT_LISTS" environment variables. [#19158](https://github.com/open-webui/open-webui/pull/19158), [#16890](https://github.com/open-webui/open-webui/issues/16890)
+- 🔍 Azure AI Search is now supported as a web search provider, enabling integration with Azure's cognitive search services via "AZURE_AI_SEARCH_API_KEY", "AZURE_AI_SEARCH_ENDPOINT", and "AZURE_AI_SEARCH_INDEX_NAME" configuration. [#19104](https://github.com/open-webui/open-webui/pull/19104)
+- ⚡ External embedding generation now processes API requests in parallel instead of sequential batches, reducing document processing time by 10-50x when using OpenAI, Azure OpenAI, or Ollama embedding providers, with large PDFs now processing in seconds instead of minutes. [#19296](https://github.com/open-webui/open-webui/pull/19296)
+- 💨 Base64 image conversion is now available for markdown content in chat responses, automatically uploading embedded images exceeding 1KB and replacing them with file URLs to reduce payload size and resource consumption, configurable via "REPLACE_IMAGE_URLS_IN_CHAT_RESPONSE" environment variable. [#19076](https://github.com/open-webui/open-webui/pull/19076)
+- 🎨 OpenAI image generation now supports additional API parameters including quality settings for GPT Image 1, configurable via "IMAGES_OPENAI_API_PARAMS" environment variable or through the admin interface, enabling cost-effective image generation with low, medium, or high quality options. [#19228](https://github.com/open-webui/open-webui/issues/19228)
+- 🖼️ Image editing can now be independently enabled or disabled via admin settings, allowing administrators to control whether sequential image prompts trigger image editing or new image generation, configurable via "ENABLE_IMAGE_EDIT" environment variable. [#19284](https://github.com/open-webui/open-webui/issues/19284)
+- 🔐 SSRF protection was implemented with a configurable URL blocklist that prevents access to cloud metadata endpoints and private networks, with default protections for AWS, Google Cloud, Azure, and Alibaba Cloud metadata services, customizable via "WEB_FETCH_FILTER_LIST" environment variable. [#19201](https://github.com/open-webui/open-webui/pull/19201)
+- ⚡ Workspace models page now supports server-side pagination dramatically improving load times and usability for instances with large numbers of workspace models.
+- 🔍 Hybrid search now indexes file metadata including filenames, titles, headings, sources, and snippets alongside document content, enabling keyword queries to surface documents where search terms appear only in metadata, configurable via "ENABLE_RAG_HYBRID_SEARCH_ENRICHED_TEXTS" environment variable. [#19095](https://github.com/open-webui/open-webui/pull/19095)
+- 📂 Knowledge base upload page now supports folder drag-and-drop with recursive directory handling, enabling batch uploads of entire directory structures instead of requiring individual file selection. [#19320](https://github.com/open-webui/open-webui/pull/19320)
+- 🤖 Model cloning is now available in admin settings, allowing administrators to quickly create workspace models based on existing base models through a "Clone" option in the model dropdown menu. [#17937](https://github.com/open-webui/open-webui/pull/17937)
+- 🎨 UI scale adjustment is now available in interface settings, allowing users to increase the size of the entire interface from 1.0x to 1.5x for improved accessibility and readability, particularly beneficial for users with visual impairments. [#19186](https://github.com/open-webui/open-webui/pull/19186)
+- 📌 Default pinned models can now be configured by administrators for all new users, mirroring the behavior of default models where admin-configured defaults apply only to users who haven't customized their pinned models, configurable via "DEFAULT_PINNED_MODELS" environment variable. [#19273](https://github.com/open-webui/open-webui/pull/19273)
+- 🎙️ Text-to-Speech and Speech-to-Text services now receive user information headers when "ENABLE_FORWARD_USER_INFO_HEADERS" is enabled, allowing external TTS and STT providers to implement user-specific personalization, rate limiting, and usage tracking. [#19323](https://github.com/open-webui/open-webui/pull/19323), [#19312](https://github.com/open-webui/open-webui/issues/19312)
+- 🎙️ Voice mode now supports custom system prompts via "VOICE_MODE_PROMPT_TEMPLATE" configuration, allowing administrators to control response style and behavior for voice interactions. [#18607](https://github.com/open-webui/open-webui/pull/18607)
+- 🔧 WebSocket and Redis configuration options are now available including debug logging controls, custom ping timeout and interval settings, and arbitrary Redis connection options via "WEBSOCKET_SERVER_LOGGING", "WEBSOCKET_SERVER_ENGINEIO_LOGGING", "WEBSOCKET_SERVER_PING_TIMEOUT", "WEBSOCKET_SERVER_PING_INTERVAL", and "WEBSOCKET_REDIS_OPTIONS" environment variables. [#19091](https://github.com/open-webui/open-webui/pull/19091)
+- 🔧 MCP OAuth dynamic client registration now automatically detects and uses the appropriate token endpoint authentication method from server-supported options, enabling compatibility with OAuth servers that only support "client_secret_basic" instead of "client_secret_post". [#19193](https://github.com/open-webui/open-webui/issues/19193)
+- 🔧 Custom headers can now be configured for remote MCP and OpenAPI tool server connections, enabling integration with services that require additional authentication headers. [#18918](https://github.com/open-webui/open-webui/issues/18918)
+- 🔍 Perplexity Search now supports custom API endpoints via "PERPLEXITY_SEARCH_API_URL" configuration and automatically forwards user information headers to enable personalized search experiences. [#19147](https://github.com/open-webui/open-webui/pull/19147)
+- 🔍 User information headers can now be optionally forwarded to external web search engines when "ENABLE_FORWARD_USER_INFO_HEADERS" is enabled. [#19043](https://github.com/open-webui/open-webui/pull/19043)
+- 📊 Daily active user metric is now available for monitoring, tracking unique users active since midnight UTC via the "webui.users.active.today" Prometheus gauge. [#19236](https://github.com/open-webui/open-webui/pull/19236), [#19234](https://github.com/open-webui/open-webui/issues/19234)
+- 📊 Audit log file path is now configurable via "AUDIT_LOGS_FILE_PATH" environment variable, enabling storage in separate volumes or custom locations. [#19173](https://github.com/open-webui/open-webui/pull/19173)
+- 🎨 Sidebar collapse states for model lists and group information are now persistent across page refreshes, remembering user preferences through browser-based storage. [#19159](https://github.com/open-webui/open-webui/issues/19159)
+- 🎨 Background image display was enhanced with semi-transparent overlays for navbar and sidebar, creating a seamless and visually cohesive design across the entire interface. [#19157](https://github.com/open-webui/open-webui/issues/19157)
+- 📋 Tables in chat messages now include a copy button that appears on hover, enabling quick copying of table content alongside the existing CSV export functionality. [#19162](https://github.com/open-webui/open-webui/issues/19162)
+- 📝 Notes can now be created directly via the "/notes/new" URL endpoint with optional title and content query parameters, enabling faster note creation through bookmarks and shortcuts. [#19195](https://github.com/open-webui/open-webui/issues/19195)
+- 🏷️ Tag suggestions are now context-aware, displaying only relevant tags when creating or editing models versus chat conversations, preventing confusion between model and chat tags. [#19135](https://github.com/open-webui/open-webui/issues/19135)
+- ✍️ Prompt autocompletion is now available independently of the rich text input setting, improving accessibility to the feature. [#19150](https://github.com/open-webui/open-webui/issues/19150)
+- 🔄 Various improvements were implemented across the frontend and backend to enhance performance, stability, and security.
+- 🌐 Translations for Simplified Chinese, Traditional Chinese, Portuguese (Brazil), Catalan, Spanish (Spain), Finnish, Irish, Farsi, Swedish, Danish, German, Korean, and Thai were improved and expanded.
+
+### Fixed
+
+- 🤖 Model update functionality now works correctly, resolving a database parameter binding error that prevented saving changes to model configurations via the Save & Update button. [#19335](https://github.com/open-webui/open-webui/issues/19335)
+- 🖼️ Multiple input images for image editing and generation are now correctly passed as an array using the "image[]" parameter syntax, enabling proper multi-image reference functionality with models like GPT Image 1. [#19339](https://github.com/open-webui/open-webui/issues/19339)
+- 📱 PWA installations on iOS now properly refresh after server container restarts, resolving freezing issues by automatically unregistering service workers when version or deployment changes are detected. [#19316](https://github.com/open-webui/open-webui/pull/19316)
+- 🗄️ S3 Vectors collection detection now correctly handles buckets with more than 2000 indexes by using direct index lookup instead of paginated list scanning, improving performance by approximately 8x and enabling RAG queries to work reliably at scale. [#19238](https://github.com/open-webui/open-webui/pull/19238), [#19233](https://github.com/open-webui/open-webui/issues/19233)
+- 📈 Feedback retrieval performance was optimized by eliminating N+1 query patterns through database joins, adding server-side pagination and sorting, significantly reducing database load for instances with large feedback datasets. [#17976](https://github.com/open-webui/open-webui/pull/17976)
+- 🔍 Chat search now works correctly with PostgreSQL when chat data contains null bytes, with comprehensive sanitization preventing null bytes during data writes, cleaning existing data on read, and stripping null bytes during search queries to ensure reliable search functionality. [#15616](https://github.com/open-webui/open-webui/issues/15616)
+- 🔍 Hybrid search with reranking now correctly handles attribute validation, preventing errors when collection results lack expected structure. [#19025](https://github.com/open-webui/open-webui/pull/19025), [#17046](https://github.com/open-webui/open-webui/issues/17046)
+- 🔎 Reranking functionality now works correctly after recent refactoring, resolving crashes caused by incorrect function argument handling. [#19270](https://github.com/open-webui/open-webui/pull/19270)
+- 🤖 Azure OpenAI models now support the "reasoning_effort" parameter, enabling proper configuration of reasoning capabilities for models like GPT-5.1 which default to no reasoning without this setting. [#19290](https://github.com/open-webui/open-webui/issues/19290)
+- 🤖 Models with very long IDs can now be deleted correctly, resolving URL length limitations that previously prevented management operations on such models. [#18230](https://github.com/open-webui/open-webui/pull/18230)
+- 🤖 Model-level streaming settings now correctly apply to API requests, ensuring "Stream Chat Response" toggle properly controls the streaming parameter. [#19154](https://github.com/open-webui/open-webui/issues/19154)
+- 🖼️ Image editing configuration now correctly preserves independent OpenAI API endpoints and keys, preventing them from being overwritten by image generation settings. [#19003](https://github.com/open-webui/open-webui/issues/19003)
+- 🎨 Gemini image edit settings now display correctly in the admin panel, fixing an incorrect configuration key reference that prevented proper rendering of edit options. [#19200](https://github.com/open-webui/open-webui/pull/19200)
+- 🖌️ Image generation settings menu now loads correctly, resolving validation errors with AUTOMATIC1111 API authentication parameters. [#19187](https://github.com/open-webui/open-webui/issues/19187), [#19246](https://github.com/open-webui/open-webui/issues/19246)
+- 📅 Date formatting in chat search and admin user chat search now correctly respects the "DEFAULT_LOCALE" environment variable, displaying dates according to the configured locale instead of always using MM/DD/YYYY format. [#19305](https://github.com/open-webui/open-webui/pull/19305), [#19020](https://github.com/open-webui/open-webui/issues/19020)
+- 📝 RAG template query placeholder escaping logic was corrected to prevent unintended replacements of context values when query placeholders appear in retrieved content. [#19102](https://github.com/open-webui/open-webui/pull/19102), [#19101](https://github.com/open-webui/open-webui/issues/19101)
+- 📄 RAG template prompt duplication was eliminated by removing redundant user query section from the default template. [#19099](https://github.com/open-webui/open-webui/pull/19099), [#19098](https://github.com/open-webui/open-webui/issues/19098)
+- 📋 MinerU local mode configuration no longer incorrectly requires an API key, allowing proper use of local content extraction without external API credentials. [#19258](https://github.com/open-webui/open-webui/issues/19258)
+- 📊 Excel file uploads now work correctly with the addition of the missing msoffcrypto-tool dependency, resolving import errors introduced by the unstructured package upgrade. [#19153](https://github.com/open-webui/open-webui/issues/19153)
+- 📑 Docling parameters now properly handle JSON serialization, preventing exceptions and ensuring configuration changes are saved correctly. [#19072](https://github.com/open-webui/open-webui/pull/19072)
+- 🛠️ UserValves configuration now correctly isolates settings per tool, preventing configuration contamination when multiple tools with UserValves are used simultaneously. [#19185](https://github.com/open-webui/open-webui/pull/19185), [#15569](https://github.com/open-webui/open-webui/issues/15569)
+- 🔧 Tool selection prompt now correctly handles user messages without duplication, removing redundant query prefixes and improving prompt clarity. [#19122](https://github.com/open-webui/open-webui/pull/19122), [#19121](https://github.com/open-webui/open-webui/issues/19121)
+- 📝 Notes chat feature now correctly submits messages to the completions endpoint, resolving errors that prevented AI model interactions. [#19079](https://github.com/open-webui/open-webui/pull/19079)
+- 📝 Note PDF downloads now sanitize HTML content using DOMPurify before rendering, preventing potential DOM-based XSS attacks from malicious content in notes. [Commit](https://github.com/open-webui/open-webui/commit/03cc6ce8eb5c055115406e2304fbf7e3338b8dce)
+- 📁 Archived chats now have their folder associations automatically removed to prevent unintended deletion when their previous folder is deleted. [#14578](https://github.com/open-webui/open-webui/issues/14578)
+- 🔐 ElevenLabs API key is now properly obfuscated in the admin settings page, preventing plain text exposure of sensitive credentials. [#19262](https://github.com/open-webui/open-webui/pull/19262), [#19260](https://github.com/open-webui/open-webui/issues/19260)
+- 🔧 MCP OAuth server metadata discovery now follows the correct specification order, ensuring proper authentication flow compliance. [#19244](https://github.com/open-webui/open-webui/pull/19244)
+- 🔒 API key endpoint restrictions now properly enforce access controls for all endpoints including SCIM, preventing unintended access when "API_KEY_ALLOWED_ENDPOINTS" is configured. [#19168](https://github.com/open-webui/open-webui/issues/19168)
+- 🔓 OAuth role claim parsing now supports both flat and nested claim structures, enabling compatibility with OAuth providers that deliver claims as direct properties on the user object rather than nested structures. [#19286](https://github.com/open-webui/open-webui/pull/19286)
+- 🔑 OAuth MCP server verification now correctly extracts the access token value for authorization headers instead of sending the entire token dictionary. [#19149](https://github.com/open-webui/open-webui/pull/19149), [#19148](https://github.com/open-webui/open-webui/issues/19148)
+- ⚙️ OAuth dynamic client registration now correctly converts empty strings to None for optional fields, preventing validation failures in MCP package integration. [#19144](https://github.com/open-webui/open-webui/pull/19144), [#19129](https://github.com/open-webui/open-webui/issues/19129)
+- 🔐 OIDC authentication now correctly passes client credentials in access token requests, ensuring compatibility with providers that require these parameters per RFC 6749. [#19132](https://github.com/open-webui/open-webui/pull/19132), [#19131](https://github.com/open-webui/open-webui/issues/19131)
+- 🔗 OAuth client creation now respects configured token endpoint authentication methods instead of defaulting to basic authentication, preventing failures with servers that don't support basic auth. [#19165](https://github.com/open-webui/open-webui/pull/19165)
+- 📋 Text copied from chat responses in Chrome now pastes without background formatting, improving readability when pasting into word processors. [#19083](https://github.com/open-webui/open-webui/issues/19083)
+
+### Changed
+
+- 🗄️ Group membership data storage was refactored from JSON arrays to a dedicated relational database table, significantly improving query performance and scalability for instances with large numbers of users and groups, while API responses now return member counts instead of full user ID arrays. [#19239](https://github.com/open-webui/open-webui/pull/19239)
+- 📄 MinerU parameter handling was refactored to pass parameters directly to the API, improving flexibility and fixing VLM backend configuration. [#19105](https://github.com/open-webui/open-webui/pull/19105), [#18446](https://github.com/open-webui/open-webui/discussions/18446)
+- 🔐 API key creation is now controlled by granular user and group permissions, with the "ENABLE_API_KEY" environment variable renamed to "ENABLE_API_KEYS" and disabled by default, requiring explicit configuration at both the global and user permission levels, while related environment variables "ENABLE_API_KEY_ENDPOINT_RESTRICTIONS" and "API_KEY_ALLOWED_ENDPOINTS" were renamed to "ENABLE_API_KEYS_ENDPOINT_RESTRICTIONS" and "API_KEYS_ALLOWED_ENDPOINTS" respectively. [#18336](https://github.com/open-webui/open-webui/pull/18336)
+
+## [0.6.36] - 2025-11-07
+
+### Added
+
+- 🔐 OAuth group parsing now supports configurable separators via the "OAUTH_GROUPS_SEPARATOR" environment variable, enabling proper handling of semicolon-separated group claims from providers like CILogon. [#18987](https://github.com/open-webui/open-webui/pull/18987), [#18979](https://github.com/open-webui/open-webui/issues/18979)
+
+### Fixed
+
+- 🛠️ Tool calling functionality is restored by correcting asynchronous function handling in tool parameter updates. [#18981](https://github.com/open-webui/open-webui/issues/18981)
+- 🖼️ The ComfyUI image edit workflow editor modal now opens correctly when clicking the Edit button. [#18978](https://github.com/open-webui/open-webui/issues/18978)
+- 🔥 Firecrawl import errors are resolved by implementing lazy loading and using the correct class name. [#18973](https://github.com/open-webui/open-webui/issues/18973)
+- 🔌 Socket.IO CORS warning is resolved by properly configuring CORS origins for Socket.IO connections. [Commit](https://github.com/open-webui/open-webui/commit/639d26252e528c9c37a5f553b11eb94376d8792d)
+
+## [0.6.35] - 2025-11-06
+
+### Added
+
+- 🖼️ Image generation system received a comprehensive overhaul with major new capabilities including full image editing support allowing users to modify existing images using text prompts with OpenAI, Gemini, or ComfyUI engines, adding Gemini 2.5 Flash Image (Nano Banana) support, Qwen Image Edit integration, resolution of base64-encoded image display issues, streamlined AUTOMATIC1111 configuration by consolidating parameters into a flexible JSON parameters field, and enhanced UI with a code editor modal for ComfyUI workflow management. [#17434](https://github.com/open-webui/open-webui/pull/17434), [#16976](https://github.com/open-webui/open-webui/issues/16976), [Commit](https://github.com/open-webui/open-webui/commit/8e5690aab4f632a57027e2acf880b8f89a8717c0), [Commit](https://github.com/open-webui/open-webui/commit/72f8539fd2e679fec0762945f22f4b8a6920afa0), [Commit](https://github.com/open-webui/open-webui/commit/8d34fcb586eeee1fac6da2f991518b8a68b00b72), [Commit](https://github.com/open-webui/open-webui/commit/72900cd686de1fa6be84b5a8a2fc857cff7b91b8)
+- 🔒 CORS origin validation was added to WebSocket connections as a defense-in-depth security measure against cross-site WebSocket hijacking attacks. [#18411](https://github.com/open-webui/open-webui/pull/18411), [#18410](https://github.com/open-webui/open-webui/issues/18410)
+- 🔄 Automatic page refresh now occurs when a version update is detected via WebSocket connection, ensuring users always run the latest version without cache issues. [Commit](https://github.com/open-webui/open-webui/commit/989f192c92d2fe55daa31336e7971e21798b96ae)
+- 🐍 Experimental initial preparations for Python 3.13 compatibility by updating dependencies with security enhancements and cryptographic improvements. [#18430](https://github.com/open-webui/open-webui/pull/18430), [#18424](https://github.com/open-webui/open-webui/pull/18424)
+- ⚡ Image compression now preserves the original image format instead of converting to PNG, significantly reducing file sizes and improving chat loading performance. [#18506](https://github.com/open-webui/open-webui/pull/18506)
+- 🎤 Mistral Voxtral model support was added for text-to-speech, including voxtral-small and voxtral-mini models with both transcription and chat completion API support. [#18934](https://github.com/open-webui/open-webui/pull/18934)
+- 🔊 Text-to-speech now uses a global audio queue system to prevent overlapping playback, ensuring only one TTS instance plays at a time with proper stop/start controls and automatic cleanup when switching between messages. [#16152](https://github.com/open-webui/open-webui/pull/16152), [#18744](https://github.com/open-webui/open-webui/pull/18744), [#16150](https://github.com/open-webui/open-webui/issues/16150)
+- 🔊 ELEVENLABS_API_BASE_URL environment variable now allows configuration of custom ElevenLabs API endpoints, enabling support for EU residency API requirements. [#18402](https://github.com/open-webui/open-webui/issues/18402)
+- 🔐 OAUTH_ROLES_SEPARATOR environment variable now allows custom role separators for OAuth roles that contain commas, useful for roles specified in LDAP syntax. [#18572](https://github.com/open-webui/open-webui/pull/18572)
+- 📄 External document loaders can now optionally forward user information headers when ENABLE_FORWARD_USER_INFO_HEADERS is enabled, enabling cost tracking, audit logs, and usage analytics for external services. [#18731](https://github.com/open-webui/open-webui/pull/18731)
+- 📄 MISTRAL_OCR_API_BASE_URL environment variable now allows configuration of custom Mistral OCR API endpoints for flexible deployment options. [Commit](https://github.com/open-webui/open-webui/commit/415b93c7c35c2e2db4425e6da1b88b3750f496b0)
+- ⌨️ Keyboard shortcut hints are now displayed on sidebar buttons with a refactored shortcuts modal that accurately reflects all available hotkeys across different keyboard layouts. [#18473](https://github.com/open-webui/open-webui/pull/18473)
+- 🛠️ Tooltips now display tool descriptions when hovering over tool names on the model edit page, improving usability and providing immediate context. [#18707](https://github.com/open-webui/open-webui/pull/18707)
+- 📝 "Create a new note" from the search modal now immediately creates a new private note and opens it in the editor instead of navigating to the generic notes page. [#18255](https://github.com/open-webui/open-webui/pull/18255)
+- 🖨️ Code block output now preserves whitespace formatting with monospace font to accurately reflect terminal behavior. [#18352](https://github.com/open-webui/open-webui/pull/18352)
+- ✏️ Edit button is now available in the three-dot menu of models in the workspace section for quick access to model editing, with the menu reorganized for better user experience and Edit, Clone, Copy Link, and Share options logically grouped. [#18574](https://github.com/open-webui/open-webui/pull/18574)
+- 📌 Sidebar models section is now collapsible, allowing users to expand and collapse the pinned models list for better sidebar organization. [Commit](https://github.com/open-webui/open-webui/commit/82c08a3b5d189f81c96b6548cc872198771015b0)
+- 🌙 Dark mode styles for select elements were added using Tailwind CSS classes, improving consistency across the interface. [#18636](https://github.com/open-webui/open-webui/pull/18636)
+- 🔄 Various improvements were implemented across the frontend and backend to enhance performance, stability, and security.
+- 🌐 Translations for Portuguese (Brazil), Greek, German, Traditional Chinese, Simplified Chinese, Spanish, Georgian, Danish, and Estonian were enhanced and expanded.
+
+### Fixed
+
+- 🔒 Server-Sent Event (SSE) code injection vulnerability in Direct Connections is resolved by blocking event emission from untrusted external model servers; event emitters from direct connected model servers are no longer supported, preventing arbitrary JavaScript execution in user browsers. [Commit](https://github.com/open-webui/open-webui/commit/8af6a4cf21b756a66cd58378a01c60f74c39b7ca)
+- 🛡️ DOM XSS vulnerability in "Insert Prompt as Rich Text" is resolved by sanitizing HTML content with DOMPurify before rendering. [Commit](https://github.com/open-webui/open-webui/commit/eb9c4c0e358c274aea35f21c2856c0a20051e5f1)
+- ⚙️ MCP server cancellation scope corruption is prevented by reversing disconnection order to follow LIFO and properly handling exceptions, resolving 100% CPU usage when resuming chats with expired tokens or using multiple streamable MCP servers. [#18537](https://github.com/open-webui/open-webui/pull/18537)
+- 🔧 UI freeze when querying models with knowledge bases containing inconsistent distance metrics is resolved by properly initializing the distances array in citations. [#18585](https://github.com/open-webui/open-webui/pull/18585)
+- 🤖 Duplicate model IDs from multiple OpenAI endpoints are now automatically deduplicated server-side, preventing frontend crashes for users with unified gateway proxies that aggregate multiple providers. [Commit](https://github.com/open-webui/open-webui/commit/fdf7ca11d4f3cc8fe63e81c98dc0d1e48e52ba36)
+- 🔐 Login failures with passwords longer than 72 bytes are resolved by safely truncating oversized passwords for bcrypt compatibility. [#18157](https://github.com/open-webui/open-webui/issues/18157)
+- 🔐 OAuth 2.1 MCP tool connections now automatically re-register clients when stored client IDs become stale, preventing unauthorized_client errors after editing tool endpoints and providing detailed error messages for callback failures. [#18415](https://github.com/open-webui/open-webui/pull/18415), [#18309](https://github.com/open-webui/open-webui/issues/18309)
+- 🔓 OAuth 2.1 discovery, metadata fetching, and dynamic client registration now correctly use HTTP proxy environment variables when trust_env is enabled. [Commit](https://github.com/open-webui/open-webui/commit/bafeb76c411483bd6b135f0edbcdce048120f264)
+- 🔌 MCP server connection failures now display clear error messages in the chat interface instead of silently failing. [#18892](https://github.com/open-webui/open-webui/pull/18892), [#18889](https://github.com/open-webui/open-webui/issues/18889)
+- 💬 Chat titles are now properly generated even when title auto-generation is disabled in interface settings, fixing an issue where chats would remain labeled as "New chat". [#18761](https://github.com/open-webui/open-webui/pull/18761), [#18717](https://github.com/open-webui/open-webui/issues/18717), [#6478](https://github.com/open-webui/open-webui/issues/6478)
+- 🔍 Chat query errors are prevented by properly validating and handling the "order_by" parameter to ensure requested columns exist. [#18400](https://github.com/open-webui/open-webui/pull/18400), [#18452](https://github.com/open-webui/open-webui/pull/18452)
+- 🔧 Root-level max_tokens parameter is no longer dropped when proxying to Ollama, properly converting to num_predict to limit output token length as intended. [#18618](https://github.com/open-webui/open-webui/issues/18618)
+- 🔑 Self-hosted Marker instances can now be used without requiring an API key, while keeping it optional for datalab Marker service users. [#18617](https://github.com/open-webui/open-webui/issues/18617)
+- 🔧 OpenAPI specification endpoint conflict between "/api/v1/models" and "/api/v1/models/" is resolved by changing the models router endpoint to "/list", preventing duplicate operationId errors when generating TypeScript API clients. [#18758](https://github.com/open-webui/open-webui/issues/18758)
+- 🏷️ Model tags are now de-duplicated case-insensitively in both the model selector and workspace models page, preventing duplicate entries with different capitalization from appearing in filter dropdowns. [#18716](https://github.com/open-webui/open-webui/pull/18716), [#18711](https://github.com/open-webui/open-webui/issues/18711)
+- 📄 Docling RAG parameter configuration is now correctly saved in the admin UI by fixing the typo in the "DOCLING_PARAMS" parameter name. [#18390](https://github.com/open-webui/open-webui/pull/18390)
+- 📃 Tika document processing now automatically detects content types instead of relying on potentially incorrect browser-provided mime-types, improving file handling accuracy for formats like RTF. [#18765](https://github.com/open-webui/open-webui/pull/18765), [#18683](https://github.com/open-webui/open-webui/issues/18683)
+- 🖼️ Image and video uploads to knowledge bases now display proper error messages instead of showing an infinite spinner when the content extraction engine does not support these file types. [#18514](https://github.com/open-webui/open-webui/issues/18514)
+- 📝 Notes PDF export now properly detects and applies dark mode styling consistently across both the notes list and individual note pages, with a shared utility function to eliminate code duplication. [#18526](https://github.com/open-webui/open-webui/issues/18526)
+- 💭 Details tags for reasoning content are now correctly identified and rendered even when the same tag is present in user messages. [#18840](https://github.com/open-webui/open-webui/pull/18840), [#18294](https://github.com/open-webui/open-webui/issues/18294)
+- 📊 Mermaid and Vega rendering errors now display inline with the code instead of showing repetitive toast notifications, improving user experience when models generate invalid diagram syntax. [Commit](https://github.com/open-webui/open-webui/commit/fdc0f04a8b7dd0bc9f9dc0e7e30854f7a0eea3e9)
+- 📈 Mermaid diagram rendering errors no longer cause UI unavailability or display error messages below the input box. [#18493](https://github.com/open-webui/open-webui/pull/18493), [#18340](https://github.com/open-webui/open-webui/issues/18340)
+- 🔗 Web search SSL verification is now asynchronous, preventing the website from hanging during web search operations. [#18714](https://github.com/open-webui/open-webui/pull/18714), [#18699](https://github.com/open-webui/open-webui/issues/18699)
+- 🌍 Web search results now correctly use HTTP proxy environment variables when WEB_SEARCH_TRUST_ENV is enabled. [#18667](https://github.com/open-webui/open-webui/pull/18667), [#7008](https://github.com/open-webui/open-webui/discussions/7008)
+- 🔍 Google Programmable Search Engine now properly includes referer headers, enabling API keys with HTTP referrer restrictions configured in Google Cloud Console. [#18871](https://github.com/open-webui/open-webui/pull/18871), [#18870](https://github.com/open-webui/open-webui/issues/18870)
+- ⚡ YouTube video transcript fetching now works correctly when using a proxy connection. [#18419](https://github.com/open-webui/open-webui/pull/18419)
+- 🎙️ Speech-to-text transcription no longer deletes or replaces existing text in the prompt input field, properly preserving any previously entered content. [#18540](https://github.com/open-webui/open-webui/issues/18540)
+- 🎙️ The "Instant Auto-Send After Voice Transcription" setting now functions correctly and automatically sends transcribed text when enabled. [#18466](https://github.com/open-webui/open-webui/issues/18466)
+- ⚙️ Chat settings now load properly when reopening a tab or starting a new session by initializing defaults when sessionStorage is empty. [#18438](https://github.com/open-webui/open-webui/pull/18438)
+- 🔎 Folder tag search in the sidebar now correctly handles folder names with multiple spaces by replacing all spaces with underscores. [Commit](https://github.com/open-webui/open-webui/commit/a8fe979af68e47e4e4bb3eb76e48d93d60cd2a45)
+- 🛠️ Functions page now updates immediately after deleting a function, removing the need for a manual page reload. [#18912](https://github.com/open-webui/open-webui/pull/18912), [#18908](https://github.com/open-webui/open-webui/issues/18908)
+- 🛠️ Native tool calling now properly supports sequential tool calls with shared context, allowing tools to access images and data from previous tool executions in the same conversation. [#18664](https://github.com/open-webui/open-webui/pull/18664)
+- 🎯 Globally enabled actions in the model editor now correctly apply as global instead of being treated as disabled. [#18577](https://github.com/open-webui/open-webui/pull/18577)
+- 📋 Clipboard images pasted via the "{{CLIPBOARD}}" prompt variable are now correctly converted to base64 format before being sent to the backend, resolving base64 encoding errors. [#18432](https://github.com/open-webui/open-webui/pull/18432), [#18425](https://github.com/open-webui/open-webui/issues/18425)
+- 📋 File list is now cleared when switching to models that do not support file uploads, preventing files from being sent to incompatible models. [#18496](https://github.com/open-webui/open-webui/pull/18496)
+- 📂 Move menu no longer displays when folders are empty. [#18484](https://github.com/open-webui/open-webui/pull/18484)
+- 📁 Folder and channel creation now validates that names are not empty, preventing creation of folders or channels with no name and showing an error toast if attempted. [#18564](https://github.com/open-webui/open-webui/pull/18564)
+- 🖊️ Rich text input no longer removes text between equals signs when pasting code with comparison operators. [#18551](https://github.com/open-webui/open-webui/issues/18551)
+- ⌨️ Keyboard shortcuts now display the correct keys for international and non-QWERTY keyboard layouts by detecting the user's layout using the Keyboard API. [#18533](https://github.com/open-webui/open-webui/pull/18533)
+- 🌐 "Attach Webpage" button now displays with correct disabled styling when a model does not support file uploads. [#18483](https://github.com/open-webui/open-webui/pull/18483)
+- 🎚️ Divider no longer displays in the integrations menu when no integrations are enabled. [#18487](https://github.com/open-webui/open-webui/pull/18487)
+- 📱 Chat controls button is now properly hidden on mobile for users without admin or explicit chat control permissions. [#18641](https://github.com/open-webui/open-webui/pull/18641)
+- 📍 User menu, download submenu, and move submenu are now repositioned to prevent overlap with the Chat Controls sidebar when it is open. [Commit](https://github.com/open-webui/open-webui/commit/414ab51cb6df1ab0d6c85ac6c1f2c5c9a5f8e2aa)
+- 🎯 Artifacts button no longer appears in the chat menu when there are no artifacts to display. [Commit](https://github.com/open-webui/open-webui/commit/ed6449d35f84f68dc75ee5c6b3f4748a3fda0096)
+- 🎨 Artifacts view now automatically displays when opening an existing conversation containing artifacts, improving user experience. [#18215](https://github.com/open-webui/open-webui/pull/18215)
+- 🖌️ Formatting toolbar is no longer hidden under images or code blocks in chat and now displays correctly above all message content.
+- 🎨 Layout shift near system instructions is prevented by properly rendering the chat component when system prompts are empty. [#18594](https://github.com/open-webui/open-webui/pull/18594)
+- 📐 Modal layout shift caused by scrollbar appearance is prevented by adding a stable scrollbar gutter. [#18591](https://github.com/open-webui/open-webui/pull/18591)
+- ✨ Spacing between icon and label in the user menu dropdown items is now consistent. [#18595](https://github.com/open-webui/open-webui/pull/18595)
+- 💬 Duplicate prompt suggestions no longer cause the webpage to freeze or throw JavaScript errors by implementing proper key management with composite keys. [#18841](https://github.com/open-webui/open-webui/pull/18841), [#18566](https://github.com/open-webui/open-webui/issues/18566)
+- 🔍 Chat preview loading in the search modal now works correctly for all search results by fixing an index boundary check that previously caused out-of-bounds errors. [#18911](https://github.com/open-webui/open-webui/pull/18911)
+- ♿ Screen reader support was enhanced by wrapping messages in semantic elements with descriptive aria-labels, adding "Assistant is typing" and "Response complete" announcements for improved accessibility. [#18735](https://github.com/open-webui/open-webui/pull/18735)
+- 🔒 Incorrect await call in the OAuth 2.1 flow is removed, eliminating a logged exception during authentication. [#18236](https://github.com/open-webui/open-webui/pull/18236)
+- 🛡️ Duplicate crossorigin attribute in the manifest file was removed. [#18413](https://github.com/open-webui/open-webui/pull/18413)
+
+### Changed
+
+- 🔄 Firecrawl integration was refactored to use the official Firecrawl SDK instead of direct HTTP requests and langchain_community FireCrawlLoader, improving reliability and performance with batch scraping support and enhanced error handling. [#18635](https://github.com/open-webui/open-webui/pull/18635)
+- 📄 MinerU content extraction engine now only supports PDF files following the upstream removal of LibreOffice document conversion in version 2.0.0; users needing to process office documents should convert them to PDF format first. [#18448](https://github.com/open-webui/open-webui/issues/18448)
+
+## [0.6.34] - 2025-10-16
+
+### Added
+
+- 📄 MinerU is now supported as a document parser backend, with support for both local and managed API deployments. [#18306](https://github.com/open-webui/open-webui/pull/18306)
+- 🔒 JWT token expiration default is now set to 4 weeks instead of never expiring, with security warnings displayed in backend logs and admin UI when set to unlimited. [#18261](https://github.com/open-webui/open-webui/pull/18261), [#18262](https://github.com/open-webui/open-webui/pull/18262)
+- ⚡ Page loading performance is improved by preventing unnecessary API requests when sidebar folders are not expanded. [#18179](https://github.com/open-webui/open-webui/pull/18179), [#17476](https://github.com/open-webui/open-webui/issues/17476)
+- 📁 File hash values are now included in the knowledge endpoint response, enabling efficient file synchronization through hash comparison. [#18284](https://github.com/open-webui/open-webui/pull/18284), [#18283](https://github.com/open-webui/open-webui/issues/18283)
+- 🎨 Chat dialog scrollbar visibility is improved by increasing its width, making it easier to use for navigation. [#18369](https://github.com/open-webui/open-webui/pull/18369), [#11782](https://github.com/open-webui/open-webui/issues/11782)
+- 🔄 Various improvements were implemented across the frontend and backend to enhance performance, stability, and security.
+- 🌐 Translations for Catalan, Chinese, Czech, Finnish, German, Kabyle, Korean, Portuguese (Brazil), Spanish, Thai, and Turkish were enhanced and expanded.
+
+### Fixed
+
+- 📚 Focused retrieval mode now works correctly, preventing the system from forcing full context mode and loading all documents in a knowledge base regardless of settings. [#18133](https://github.com/open-webui/open-webui/issues/18133)
+- 🔧 Filter inlet functions now correctly execute on tool call continuations, ensuring parameter persistence throughout tool interactions. [#18222](https://github.com/open-webui/open-webui/issues/18222)
+- 🛠️ External tool servers now properly support DELETE requests with body data. [#18289](https://github.com/open-webui/open-webui/pull/18289), [#18287](https://github.com/open-webui/open-webui/issues/18287)
+- 🗄️ Oracle23ai vector database client now correctly handles variable initialization, resolving UnboundLocalError when retrieving items from collections. [#18356](https://github.com/open-webui/open-webui/issues/18356)
+- 🔧 Model auto-pull functionality now works correctly even when user settings remain unmodified. [#18324](https://github.com/open-webui/open-webui/pull/18324)
+- 🎨 Duplicate HTML content in artifacts is now prevented by improving code block detection logic. [#18195](https://github.com/open-webui/open-webui/pull/18195), [#6154](https://github.com/open-webui/open-webui/issues/6154)
+- 💬 Pinned chats now appear in the Reference Chats list and can be referenced in conversations. [#18288](https://github.com/open-webui/open-webui/issues/18288)
+- 📝 Misleading knowledge base warning text in documents settings is clarified to correctly instruct users about reindexing vectors. [#18263](https://github.com/open-webui/open-webui/pull/18263)
+- 🔔 Toast notifications can now be dismissed even when a modal is open. [#18260](https://github.com/open-webui/open-webui/pull/18260)
+- 🔘 The "Chats" button in the sidebar now correctly toggles chat list visibility without navigating away from the current page. [#18232](https://github.com/open-webui/open-webui/pull/18232)
+- 🎯 The Integrations menu no longer closes prematurely when clicking outside the Valves modal. [#18310](https://github.com/open-webui/open-webui/pull/18310)
+- 🛠️ Tool ID display issues where "undefined" was incorrectly shown in the interface are now resolved. [#18178](https://github.com/open-webui/open-webui/pull/18178)
+- 🛠️ Model management issues caused by excessively long model IDs are now prevented through validation that limits model IDs to 256 characters. [#18125](https://github.com/open-webui/open-webui/issues/18125)
+
 ## [0.6.33] - 2025-10-08
 
 ### Added

Dockerfile

@@ -13,6 +13,7 @@ ARG USE_CUDA_VER=cu128
 # IMPORTANT: If you change the embedding model (sentence-transformers/all-MiniLM-L6-v2) and vice versa, you aren't able to use RAG Chat with your previous documents loaded in the WebUI! You need to re-embed them.
 ARG USE_EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2
 ARG USE_RERANKING_MODEL=""
+ARG USE_AUXILIARY_EMBEDDING_MODEL=TaylorAI/bge-micro-v2
 
 # Tiktoken encoding name; models to use can be found at https://huggingface.co/models?library=tiktoken
 ARG USE_TIKTOKEN_ENCODING_NAME="cl100k_base"
@@ -42,7 +43,7 @@ ENV APP_BUILD_HASH=${BUILD_HASH}
 RUN npm run build
 
 ######## WebUI backend ########
-FROM python:3.11-slim-bookworm AS base
+FROM python:3.11.14-slim-bookworm AS base
 
 # Use args
 ARG USE_CUDA
@@ -52,9 +53,13 @@ ARG USE_SLIM
 ARG USE_PERMISSION_HARDENING
 ARG USE_EMBEDDING_MODEL
 ARG USE_RERANKING_MODEL
+ARG USE_AUXILIARY_EMBEDDING_MODEL
 ARG UID
 ARG GID
 
+# Python settings
+ENV PYTHONUNBUFFERED=1
+
 ## Basis ##
 ENV ENV=prod \
     PORT=8080 \
@@ -64,7 +69,8 @@ ENV ENV=prod \
     USE_SLIM_DOCKER=${USE_SLIM} \
     USE_CUDA_DOCKER_VER=${USE_CUDA_VER} \
     USE_EMBEDDING_MODEL_DOCKER=${USE_EMBEDDING_MODEL} \
-    USE_RERANKING_MODEL_DOCKER=${USE_RERANKING_MODEL}
+    USE_RERANKING_MODEL_DOCKER=${USE_RERANKING_MODEL} \
+    USE_AUXILIARY_EMBEDDING_MODEL_DOCKER=${USE_AUXILIARY_EMBEDDING_MODEL}
 
 ## Basis URL Config ##
 ENV OLLAMA_BASE_URL="/ollama" \
@@ -85,6 +91,7 @@ ENV WHISPER_MODEL="base" \
 ## RAG Embedding model settings ##
 ENV RAG_EMBEDDING_MODEL="$USE_EMBEDDING_MODEL_DOCKER" \
     RAG_RERANKING_MODEL="$USE_RERANKING_MODEL_DOCKER" \
+    AUXILIARY_EMBEDDING_MODEL="$USE_AUXILIARY_EMBEDDING_MODEL_DOCKER" \
     SENTENCE_TRANSFORMERS_HOME="/app/backend/data/cache/embedding/models"
 
 ## Tiktoken model settings ##
@@ -121,7 +128,7 @@ RUN apt-get update && \
     apt-get install -y --no-install-recommends \
     git build-essential pandoc gcc netcat-openbsd curl jq \
     python3-dev \
-    ffmpeg libsm6 libxext6 \
+    ffmpeg libsm6 libxext6 zstd \
     && rm -rf /var/lib/apt/lists/*
 
 # install python dependencies
@@ -130,18 +137,23 @@ COPY --chown=$UID:$GID ./backend/requirements.txt ./requirements.txt
 RUN pip3 install --no-cache-dir uv && \
     if [ "$USE_CUDA" = "true" ]; then \
     # If you use CUDA the whisper and embedding model will be downloaded on first use
-    pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/$USE_CUDA_DOCKER_VER --no-cache-dir && \
+    # fix: pin torch<=2.9.1 - torch 2.10.0 aarch64 wheels cause SIGILL on ARM devices (RPi 4 Cortex-A72) #21349
+    pip3 install 'torch<=2.9.1' torchvision torchaudio --index-url https://download.pytorch.org/whl/$USE_CUDA_DOCKER_VER --no-cache-dir && \
     uv pip install --system -r requirements.txt --no-cache-dir && \
     python -c "import os; from sentence_transformers import SentenceTransformer; SentenceTransformer(os.environ['RAG_EMBEDDING_MODEL'], device='cpu')" && \
+    python -c "import os; from sentence_transformers import SentenceTransformer; SentenceTransformer(os.environ.get('AUXILIARY_EMBEDDING_MODEL', 'TaylorAI/bge-micro-v2'), device='cpu')" && \
     python -c "import os; from faster_whisper import WhisperModel; WhisperModel(os.environ['WHISPER_MODEL'], device='cpu', compute_type='int8', download_root=os.environ['WHISPER_MODEL_DIR'])"; \
     python -c "import os; import tiktoken; tiktoken.get_encoding(os.environ['TIKTOKEN_ENCODING_NAME'])"; \
+    python -c "import nltk; nltk.download('punkt_tab')"; \
     else \
-    pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu --no-cache-dir && \
+    pip3 install 'torch<=2.9.1' torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu --no-cache-dir && \
     uv pip install --system -r requirements.txt --no-cache-dir && \
     if [ "$USE_SLIM" != "true" ]; then \
     python -c "import os; from sentence_transformers import SentenceTransformer; SentenceTransformer(os.environ['RAG_EMBEDDING_MODEL'], device='cpu')" && \
+    python -c "import os; from sentence_transformers import SentenceTransformer; SentenceTransformer(os.environ.get('AUXILIARY_EMBEDDING_MODEL', 'TaylorAI/bge-micro-v2'), device='cpu')" && \
     python -c "import os; from faster_whisper import WhisperModel; WhisperModel(os.environ['WHISPER_MODEL'], device='cpu', compute_type='int8', download_root=os.environ['WHISPER_MODEL_DIR'])"; \
     python -c "import os; import tiktoken; tiktoken.get_encoding(os.environ['TIKTOKEN_ENCODING_NAME'])"; \
+    python -c "import nltk; nltk.download('punkt_tab')"; \
     fi; \
     fi; \
     mkdir -p /app/backend/data && chown -R $UID:$GID /app/backend/data/ && \

INSTALLATION.md

@@ -1,35 +0,0 @@
-### Installing Both Ollama and Open WebUI Using Kustomize
-
-For cpu-only pod
-
-```bash
-kubectl apply -f ./kubernetes/manifest/base
-```
-
-For gpu-enabled pod
-
-```bash
-kubectl apply -k ./kubernetes/manifest
-```
-
-### Installing Both Ollama and Open WebUI Using Helm
-
-Package Helm file first
-
-```bash
-helm package ./kubernetes/helm/
-```
-
-For cpu-only pod
-
-```bash
-helm install ollama-webui ./ollama-webui-*.tgz
-```
-
-For gpu-enabled pod
-
-```bash
-helm install ollama-webui ./ollama-webui-*.tgz --set ollama.resources.limits.nvidia.com/gpu="1"
-```
-
-Check the `kubernetes/helm/values.yaml` file to know which parameters are available for customization

LICENSE

@@ -1,4 +1,4 @@
-Copyright (c) 2023-2025 Timothy Jaeryang Baek (Open WebUI)
+Copyright (c) 2023- Open WebUI Inc. [Created by Timothy Jaeryang Baek]
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without

README.md

@@ -10,14 +10,16 @@
 [![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)
 
+![Open WebUI Banner](./banner.png)
+
 **Open WebUI is an [extensible](https://docs.openwebui.com/features/plugin/), feature-rich, and user-friendly self-hosted AI platform designed to operate entirely offline.** It supports various LLM runners like **Ollama** and **OpenAI-compatible APIs**, with **built-in inference engine** for RAG, making it a **powerful AI deployment solution**.
 
 Passionate about open-source AI? [Join our team →](https://careers.openwebui.com/)
 
-![Open WebUI Demo](./demo.gif)
+![Open WebUI Demo](./demo.png)
 
 > [!TIP]  
-> **Looking for an [Enterprise Plan](https://docs.openwebui.com/enterprise)?** – **[Speak with Our Sales Team Today!](mailto:sales@openwebui.com)**
+> **Looking for an [Enterprise Plan](https://docs.openwebui.com/enterprise)?** – **[Speak with Our Sales Team Today!](https://docs.openwebui.com/enterprise)**
 >
 > Get **enhanced capabilities**, including **custom theming and branding**, **Service Level Agreement (SLA) support**, **Long-Term Support (LTS) versions**, and **more!**
 
@@ -31,32 +33,44 @@ For more information, be sure to check out our [Open WebUI Documentation](https:
 
 - 🛡️ **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.
 
-- 🔄 **SCIM 2.0 Support**: 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.
-
 - 📱 **Responsive Design**: Enjoy a seamless experience across Desktop PC, Laptop, and Mobile devices.
 
 - 📱 **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.
 
 - ✒️🔢 **Full Markdown and LaTeX Support**: Elevate your LLM experience with comprehensive Markdown and LaTeX capabilities for enriched interaction.
 
-- 🎤📹 **Hands-Free Voice/Video Call**: Experience seamless communication with integrated hands-free voice and video call features, allowing for a more dynamic and interactive chat environment.
+- 🎤📹 **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.
 
 - 🛠️ **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.
 
 - 🐍 **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.
 
-- 📚 **Local RAG Integration**: Dive into the future of chat interactions with groundbreaking Retrieval Augmented Generation (RAG) support. This feature seamlessly integrates document interactions into your chat experience. You can load documents directly into the chat or add files to your document library, effortlessly accessing them using the `#` command before a query.
+- 💾 **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.
 
-- 🔍 **Web Search for RAG**: Perform web searches using providers like `SearXNG`, `Google PSE`, `Brave Search`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `TavilySearch`, `SearchApi` and `Bing` and inject the results directly into your chat experience.
+- 📚 **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, External loaders). Load documents directly into chat or add files to your document library, effortlessly accessing them using the `#` command before a query.
+
+- 🔍 **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.
 
 - 🌐 **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.
 
-- 🎨 **Image Generation Integration**: Seamlessly incorporate image generation capabilities using options such as AUTOMATIC1111 API or ComfyUI (local), and OpenAI's DALL-E (external), enriching your chat experience with dynamic visual content.
+- 🎨 **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.
 
 - ⚙️ **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.
 
 - 🔐 **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.
 
+- 🗄️ **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.
+
+- 🔍 **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.
+
+- 🔐 **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.
+
+- ☁️ **Cloud-Native Integration**: Native support for Google Drive and OneDrive/SharePoint file picking, enabling seamless document import from enterprise cloud storage.
+
+- 📊 **Production Observability**: Built-in OpenTelemetry support for traces, metrics, and logs, enabling comprehensive monitoring with your existing observability stack.
+
+- ⚖️ **Horizontal Scalability**: Redis-backed session management and WebSocket support for multi-worker and multi-node deployments behind load balancers.
+
 - 🌐🌍 **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!
 
 - 🧩 **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.
@@ -65,43 +79,6 @@ For more information, be sure to check out our [Open WebUI Documentation](https:
 
 Want to learn more about Open WebUI's features? Check out our [Open WebUI documentation](https://docs.openwebui.com/features) for a comprehensive overview!
 
-## Sponsors 🙌
-
-#### Emerald
-
-<table>
-  <!-- <tr>
-    <td>
-      <a href="https://n8n.io/" target="_blank">
-        <img src="https://docs.openwebui.com/sponsors/logos/n8n.png" alt="n8n" style="width: 8rem; height: 8rem; border-radius: .75rem;" />
-      </a>
-    </td>
-    <td>
-      <a href="https://n8n.io/">n8n</a> • Does your interface have a backend yet?<br>Try <a href="https://n8n.io/">n8n</a>
-    </td>
-  </tr> -->
-  <tr>
-    <td>
-      <a href="https://tailscale.com/blog/self-host-a-local-ai-stack/?utm_source=OpenWebUI&utm_medium=paid-ad-placement&utm_campaign=OpenWebUI-Docs" target="_blank">
-        <img src="https://docs.openwebui.com/sponsors/logos/tailscale.png" alt="Tailscale" style="width: 8rem; height: 8rem; border-radius: .75rem;" />
-      </a>
-    </td>
-    <td>
-      <a href="https://tailscale.com/blog/self-host-a-local-ai-stack/?utm_source=OpenWebUI&utm_medium=paid-ad-placement&utm_campaign=OpenWebUI-Docs">Tailscale</a> • Connect self-hosted AI to any device with Tailscale
-    </td>
-  </tr>
-   <tr>
-    <td>
-      <a href="https://warp.dev/open-webui" target="_blank">
-        <img src="https://docs.openwebui.com/sponsors/logos/warp.png" alt="Warp" style="width: 8rem; height: 8rem; border-radius: .75rem;" />
-      </a>
-    </td>
-    <td>
-      <a href="https://warp.dev/open-webui">Warp</a> • The intelligent terminal for developers
-    </td>
-  </tr>
-</table>
-
 ---
 
 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!
@@ -213,14 +190,6 @@ docker run -d --network=host -v open-webui:/app/backend/data -e OLLAMA_BASE_URL=
 
 ### Keeping Your Docker Installation Up-to-Date
 
-In case you want to update your local Docker installation to the latest version, you can do it with [Watchtower](https://containrrr.dev/watchtower/):
-
-```bash
-docker run --rm --volume /var/run/docker.sock:/var/run/docker.sock containrrr/watchtower --run-once open-webui
-```
-
-In the last part of the command, replace `open-webui` with your container name if it is different.
-
 Check our Updating Guide available in our [Open WebUI Documentation](https://docs.openwebui.com/getting-started/updating).
 
 ### Using the Dev Branch 🌙

backend/open_webui/constants.py

@@ -38,13 +38,14 @@ class ERROR_MESSAGES(str, Enum):
     ID_TAKEN = "Uh-oh! This id is already registered. Please choose another id string."
     MODEL_ID_TAKEN = "Uh-oh! This model id is already registered. Please choose another model id string."
     NAME_TAG_TAKEN = "Uh-oh! This name tag is already registered. Please choose another name tag string."
+    MODEL_ID_TOO_LONG = "The model id is too long. Please make sure your model id is less than 256 characters long."
 
     INVALID_TOKEN = (
         "Your session has expired or the token is invalid. Please sign in again."
     )
     INVALID_CRED = "The email or password provided is incorrect. Please check for typos and try logging in again."
     INVALID_EMAIL_FORMAT = "The email format you entered is invalid. Please double-check and make sure you're using a valid email address (e.g., yourname@example.com)."
-    INVALID_PASSWORD = (
+    INCORRECT_PASSWORD = (
         "The password provided is incorrect. Please check for typos and try again."
     )
     INVALID_TRUSTED_HEADER = "Your provider has not provided a trusted header. Please contact your administrator for assistance."
@@ -104,6 +105,10 @@ class ERROR_MESSAGES(str, Enum):
     )
     FILE_NOT_PROCESSED = "Extracted content is not available for this file. Please ensure that the file is processed before proceeding."
 
+    INVALID_PASSWORD = lambda err="": (
+        err if err else "The password does not meet the required validation criteria."
+    )
+
 
 class TASKS(str, Enum):
     def __str__(self) -> str:

backend/open_webui/env.py

@@ -8,6 +8,8 @@ import shutil
 from uuid import uuid4
 from pathlib import Path
 from cryptography.hazmat.primitives import serialization
+import re
+
 
 import markdown
 from bs4 import BeautifulSoup
@@ -83,32 +85,7 @@ if "cuda_error" in locals():
     log.exception(cuda_error)
     del cuda_error
 
-log_sources = [
-    "AUDIO",
-    "COMFYUI",
-    "CONFIG",
-    "DB",
-    "IMAGES",
-    "MAIN",
-    "MODELS",
-    "OLLAMA",
-    "OPENAI",
-    "RAG",
-    "WEBHOOK",
-    "SOCKET",
-    "OAUTH",
-]
-
-SRC_LOG_LEVELS = {}
-
-for source in log_sources:
-    log_env_var = source + "_LOG_LEVEL"
-    SRC_LOG_LEVELS[source] = os.environ.get(log_env_var, "").upper()
-    if SRC_LOG_LEVELS[source] not in logging.getLevelNamesMapping():
-        SRC_LOG_LEVELS[source] = GLOBAL_LOG_LEVEL
-    log.info(f"{log_env_var}: {SRC_LOG_LEVELS[source]}")
-
-log.setLevel(SRC_LOG_LEVELS["CONFIG"])
+SRC_LOG_LEVELS = {}  # Legacy variable, do not remove
 
 WEBUI_NAME = os.environ.get("WEBUI_NAME", "Open WebUI")
 if WEBUI_NAME != "Open WebUI":
@@ -135,8 +112,13 @@ else:
         PACKAGE_DATA = {"version": "0.0.0"}
 
 VERSION = PACKAGE_DATA["version"]
+
+
+DEPLOYMENT_ID = os.environ.get("DEPLOYMENT_ID", "")
 INSTANCE_ID = os.environ.get("INSTANCE_ID", str(uuid4()))
 
+ENABLE_DB_MIGRATIONS = os.environ.get("ENABLE_DB_MIGRATIONS", "True").lower() == "true"
+
 
 # Function to parse each section
 def parse_section(section):
@@ -212,11 +194,35 @@ ENABLE_FORWARD_USER_INFO_HEADERS = (
     os.environ.get("ENABLE_FORWARD_USER_INFO_HEADERS", "False").lower() == "true"
 )
 
+# Header names for user info forwarding (customizable via environment variables)
+FORWARD_USER_INFO_HEADER_USER_NAME = os.environ.get(
+    "FORWARD_USER_INFO_HEADER_USER_NAME", "X-OpenWebUI-User-Name"
+)
+FORWARD_USER_INFO_HEADER_USER_ID = os.environ.get(
+    "FORWARD_USER_INFO_HEADER_USER_ID", "X-OpenWebUI-User-Id"
+)
+FORWARD_USER_INFO_HEADER_USER_EMAIL = os.environ.get(
+    "FORWARD_USER_INFO_HEADER_USER_EMAIL", "X-OpenWebUI-User-Email"
+)
+FORWARD_USER_INFO_HEADER_USER_ROLE = os.environ.get(
+    "FORWARD_USER_INFO_HEADER_USER_ROLE", "X-OpenWebUI-User-Role"
+)
+
+# Header name for chat ID forwarding (customizable via environment variable)
+FORWARD_SESSION_INFO_HEADER_MESSAGE_ID = os.environ.get(
+    "FORWARD_SESSION_INFO_HEADER_MESSAGE_ID", "X-OpenWebUI-Message-Id"
+)
+FORWARD_SESSION_INFO_HEADER_CHAT_ID = os.environ.get(
+    "FORWARD_SESSION_INFO_HEADER_CHAT_ID", "X-OpenWebUI-Chat-Id"
+)
+
 # Experimental feature, may be removed in future
 ENABLE_STAR_SESSIONS_MIDDLEWARE = (
     os.environ.get("ENABLE_STAR_SESSIONS_MIDDLEWARE", "False").lower() == "true"
 )
 
+ENABLE_EASTER_EGGS = os.environ.get("ENABLE_EASTER_EGGS", "True").lower() == "true"
+
 ####################################
 # WEBUI_BUILD_HASH
 ####################################
@@ -359,6 +365,16 @@ if DATABASE_USER_ACTIVE_STATUS_UPDATE_INTERVAL is not None:
     except Exception:
         DATABASE_USER_ACTIVE_STATUS_UPDATE_INTERVAL = 0.0
 
+# When enabled, get_db_context reuses existing sessions; set to False to always create new sessions
+DATABASE_ENABLE_SESSION_SHARING = (
+    os.environ.get("DATABASE_ENABLE_SESSION_SHARING", "False").lower() == "true"
+)
+
+# Enable public visibility of active user count (when disabled, only admins can see it)
+ENABLE_PUBLIC_ACTIVE_USERS_COUNT = (
+    os.environ.get("ENABLE_PUBLIC_ACTIVE_USERS_COUNT", "True").lower() == "true"
+)
+
 RESET_CONFIG_ON_START = (
     os.environ.get("RESET_CONFIG_ON_START", "False").lower() == "true"
 )
@@ -369,6 +385,8 @@ ENABLE_REALTIME_CHAT_SAVE = (
 
 ENABLE_QUERIES_CACHE = os.environ.get("ENABLE_QUERIES_CACHE", "False").lower() == "true"
 
+RAG_SYSTEM_CONTEXT = os.environ.get("RAG_SYSTEM_CONTEXT", "False").lower() == "true"
+
 ####################################
 # REDIS
 ####################################
@@ -390,6 +408,25 @@ try:
 except ValueError:
     REDIS_SENTINEL_MAX_RETRY_COUNT = 2
 
+
+REDIS_SOCKET_CONNECT_TIMEOUT = os.environ.get("REDIS_SOCKET_CONNECT_TIMEOUT", "")
+try:
+    REDIS_SOCKET_CONNECT_TIMEOUT = float(REDIS_SOCKET_CONNECT_TIMEOUT)
+except ValueError:
+    REDIS_SOCKET_CONNECT_TIMEOUT = None
+
+REDIS_RECONNECT_DELAY = os.environ.get("REDIS_RECONNECT_DELAY", "")
+
+if REDIS_RECONNECT_DELAY == "":
+    REDIS_RECONNECT_DELAY = None
+else:
+    try:
+        REDIS_RECONNECT_DELAY = float(REDIS_RECONNECT_DELAY)
+        if REDIS_RECONNECT_DELAY < 0:
+            REDIS_RECONNECT_DELAY = None
+    except Exception:
+        REDIS_RECONNECT_DELAY = None
+
 ####################################
 # UVICORN WORKERS
 ####################################
@@ -417,6 +454,16 @@ ENABLE_SIGNUP_PASSWORD_CONFIRMATION = (
     os.environ.get("ENABLE_SIGNUP_PASSWORD_CONFIRMATION", "False").lower() == "true"
 )
 
+####################################
+# Admin Account Runtime Creation
+####################################
+
+# Optional env vars for creating an admin account on startup
+# Useful for headless/automated deployments
+WEBUI_ADMIN_EMAIL = os.environ.get("WEBUI_ADMIN_EMAIL", "")
+WEBUI_ADMIN_PASSWORD = os.environ.get("WEBUI_ADMIN_PASSWORD", "")
+WEBUI_ADMIN_NAME = os.environ.get("WEBUI_ADMIN_NAME", "Admin")
+
 WEBUI_AUTH_TRUSTED_EMAIL_HEADER = os.environ.get(
     "WEBUI_AUTH_TRUSTED_EMAIL_HEADER", None
 )
@@ -426,6 +473,27 @@ WEBUI_AUTH_TRUSTED_GROUPS_HEADER = os.environ.get(
 )
 
 
+ENABLE_PASSWORD_VALIDATION = (
+    os.environ.get("ENABLE_PASSWORD_VALIDATION", "False").lower() == "true"
+)
+PASSWORD_VALIDATION_REGEX_PATTERN = os.environ.get(
+    "PASSWORD_VALIDATION_REGEX_PATTERN",
+    r"^(?=.*[a-z])(?=.*[A-Z])(?=.*\d)(?=.*[^\w\s]).{8,}$",
+)
+
+
+try:
+    PASSWORD_VALIDATION_REGEX_PATTERN = rf"{PASSWORD_VALIDATION_REGEX_PATTERN}"
+    PASSWORD_VALIDATION_REGEX_PATTERN = re.compile(PASSWORD_VALIDATION_REGEX_PATTERN)
+except Exception as e:
+    log.error(f"Invalid PASSWORD_VALIDATION_REGEX_PATTERN: {e}")
+    PASSWORD_VALIDATION_REGEX_PATTERN = re.compile(
+        r"^(?=.*[a-z])(?=.*[A-Z])(?=.*\d)(?=.*[^\w\s]).{8,}$"
+    )
+
+PASSWORD_VALIDATION_HINT = os.environ.get("PASSWORD_VALIDATION_HINT", "")
+
+
 BYPASS_MODEL_ACCESS_CONTROL = (
     os.environ.get("BYPASS_MODEL_ACCESS_CONTROL", "False").lower() == "true"
 )
@@ -489,12 +557,29 @@ OAUTH_SESSION_TOKEN_ENCRYPTION_KEY = os.environ.get(
     "OAUTH_SESSION_TOKEN_ENCRYPTION_KEY", WEBUI_SECRET_KEY
 )
 
+# Token Exchange Configuration
+# Allows external apps to exchange OAuth tokens for OpenWebUI tokens
+ENABLE_OAUTH_TOKEN_EXCHANGE = (
+    os.environ.get("ENABLE_OAUTH_TOKEN_EXCHANGE", "False").lower() == "true"
+)
+
 ####################################
 # SCIM Configuration
 ####################################
 
-SCIM_ENABLED = os.environ.get("SCIM_ENABLED", "False").lower() == "true"
+ENABLE_SCIM = (
+    os.environ.get("ENABLE_SCIM", os.environ.get("SCIM_ENABLED", "False")).lower()
+    == "true"
+)
 SCIM_TOKEN = os.environ.get("SCIM_TOKEN", "")
+SCIM_AUTH_PROVIDER = os.environ.get("SCIM_AUTH_PROVIDER", "")
+
+if ENABLE_SCIM and not SCIM_AUTH_PROVIDER:
+    log.warning(
+        "SCIM is enabled but SCIM_AUTH_PROVIDER is not set. "
+        "Set SCIM_AUTH_PROVIDER to the OAuth provider name (e.g. 'microsoft', 'oidc') "
+        "to enable externalId storage."
+    )
 
 ####################################
 # LICENSE_KEY
@@ -512,21 +597,21 @@ LICENSE_PUBLIC_KEY = os.environ.get("LICENSE_PUBLIC_KEY", "")
 
 pk = None
 if LICENSE_PUBLIC_KEY:
-    pk = serialization.load_pem_public_key(
-        f"""
+    pk = serialization.load_pem_public_key(f"""
 -----BEGIN PUBLIC KEY-----
 {LICENSE_PUBLIC_KEY}
 -----END PUBLIC KEY-----
-""".encode(
-            "utf-8"
-        )
-    )
+""".encode("utf-8"))
 
 
 ####################################
 # MODELS
 ####################################
 
+ENABLE_CUSTOM_MODEL_FALLBACK = (
+    os.environ.get("ENABLE_CUSTOM_MODEL_FALLBACK", "False").lower() == "true"
+)
+
 MODELS_CACHE_TTL = os.environ.get("MODELS_CACHE_TTL", "1")
 if MODELS_CACHE_TTL == "":
     MODELS_CACHE_TTL = None
@@ -541,6 +626,11 @@ else:
 # CHAT
 ####################################
 
+ENABLE_CHAT_RESPONSE_BASE64_IMAGE_URL_CONVERSION = (
+    os.environ.get("ENABLE_CHAT_RESPONSE_BASE64_IMAGE_URL_CONVERSION", "False").lower()
+    == "true"
+)
+
 CHAT_RESPONSE_STREAM_DELTA_CHUNK_SIZE = os.environ.get(
     "CHAT_RESPONSE_STREAM_DELTA_CHUNK_SIZE", "1"
 )
@@ -569,6 +659,21 @@ else:
         CHAT_RESPONSE_MAX_TOOL_CALL_RETRIES = 30
 
 
+CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE = os.environ.get(
+    "CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE", ""
+)
+
+if CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE == "":
+    CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE = None
+else:
+    try:
+        CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE = int(
+            CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE
+        )
+    except Exception:
+        CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE = None
+
+
 ####################################
 # WEBSOCKET SUPPORT
 ####################################
@@ -580,6 +685,24 @@ ENABLE_WEBSOCKET_SUPPORT = (
 
 WEBSOCKET_MANAGER = os.environ.get("WEBSOCKET_MANAGER", "")
 
+WEBSOCKET_REDIS_OPTIONS = os.environ.get("WEBSOCKET_REDIS_OPTIONS", "")
+
+
+if WEBSOCKET_REDIS_OPTIONS == "":
+    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
+else:
+    try:
+        WEBSOCKET_REDIS_OPTIONS = json.loads(WEBSOCKET_REDIS_OPTIONS)
+    except Exception:
+        log.warning("Invalid WEBSOCKET_REDIS_OPTIONS, defaulting to None")
+        WEBSOCKET_REDIS_OPTIONS = None
+
 WEBSOCKET_REDIS_URL = os.environ.get("WEBSOCKET_REDIS_URL", REDIS_URL)
 WEBSOCKET_REDIS_CLUSTER = (
     os.environ.get("WEBSOCKET_REDIS_CLUSTER", str(REDIS_CLUSTER)).lower() == "true"
@@ -594,7 +717,30 @@ except ValueError:
 
 WEBSOCKET_SENTINEL_HOSTS = os.environ.get("WEBSOCKET_SENTINEL_HOSTS", "")
 WEBSOCKET_SENTINEL_PORT = os.environ.get("WEBSOCKET_SENTINEL_PORT", "26379")
+WEBSOCKET_SERVER_LOGGING = (
+    os.environ.get("WEBSOCKET_SERVER_LOGGING", "False").lower() == "true"
+)
+WEBSOCKET_SERVER_ENGINEIO_LOGGING = (
+    os.environ.get(
+        "WEBSOCKET_SERVER_ENGINEIO_LOGGING",
+        os.environ.get("WEBSOCKET_SERVER_LOGGING", "False"),
+    ).lower()
+    == "true"
+)
+WEBSOCKET_SERVER_PING_TIMEOUT = os.environ.get("WEBSOCKET_SERVER_PING_TIMEOUT", "20")
+try:
+    WEBSOCKET_SERVER_PING_TIMEOUT = int(WEBSOCKET_SERVER_PING_TIMEOUT)
+except ValueError:
+    WEBSOCKET_SERVER_PING_TIMEOUT = 20
 
+WEBSOCKET_SERVER_PING_INTERVAL = os.environ.get("WEBSOCKET_SERVER_PING_INTERVAL", "25")
+try:
+    WEBSOCKET_SERVER_PING_INTERVAL = int(WEBSOCKET_SERVER_PING_INTERVAL)
+except ValueError:
+    WEBSOCKET_SERVER_PING_INTERVAL = 25
+
+
+REQUESTS_VERIFY = os.environ.get("REQUESTS_VERIFY", "True").lower() == "true"
 
 AIOHTTP_CLIENT_TIMEOUT = os.environ.get("AIOHTTP_CLIENT_TIMEOUT", "")
 
@@ -645,6 +791,17 @@ AIOHTTP_CLIENT_SESSION_TOOL_SERVER_SSL = (
 )
 
 
+RAG_EMBEDDING_TIMEOUT = os.environ.get("RAG_EMBEDDING_TIMEOUT", "")
+
+if RAG_EMBEDDING_TIMEOUT == "":
+    RAG_EMBEDDING_TIMEOUT = None
+else:
+    try:
+        RAG_EMBEDDING_TIMEOUT = int(RAG_EMBEDDING_TIMEOUT)
+    except Exception:
+        RAG_EMBEDDING_TIMEOUT = None
+
+
 ####################################
 # SENTENCE TRANSFORMERS
 ####################################
@@ -689,6 +846,16 @@ else:
     except Exception:
         SENTENCE_TRANSFORMERS_CROSS_ENCODER_MODEL_KWARGS = None
 
+# Whether to apply sigmoid normalization to CrossEncoder reranking scores.
+# When enabled (default), scores are normalized to 0-1 range for proper
+# relevance threshold behavior with MS MARCO models.
+SENTENCE_TRANSFORMERS_CROSS_ENCODER_SIGMOID_ACTIVATION_FUNCTION = (
+    os.environ.get(
+        "SENTENCE_TRANSFORMERS_CROSS_ENCODER_SIGMOID_ACTIVATION_FUNCTION", "True"
+    ).lower()
+    == "true"
+)
+
 ####################################
 # OFFLINE_MODE
 ####################################
@@ -705,8 +872,15 @@ if OFFLINE_MODE:
 ####################################
 # AUDIT LOGGING
 ####################################
+
+
+ENABLE_AUDIT_STDOUT = os.getenv("ENABLE_AUDIT_STDOUT", "False").lower() == "true"
+ENABLE_AUDIT_LOGS_FILE = os.getenv("ENABLE_AUDIT_LOGS_FILE", "True").lower() == "true"
+
 # Where to store log file
-AUDIT_LOGS_FILE_PATH = f"{DATA_DIR}/audit.log"
+# Defaults to the DATA_DIR/audit.log. To set AUDIT_LOGS_FILE_PATH you need to
+# provide the whole path, like: /app/audit.log
+AUDIT_LOGS_FILE_PATH = os.getenv("AUDIT_LOGS_FILE_PATH", f"{DATA_DIR}/audit.log")
 # Maximum size of a file before rotating into a new log file
 AUDIT_LOG_FILE_ROTATION_SIZE = os.getenv("AUDIT_LOG_FILE_ROTATION_SIZE", "10MB")
 

backend/open_webui/functions.py

@@ -35,9 +35,8 @@ from open_webui.utils.plugin import (
     get_function_module_from_cache,
 )
 from open_webui.utils.tools import get_tools
-from open_webui.utils.access_control import has_access
 
-from open_webui.env import SRC_LOG_LEVELS, GLOBAL_LOG_LEVEL
+from open_webui.env import GLOBAL_LOG_LEVEL
 
 from open_webui.utils.misc import (
     add_or_update_system_message,
@@ -51,10 +50,8 @@ from open_webui.utils.payload import (
     apply_system_prompt_to_body,
 )
 
-
 logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MAIN"])
 
 
 def get_function_module_by_id(request: Request, pipe_id: str):

backend/open_webui/internal/db.py

@@ -9,23 +9,23 @@ from open_webui.env import (
     OPEN_WEBUI_DIR,
     DATABASE_URL,
     DATABASE_SCHEMA,
-    SRC_LOG_LEVELS,
     DATABASE_POOL_MAX_OVERFLOW,
     DATABASE_POOL_RECYCLE,
     DATABASE_POOL_SIZE,
     DATABASE_POOL_TIMEOUT,
     DATABASE_ENABLE_SQLITE_WAL,
+    DATABASE_ENABLE_SESSION_SHARING,
+    ENABLE_DB_MIGRATIONS,
 )
 from peewee_migrate import Router
 from sqlalchemy import Dialect, create_engine, MetaData, event, types
 from sqlalchemy.ext.declarative import declarative_base
-from sqlalchemy.orm import scoped_session, sessionmaker
+from sqlalchemy.orm import scoped_session, sessionmaker, Session
 from sqlalchemy.pool import QueuePool, NullPool
 from sqlalchemy.sql.type_api import _T
 from typing_extensions import Self
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["DB"])
 
 
 class JSONField(types.TypeDecorator):
@@ -77,7 +77,8 @@ def handle_peewee_migration(DATABASE_URL):
         assert db.is_closed(), "Database connection is still open."
 
 
-handle_peewee_migration(DATABASE_URL)
+if ENABLE_DB_MIGRATIONS:
+    handle_peewee_migration(DATABASE_URL)
 
 
 SQLALCHEMY_DATABASE_URL = DATABASE_URL
@@ -92,8 +93,6 @@ if SQLALCHEMY_DATABASE_URL.startswith("sqlite+sqlcipher://"):
 
     # Extract database path from SQLCipher URL
     db_path = SQLALCHEMY_DATABASE_URL.replace("sqlite+sqlcipher://", "")
-    if db_path.startswith("/"):
-        db_path = db_path[1:]  # Remove leading slash for relative paths
 
     # Create a custom creator function that uses sqlcipher3
     def create_sqlcipher_connection():
@@ -150,7 +149,7 @@ SessionLocal = sessionmaker(
 )
 metadata_obj = MetaData(schema=DATABASE_SCHEMA)
 Base = declarative_base(metadata=metadata_obj)
-Session = scoped_session(SessionLocal)
+ScopedSession = scoped_session(SessionLocal)
 
 
 def get_session():
@@ -162,3 +161,12 @@ def get_session():
 
 
 get_db = contextmanager(get_session)
+
+
+@contextmanager
+def get_db_context(db: Optional[Session] = None):
+    if isinstance(db, Session) and DATABASE_ENABLE_SESSION_SHARING:
+        yield db
+    else:
+        with get_db() as session:
+            yield session

backend/open_webui/internal/migrations/001_initial_schema.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/002_add_local_sharing.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/003_add_auth_api_key.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/004_add_archived.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/005_add_updated_at.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/006_migrate_timestamps_and_charfields.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/007_add_user_last_active_at.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/008_add_memory.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/009_add_models.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/011_add_user_settings.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/012_add_tools.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/013_add_user_info.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/014_add_files.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/015_add_functions.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/016_add_valves_and_is_active.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/017_add_user_oauth_sub.py

@@ -25,7 +25,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/migrations/018_add_function_is_global.py

@@ -29,7 +29,6 @@ from contextlib import suppress
 import peewee as pw
 from peewee_migrate import Migrator
 
-
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext
 

backend/open_webui/internal/wrappers.py

@@ -2,7 +2,6 @@ import logging
 import os
 from contextvars import ContextVar
 
-from open_webui.env import SRC_LOG_LEVELS
 from peewee import *
 from peewee import InterfaceError as PeeWeeInterfaceError
 from peewee import PostgresqlDatabase
@@ -10,7 +9,6 @@ from playhouse.db_url import connect, parse
 from playhouse.shortcuts import ReconnectMixin
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["DB"])
 
 db_state_default = {"closed": None, "conn": None, "ctx": None, "transactions": None}
 db_state = ContextVar("db_state", default=db_state_default.copy())
@@ -56,8 +54,6 @@ def register_connection(db_url):
         # Parse the database path from SQLCipher URL
         # Convert sqlite+sqlcipher:///path/to/db.sqlite to /path/to/db.sqlite
         db_path = db_url.replace("sqlite+sqlcipher://", "")
-        if db_path.startswith("/"):
-            db_path = db_path[1:]  # Remove leading slash for relative paths
 
         # Use Peewee's native SqlCipherDatabase with encryption
         db = SqlCipherDatabase(db_path, passphrase=database_password)

backend/open_webui/migrations/versions/2f1211949ecc_update_message_and_channel_member_table.py

@@ -0,0 +1,102 @@
+"""Update messages and channel member table
+
+Revision ID: 2f1211949ecc
+Revises: 37f288994c47
+Create Date: 2025-11-27 03:07:56.200231
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+import open_webui.internal.db
+
+# revision identifiers, used by Alembic.
+revision: str = "2f1211949ecc"
+down_revision: Union[str, None] = "37f288994c47"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # New columns to be added to channel_member table
+    op.add_column("channel_member", sa.Column("status", sa.Text(), nullable=True))
+    op.add_column(
+        "channel_member",
+        sa.Column(
+            "is_active",
+            sa.Boolean(),
+            nullable=False,
+            default=True,
+            server_default=sa.sql.expression.true(),
+        ),
+    )
+
+    op.add_column(
+        "channel_member",
+        sa.Column(
+            "is_channel_muted",
+            sa.Boolean(),
+            nullable=False,
+            default=False,
+            server_default=sa.sql.expression.false(),
+        ),
+    )
+    op.add_column(
+        "channel_member",
+        sa.Column(
+            "is_channel_pinned",
+            sa.Boolean(),
+            nullable=False,
+            default=False,
+            server_default=sa.sql.expression.false(),
+        ),
+    )
+
+    op.add_column("channel_member", sa.Column("data", sa.JSON(), nullable=True))
+    op.add_column("channel_member", sa.Column("meta", sa.JSON(), nullable=True))
+
+    op.add_column(
+        "channel_member", sa.Column("joined_at", sa.BigInteger(), nullable=False)
+    )
+    op.add_column(
+        "channel_member", sa.Column("left_at", sa.BigInteger(), nullable=True)
+    )
+
+    op.add_column(
+        "channel_member", sa.Column("last_read_at", sa.BigInteger(), nullable=True)
+    )
+
+    op.add_column(
+        "channel_member", sa.Column("updated_at", sa.BigInteger(), nullable=True)
+    )
+
+    # New columns to be added to message table
+    op.add_column(
+        "message",
+        sa.Column(
+            "is_pinned",
+            sa.Boolean(),
+            nullable=False,
+            default=False,
+            server_default=sa.sql.expression.false(),
+        ),
+    )
+    op.add_column("message", sa.Column("pinned_at", sa.BigInteger(), nullable=True))
+    op.add_column("message", sa.Column("pinned_by", sa.Text(), nullable=True))
+
+
+def downgrade() -> None:
+    op.drop_column("channel_member", "updated_at")
+    op.drop_column("channel_member", "last_read_at")
+
+    op.drop_column("channel_member", "meta")
+    op.drop_column("channel_member", "data")
+
+    op.drop_column("channel_member", "is_channel_pinned")
+    op.drop_column("channel_member", "is_channel_muted")
+
+    op.drop_column("message", "pinned_by")
+    op.drop_column("message", "pinned_at")
+    op.drop_column("message", "is_pinned")

backend/open_webui/migrations/versions/374d2f66af06_add_prompt_history_table.py

@@ -0,0 +1,247 @@
+"""Add prompt history table
+
+Revision ID: 374d2f66af06
+Revises: c440947495f3
+Create Date: 2026-01-23 17:15:00.000000
+
+"""
+
+from typing import Sequence, Union
+import uuid
+
+from alembic import op
+import sqlalchemy as sa
+
+revision: str = "374d2f66af06"
+down_revision: Union[str, None] = "c440947495f3"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+
+    # Step 1: Read existing data from OLD table (schema likely command as PK)
+    # We use batch_alter previously, but we want to move to new table.
+    # We need to assume the OLD structure.
+
+    old_prompt_table = sa.table(
+        "prompt",
+        sa.column("command", sa.Text()),
+        sa.column("user_id", sa.Text()),
+        sa.column("title", sa.Text()),
+        sa.column("content", sa.Text()),
+        sa.column("timestamp", sa.BigInteger()),
+        sa.column("access_control", sa.JSON()),
+    )
+
+    # Check if table exists/read data
+    try:
+        existing_prompts = conn.execute(
+            sa.select(
+                old_prompt_table.c.command,
+                old_prompt_table.c.user_id,
+                old_prompt_table.c.title,
+                old_prompt_table.c.content,
+                old_prompt_table.c.timestamp,
+                old_prompt_table.c.access_control,
+            )
+        ).fetchall()
+    except Exception:
+        # Fallback if table doesn't exist (new install)
+        existing_prompts = []
+
+    # Step 2: Create new prompt table with 'id' as PRIMARY KEY
+    op.create_table(
+        "prompt_new",
+        sa.Column("id", sa.Text(), primary_key=True),
+        sa.Column("command", sa.String(), unique=True, index=True),
+        sa.Column("user_id", sa.String(), nullable=False),
+        sa.Column("name", sa.Text(), nullable=False),
+        sa.Column("content", sa.Text(), nullable=False),
+        sa.Column("data", sa.JSON(), nullable=True),
+        sa.Column("meta", sa.JSON(), nullable=True),
+        sa.Column("access_control", sa.JSON(), nullable=True),
+        sa.Column("is_active", sa.Boolean(), nullable=False, server_default="1"),
+        sa.Column("version_id", sa.Text(), nullable=True),
+        sa.Column("tags", sa.JSON(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+        sa.Column("updated_at", sa.BigInteger(), nullable=False),
+    )
+
+    # Step 3: Create prompt_history table
+    op.create_table(
+        "prompt_history",
+        sa.Column("id", sa.Text(), primary_key=True),
+        sa.Column("prompt_id", sa.Text(), nullable=False, index=True),
+        sa.Column("parent_id", sa.Text(), nullable=True),
+        sa.Column("snapshot", sa.JSON(), nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=False),
+        sa.Column("commit_message", sa.Text(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+    )
+
+    # Step 4: Migrate data
+    prompt_new_table = sa.table(
+        "prompt_new",
+        sa.column("id", sa.Text()),
+        sa.column("command", sa.String()),
+        sa.column("user_id", sa.String()),
+        sa.column("name", sa.Text()),
+        sa.column("content", sa.Text()),
+        sa.column("data", sa.JSON()),
+        sa.column("meta", sa.JSON()),
+        sa.column("access_control", sa.JSON()),
+        sa.column("is_active", sa.Boolean()),
+        sa.column("version_id", sa.Text()),
+        sa.column("tags", sa.JSON()),
+        sa.column("created_at", sa.BigInteger()),
+        sa.column("updated_at", sa.BigInteger()),
+    )
+
+    prompt_history_table = sa.table(
+        "prompt_history",
+        sa.column("id", sa.Text()),
+        sa.column("prompt_id", sa.Text()),
+        sa.column("parent_id", sa.Text()),
+        sa.column("snapshot", sa.JSON()),
+        sa.column("user_id", sa.Text()),
+        sa.column("commit_message", sa.Text()),
+        sa.column("created_at", sa.BigInteger()),
+    )
+
+    for row in existing_prompts:
+        command = row[0]
+        user_id = row[1]
+        title = row[2]
+        content = row[3]
+        timestamp = row[4]
+        access_control = row[5]
+
+        new_uuid = str(uuid.uuid4())
+        history_uuid = str(uuid.uuid4())
+        clean_command = command[1:] if command and command.startswith("/") else command
+
+        # Insert into prompt_new
+        conn.execute(
+            sa.insert(prompt_new_table).values(
+                id=new_uuid,
+                command=clean_command,
+                user_id=user_id,
+                name=title,
+                content=content,
+                data={},
+                meta={},
+                access_control=access_control,
+                is_active=True,
+                version_id=history_uuid,
+                tags=[],
+                created_at=timestamp,
+                updated_at=timestamp,
+            )
+        )
+
+        # Create initial history entry
+        conn.execute(
+            sa.insert(prompt_history_table).values(
+                id=history_uuid,
+                prompt_id=new_uuid,
+                parent_id=None,
+                snapshot={
+                    "name": title,
+                    "content": content,
+                    "command": clean_command,
+                    "data": {},
+                    "meta": {},
+                    "access_control": access_control,
+                },
+                user_id=user_id,
+                commit_message=None,
+                created_at=timestamp,
+            )
+        )
+
+    # Step 5: Replace old table with new one
+    op.drop_table("prompt")
+    op.rename_table("prompt_new", "prompt")
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+
+    # Step 1: Read new data
+    prompt_table = sa.table(
+        "prompt",
+        sa.column("command", sa.String()),
+        sa.column("name", sa.Text()),
+        sa.column("created_at", sa.BigInteger()),
+        sa.column("user_id", sa.Text()),
+        sa.column("content", sa.Text()),
+        sa.column("access_control", sa.JSON()),
+    )
+
+    try:
+        current_data = conn.execute(
+            sa.select(
+                prompt_table.c.command,
+                prompt_table.c.name,
+                prompt_table.c.created_at,
+                prompt_table.c.user_id,
+                prompt_table.c.content,
+                prompt_table.c.access_control,
+            )
+        ).fetchall()
+    except Exception:
+        current_data = []
+
+    # Step 2: Drop history and table
+    op.drop_table("prompt_history")
+    op.drop_table("prompt")
+
+    # Step 3: Recreate old table (command as PK?)
+    # Assuming old schema:
+    op.create_table(
+        "prompt",
+        sa.Column("command", sa.String(), primary_key=True),
+        sa.Column("user_id", sa.String()),
+        sa.Column("title", sa.Text()),
+        sa.Column("content", sa.Text()),
+        sa.Column("timestamp", sa.BigInteger()),
+        sa.Column("access_control", sa.JSON()),
+        sa.Column("id", sa.Integer(), nullable=True),
+    )
+
+    # Step 4: Restore data
+    old_prompt_table = sa.table(
+        "prompt",
+        sa.column("command", sa.String()),
+        sa.column("user_id", sa.String()),
+        sa.column("title", sa.Text()),
+        sa.column("content", sa.Text()),
+        sa.column("timestamp", sa.BigInteger()),
+        sa.column("access_control", sa.JSON()),
+    )
+
+    for row in current_data:
+        command = row[0]
+        name = row[1]
+        created_at = row[2]
+        user_id = row[3]
+        content = row[4]
+        access_control = row[5]
+
+        # Restore leading /
+        old_command = (
+            "/" + command if command and not command.startswith("/") else command
+        )
+
+        conn.execute(
+            sa.insert(old_prompt_table).values(
+                command=old_command,
+                user_id=user_id,
+                title=name,
+                content=content,
+                timestamp=created_at,
+                access_control=access_control,
+            )
+        )

backend/open_webui/migrations/versions/37f288994c47_add_group_member_table.py

@@ -0,0 +1,145 @@
+"""add_group_member_table
+
+Revision ID: 37f288994c47
+Revises: a5c220713937
+Create Date: 2025-11-17 03:45:25.123939
+
+"""
+
+import uuid
+import time
+import json
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision: str = "37f288994c47"
+down_revision: Union[str, None] = "a5c220713937"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # 1. Create new table
+    op.create_table(
+        "group_member",
+        sa.Column("id", sa.Text(), primary_key=True, unique=True, nullable=False),
+        sa.Column(
+            "group_id",
+            sa.Text(),
+            sa.ForeignKey("group.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column(
+            "user_id",
+            sa.Text(),
+            sa.ForeignKey("user.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column("created_at", sa.BigInteger(), nullable=True),
+        sa.Column("updated_at", sa.BigInteger(), nullable=True),
+        sa.UniqueConstraint("group_id", "user_id", name="uq_group_member_group_user"),
+    )
+
+    connection = op.get_bind()
+
+    # 2. Read existing group with user_ids JSON column
+    group_table = sa.Table(
+        "group",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("user_ids", sa.JSON()),  # JSON stored as text in SQLite + PG
+    )
+
+    results = connection.execute(
+        sa.select(group_table.c.id, group_table.c.user_ids)
+    ).fetchall()
+
+    print(results)
+
+    # 3. Insert members into group_member table
+    gm_table = sa.Table(
+        "group_member",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("group_id", sa.Text()),
+        sa.Column("user_id", sa.Text()),
+        sa.Column("created_at", sa.BigInteger()),
+        sa.Column("updated_at", sa.BigInteger()),
+    )
+
+    now = int(time.time())
+    for group_id, user_ids in results:
+        if not user_ids:
+            continue
+
+        if isinstance(user_ids, str):
+            try:
+                user_ids = json.loads(user_ids)
+            except Exception:
+                continue  # skip invalid JSON
+
+        if not isinstance(user_ids, list):
+            continue
+
+        rows = [
+            {
+                "id": str(uuid.uuid4()),
+                "group_id": group_id,
+                "user_id": uid,
+                "created_at": now,
+                "updated_at": now,
+            }
+            for uid in user_ids
+        ]
+
+        if rows:
+            connection.execute(gm_table.insert(), rows)
+
+    # 4. Optionally drop the old column
+    with op.batch_alter_table("group") as batch:
+        batch.drop_column("user_ids")
+
+
+def downgrade():
+    # Reverse: restore user_ids column
+    with op.batch_alter_table("group") as batch:
+        batch.add_column(sa.Column("user_ids", sa.JSON()))
+
+    connection = op.get_bind()
+    gm_table = sa.Table(
+        "group_member",
+        sa.MetaData(),
+        sa.Column("group_id", sa.Text()),
+        sa.Column("user_id", sa.Text()),
+        sa.Column("created_at", sa.BigInteger()),
+        sa.Column("updated_at", sa.BigInteger()),
+    )
+
+    group_table = sa.Table(
+        "group",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("user_ids", sa.JSON()),
+    )
+
+    # Build JSON arrays again
+    results = connection.execute(sa.select(group_table.c.id)).fetchall()
+
+    for (group_id,) in results:
+        members = connection.execute(
+            sa.select(gm_table.c.user_id).where(gm_table.c.group_id == group_id)
+        ).fetchall()
+
+        member_ids = [m[0] for m in members]
+
+        connection.execute(
+            group_table.update()
+            .where(group_table.c.id == group_id)
+            .values(user_ids=member_ids)
+        )
+
+    # Drop the new table
+    op.drop_table("group_member")

backend/open_webui/migrations/versions/38d63c18f30f_add_oauth_session_table.py

@@ -11,7 +11,6 @@ from typing import Sequence, Union
 from alembic import op
 import sqlalchemy as sa
 
-
 # revision identifiers, used by Alembic.
 revision: str = "38d63c18f30f"
 down_revision: Union[str, None] = "3af16a1c9fb6"
@@ -20,18 +19,46 @@ depends_on: Union[str, Sequence[str], None] = None
 
 
 def upgrade() -> None:
+    # Ensure 'id' column in 'user' table is unique and primary key (ForeignKey constraint)
+    inspector = sa.inspect(op.get_bind())
+    columns = inspector.get_columns("user")
+
+    pk_columns = inspector.get_pk_constraint("user")["constrained_columns"]
+    id_column = next((col for col in columns if col["name"] == "id"), None)
+
+    if id_column and not id_column.get("unique", False):
+        unique_constraints = inspector.get_unique_constraints("user")
+        unique_columns = {tuple(u["column_names"]) for u in unique_constraints}
+
+        with op.batch_alter_table("user") as batch_op:
+            # If primary key is wrong, drop it
+            if pk_columns and pk_columns != ["id"]:
+                batch_op.drop_constraint(
+                    inspector.get_pk_constraint("user")["name"], type_="primary"
+                )
+
+            # Add unique constraint if missing
+            if ("id",) not in unique_columns:
+                batch_op.create_unique_constraint("uq_user_id", ["id"])
+
+            # Re-create correct primary key
+            batch_op.create_primary_key("pk_user_id", ["id"])
+
     # Create oauth_session table
     op.create_table(
         "oauth_session",
-        sa.Column("id", sa.Text(), nullable=False),
-        sa.Column("user_id", sa.Text(), nullable=False),
+        sa.Column("id", sa.Text(), primary_key=True, nullable=False, unique=True),
+        sa.Column(
+            "user_id",
+            sa.Text(),
+            sa.ForeignKey("user.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
         sa.Column("provider", sa.Text(), nullable=False),
         sa.Column("token", sa.Text(), nullable=False),
         sa.Column("expires_at", sa.BigInteger(), nullable=False),
         sa.Column("created_at", sa.BigInteger(), nullable=False),
         sa.Column("updated_at", sa.BigInteger(), nullable=False),
-        sa.PrimaryKeyConstraint("id"),
-        sa.ForeignKeyConstraint(["user_id"], ["user.id"], ondelete="CASCADE"),
     )
 
     # Create indexes for better performance

backend/open_webui/migrations/versions/3e0e00844bb0_add_knowledge_file_table.py

@@ -0,0 +1,169 @@
+"""Add knowledge_file table
+
+Revision ID: 3e0e00844bb0
+Revises: 90ef40d4714e
+Create Date: 2025-12-02 06:54:19.401334
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy import inspect
+import open_webui.internal.db
+
+import time
+import json
+import uuid
+
+# revision identifiers, used by Alembic.
+revision: str = "3e0e00844bb0"
+down_revision: Union[str, None] = "90ef40d4714e"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "knowledge_file",
+        sa.Column("id", sa.Text(), primary_key=True),
+        sa.Column("user_id", sa.Text(), nullable=False),
+        sa.Column(
+            "knowledge_id",
+            sa.Text(),
+            sa.ForeignKey("knowledge.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column(
+            "file_id",
+            sa.Text(),
+            sa.ForeignKey("file.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+        sa.Column("updated_at", sa.BigInteger(), nullable=False),
+        # indexes
+        sa.Index("ix_knowledge_file_knowledge_id", "knowledge_id"),
+        sa.Index("ix_knowledge_file_file_id", "file_id"),
+        sa.Index("ix_knowledge_file_user_id", "user_id"),
+        # unique constraints
+        sa.UniqueConstraint(
+            "knowledge_id", "file_id", name="uq_knowledge_file_knowledge_file"
+        ),  # prevent duplicate entries
+    )
+
+    connection = op.get_bind()
+
+    # 2. Read existing group with user_ids JSON column
+    knowledge_table = sa.Table(
+        "knowledge",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("user_id", sa.Text()),
+        sa.Column("data", sa.JSON()),  # JSON stored as text in SQLite + PG
+    )
+
+    results = connection.execute(
+        sa.select(
+            knowledge_table.c.id, knowledge_table.c.user_id, knowledge_table.c.data
+        )
+    ).fetchall()
+
+    # 3. Insert members into group_member table
+    kf_table = sa.Table(
+        "knowledge_file",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("user_id", sa.Text()),
+        sa.Column("knowledge_id", sa.Text()),
+        sa.Column("file_id", sa.Text()),
+        sa.Column("created_at", sa.BigInteger()),
+        sa.Column("updated_at", sa.BigInteger()),
+    )
+
+    file_table = sa.Table(
+        "file",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+    )
+
+    now = int(time.time())
+    for knowledge_id, user_id, data in results:
+        if not data:
+            continue
+
+        if isinstance(data, str):
+            try:
+                data = json.loads(data)
+            except Exception:
+                continue  # skip invalid JSON
+
+        if not isinstance(data, dict):
+            continue
+
+        file_ids = data.get("file_ids", [])
+
+        for file_id in file_ids:
+            file_exists = connection.execute(
+                sa.select(file_table.c.id).where(file_table.c.id == file_id)
+            ).fetchone()
+
+            if not file_exists:
+                continue  # skip non-existing files
+
+            row = {
+                "id": str(uuid.uuid4()),
+                "user_id": user_id,
+                "knowledge_id": knowledge_id,
+                "file_id": file_id,
+                "created_at": now,
+                "updated_at": now,
+            }
+            connection.execute(kf_table.insert().values(**row))
+
+    with op.batch_alter_table("knowledge") as batch:
+        batch.drop_column("data")
+
+
+def downgrade() -> None:
+    # 1. Add back the old data column
+    op.add_column("knowledge", sa.Column("data", sa.JSON(), nullable=True))
+
+    connection = op.get_bind()
+
+    # 2. Read knowledge_file entries and reconstruct data JSON
+    knowledge_table = sa.Table(
+        "knowledge",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("data", sa.JSON()),
+    )
+
+    kf_table = sa.Table(
+        "knowledge_file",
+        sa.MetaData(),
+        sa.Column("id", sa.Text()),
+        sa.Column("knowledge_id", sa.Text()),
+        sa.Column("file_id", sa.Text()),
+    )
+
+    results = connection.execute(sa.select(knowledge_table.c.id)).fetchall()
+
+    for (knowledge_id,) in results:
+        file_ids = connection.execute(
+            sa.select(kf_table.c.file_id).where(kf_table.c.knowledge_id == knowledge_id)
+        ).fetchall()
+
+        file_ids_list = [fid for (fid,) in file_ids]
+
+        data_json = {"file_ids": file_ids_list}
+
+        connection.execute(
+            knowledge_table.update()
+            .where(knowledge_table.c.id == knowledge_id)
+            .values(data=data_json)
+        )
+
+    # 3. Drop the knowledge_file table
+    op.drop_table("knowledge_file")

backend/open_webui/migrations/versions/6283dc0e4d8d_add_channel_file_table.py

@@ -0,0 +1,53 @@
+"""Add channel file table
+
+Revision ID: 6283dc0e4d8d
+Revises: 3e0e00844bb0
+Create Date: 2025-12-10 15:11:39.424601
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+import open_webui.internal.db
+
+# revision identifiers, used by Alembic.
+revision: str = "6283dc0e4d8d"
+down_revision: Union[str, None] = "3e0e00844bb0"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "channel_file",
+        sa.Column("id", sa.Text(), primary_key=True),
+        sa.Column("user_id", sa.Text(), nullable=False),
+        sa.Column(
+            "channel_id",
+            sa.Text(),
+            sa.ForeignKey("channel.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column(
+            "file_id",
+            sa.Text(),
+            sa.ForeignKey("file.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+        sa.Column("updated_at", sa.BigInteger(), nullable=False),
+        # indexes
+        sa.Index("ix_channel_file_channel_id", "channel_id"),
+        sa.Index("ix_channel_file_file_id", "file_id"),
+        sa.Index("ix_channel_file_user_id", "user_id"),
+        # unique constraints
+        sa.UniqueConstraint(
+            "channel_id", "file_id", name="uq_channel_file_channel_file"
+        ),  # prevent duplicate entries
+    )
+
+
+def downgrade() -> None:
+    op.drop_table("channel_file")

backend/open_webui/migrations/versions/6a39f3d8e55c_add_knowledge_table.py

@@ -11,7 +11,6 @@ import sqlalchemy as sa
 from sqlalchemy.sql import table, column, select
 import json
 
-
 revision = "6a39f3d8e55c"
 down_revision = "c0fbf31ca0db"
 branch_labels = None

backend/open_webui/migrations/versions/81cc2ce44d79_update_channel_file_and_knowledge_table.py

@@ -0,0 +1,48 @@
+"""Update channel file and knowledge table
+
+Revision ID: 81cc2ce44d79
+Revises: 6283dc0e4d8d
+Create Date: 2025-12-10 16:07:58.001282
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+import open_webui.internal.db
+
+# revision identifiers, used by Alembic.
+revision: str = "81cc2ce44d79"
+down_revision: Union[str, None] = "6283dc0e4d8d"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # Add message_id column to channel_file table
+    with op.batch_alter_table("channel_file", schema=None) as batch_op:
+        batch_op.add_column(
+            sa.Column(
+                "message_id",
+                sa.Text(),
+                sa.ForeignKey(
+                    "message.id", ondelete="CASCADE", name="fk_channel_file_message_id"
+                ),
+                nullable=True,
+            )
+        )
+
+    # Add data column to knowledge table
+    with op.batch_alter_table("knowledge", schema=None) as batch_op:
+        batch_op.add_column(sa.Column("data", sa.JSON(), nullable=True))
+
+
+def downgrade() -> None:
+    # Remove message_id column from channel_file table
+    with op.batch_alter_table("channel_file", schema=None) as batch_op:
+        batch_op.drop_column("message_id")
+
+    # Remove data column from knowledge table
+    with op.batch_alter_table("knowledge", schema=None) as batch_op:
+        batch_op.drop_column("data")

backend/open_webui/migrations/versions/8452d01d26d7_add_chat_message_table.py

@@ -0,0 +1,177 @@
+"""Add chat_message table
+
+Revision ID: 8452d01d26d7
+Revises: 374d2f66af06
+Create Date: 2026-02-01 04:00:00.000000
+
+"""
+
+import time
+import json
+import logging
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+log = logging.getLogger(__name__)
+
+revision: str = "8452d01d26d7"
+down_revision: Union[str, None] = "374d2f66af06"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # Step 1: Create table
+    op.create_table(
+        "chat_message",
+        sa.Column("id", sa.Text(), primary_key=True),
+        sa.Column("chat_id", sa.Text(), nullable=False, index=True),
+        sa.Column("user_id", sa.Text(), index=True),
+        sa.Column("role", sa.Text(), nullable=False),
+        sa.Column("parent_id", sa.Text(), nullable=True),
+        sa.Column("content", sa.JSON(), nullable=True),
+        sa.Column("output", sa.JSON(), nullable=True),
+        sa.Column("model_id", sa.Text(), nullable=True, index=True),
+        sa.Column("files", sa.JSON(), nullable=True),
+        sa.Column("sources", sa.JSON(), nullable=True),
+        sa.Column("embeds", sa.JSON(), nullable=True),
+        sa.Column("done", sa.Boolean(), default=True),
+        sa.Column("status_history", sa.JSON(), nullable=True),
+        sa.Column("error", sa.JSON(), nullable=True),
+        sa.Column("usage", sa.JSON(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), index=True),
+        sa.Column("updated_at", sa.BigInteger()),
+        sa.ForeignKeyConstraint(["chat_id"], ["chat.id"], ondelete="CASCADE"),
+    )
+
+    # Create composite indexes
+    op.create_index(
+        "chat_message_chat_parent_idx", "chat_message", ["chat_id", "parent_id"]
+    )
+    op.create_index(
+        "chat_message_model_created_idx", "chat_message", ["model_id", "created_at"]
+    )
+    op.create_index(
+        "chat_message_user_created_idx", "chat_message", ["user_id", "created_at"]
+    )
+
+    # Step 2: Backfill from existing chats
+    conn = op.get_bind()
+
+    chat_table = sa.table(
+        "chat",
+        sa.column("id", sa.Text()),
+        sa.column("user_id", sa.Text()),
+        sa.column("chat", sa.JSON()),
+    )
+
+    chat_message_table = sa.table(
+        "chat_message",
+        sa.column("id", sa.Text()),
+        sa.column("chat_id", sa.Text()),
+        sa.column("user_id", sa.Text()),
+        sa.column("role", sa.Text()),
+        sa.column("parent_id", sa.Text()),
+        sa.column("content", sa.JSON()),
+        sa.column("output", sa.JSON()),
+        sa.column("model_id", sa.Text()),
+        sa.column("files", sa.JSON()),
+        sa.column("sources", sa.JSON()),
+        sa.column("embeds", sa.JSON()),
+        sa.column("done", sa.Boolean()),
+        sa.column("status_history", sa.JSON()),
+        sa.column("error", sa.JSON()),
+        sa.column("usage", sa.JSON()),
+        sa.column("created_at", sa.BigInteger()),
+        sa.column("updated_at", sa.BigInteger()),
+    )
+
+    # Fetch all chats (excluding shared chats which have user_id starting with 'shared-')
+    chats = conn.execute(
+        sa.select(chat_table.c.id, chat_table.c.user_id, chat_table.c.chat).where(
+            ~chat_table.c.user_id.like("shared-%")
+        )
+    ).fetchall()
+
+    now = int(time.time())
+    messages_inserted = 0
+    messages_failed = 0
+
+    for chat_row in chats:
+        chat_id = chat_row[0]
+        user_id = chat_row[1]
+        chat_data = chat_row[2]
+
+        if not chat_data:
+            continue
+
+        # Handle both string and dict chat data
+        if isinstance(chat_data, str):
+            try:
+                chat_data = json.loads(chat_data)
+            except Exception:
+                continue
+
+        history = chat_data.get("history", {})
+        messages = history.get("messages", {})
+
+        for message_id, message in messages.items():
+            if not isinstance(message, dict):
+                continue
+
+            role = message.get("role")
+            if not role:
+                continue
+
+            timestamp = message.get("timestamp", now)
+
+            # Normalize timestamp: convert ms to seconds, validate range
+            if timestamp > 10_000_000_000:
+                timestamp = timestamp // 1000
+            # Must be after 2020 and not too far in the future
+            if timestamp < 1577836800 or timestamp > now + 86400:
+                timestamp = now
+
+            # Use savepoint to allow individual insert failures without aborting transaction
+            savepoint = conn.begin_nested()
+            try:
+                conn.execute(
+                    sa.insert(chat_message_table).values(
+                        id=f"{chat_id}-{message_id}",
+                        chat_id=chat_id,
+                        user_id=user_id,
+                        role=role,
+                        parent_id=message.get("parentId"),
+                        content=message.get("content"),
+                        output=message.get("output"),
+                        model_id=message.get("model"),
+                        files=message.get("files"),
+                        sources=message.get("sources"),
+                        embeds=message.get("embeds"),
+                        done=message.get("done", True),
+                        status_history=message.get("statusHistory"),
+                        error=message.get("error"),
+                        created_at=timestamp,
+                        updated_at=timestamp,
+                    )
+                )
+                savepoint.commit()
+                messages_inserted += 1
+            except Exception as e:
+                savepoint.rollback()
+                messages_failed += 1
+                log.warning(f"Failed to insert message {message_id}: {e}")
+                continue
+
+    log.info(
+        f"Backfilled {messages_inserted} messages into chat_message table ({messages_failed} failed)"
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("chat_message_user_created_idx", table_name="chat_message")
+    op.drop_index("chat_message_model_created_idx", table_name="chat_message")
+    op.drop_index("chat_message_chat_parent_idx", table_name="chat_message")
+    op.drop_table("chat_message")

backend/open_webui/migrations/versions/90ef40d4714e_update_channel_and_channel_members_table.py

@@ -0,0 +1,80 @@
+"""Update channel and channel members table
+
+Revision ID: 90ef40d4714e
+Revises: b10670c03dd5
+Create Date: 2025-11-30 06:33:38.790341
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+import open_webui.internal.db
+
+# revision identifiers, used by Alembic.
+revision: str = "90ef40d4714e"
+down_revision: Union[str, None] = "b10670c03dd5"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # Update 'channel' table
+    op.add_column("channel", sa.Column("is_private", sa.Boolean(), nullable=True))
+
+    op.add_column("channel", sa.Column("archived_at", sa.BigInteger(), nullable=True))
+    op.add_column("channel", sa.Column("archived_by", sa.Text(), nullable=True))
+
+    op.add_column("channel", sa.Column("deleted_at", sa.BigInteger(), nullable=True))
+    op.add_column("channel", sa.Column("deleted_by", sa.Text(), nullable=True))
+
+    op.add_column("channel", sa.Column("updated_by", sa.Text(), nullable=True))
+
+    # Update 'channel_member' table
+    op.add_column("channel_member", sa.Column("role", sa.Text(), nullable=True))
+    op.add_column("channel_member", sa.Column("invited_by", sa.Text(), nullable=True))
+    op.add_column(
+        "channel_member", sa.Column("invited_at", sa.BigInteger(), nullable=True)
+    )
+
+    #  Create 'channel_webhook' table
+    op.create_table(
+        "channel_webhook",
+        sa.Column("id", sa.Text(), primary_key=True, unique=True, nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=False),
+        sa.Column(
+            "channel_id",
+            sa.Text(),
+            sa.ForeignKey("channel.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column("name", sa.Text(), nullable=False),
+        sa.Column("profile_image_url", sa.Text(), nullable=True),
+        sa.Column("token", sa.Text(), nullable=False),
+        sa.Column("last_used_at", sa.BigInteger(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+        sa.Column("updated_at", sa.BigInteger(), nullable=False),
+    )
+
+    pass
+
+
+def downgrade() -> None:
+    # Downgrade 'channel' table
+    op.drop_column("channel", "is_private")
+    op.drop_column("channel", "archived_at")
+    op.drop_column("channel", "archived_by")
+    op.drop_column("channel", "deleted_at")
+    op.drop_column("channel", "deleted_by")
+    op.drop_column("channel", "updated_by")
+
+    # Downgrade 'channel_member' table
+    op.drop_column("channel_member", "role")
+    op.drop_column("channel_member", "invited_by")
+    op.drop_column("channel_member", "invited_at")
+
+    # Drop 'channel_webhook' table
+    op.drop_table("channel_webhook")
+
+    pass

backend/open_webui/migrations/versions/a1b2c3d4e5f6_add_skill_table.py

@@ -0,0 +1,45 @@
+"""Add skill table
+
+Revision ID: a1b2c3d4e5f6
+Revises: f1e2d3c4b5a6
+Create Date: 2026-02-11 09:30:00.000000
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+from open_webui.migrations.util import get_existing_tables
+
+revision: str = "a1b2c3d4e5f6"
+down_revision: Union[str, None] = "f1e2d3c4b5a6"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    existing_tables = set(get_existing_tables())
+
+    if "skill" not in existing_tables:
+        op.create_table(
+            "skill",
+            sa.Column("id", sa.String(), nullable=False, primary_key=True),
+            sa.Column("user_id", sa.String(), nullable=False),
+            sa.Column("name", sa.Text(), nullable=False, unique=True),
+            sa.Column("description", sa.Text(), nullable=True),
+            sa.Column("content", sa.Text(), nullable=False),
+            sa.Column("meta", sa.JSON(), nullable=True),
+            sa.Column("is_active", sa.Boolean(), nullable=False),
+            sa.Column("updated_at", sa.BigInteger(), nullable=False),
+            sa.Column("created_at", sa.BigInteger(), nullable=False),
+        )
+        op.create_index("idx_skill_user_id", "skill", ["user_id"])
+        op.create_index("idx_skill_updated_at", "skill", ["updated_at"])
+
+
+def downgrade() -> None:
+    op.drop_index("idx_skill_updated_at", table_name="skill")
+    op.drop_index("idx_skill_user_id", table_name="skill")
+    op.drop_table("skill")

backend/open_webui/migrations/versions/b10670c03dd5_update_user_table.py

@@ -0,0 +1,249 @@
+"""Update user table
+
+Revision ID: b10670c03dd5
+Revises: 2f1211949ecc
+Create Date: 2025-11-28 04:55:31.737538
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+import open_webui.internal.db
+import json
+import time
+
+# revision identifiers, used by Alembic.
+revision: str = "b10670c03dd5"
+down_revision: Union[str, None] = "2f1211949ecc"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def _drop_sqlite_indexes_for_column(table_name, column_name, conn):
+    """
+    SQLite requires manual removal of any indexes referencing a column
+    before ALTER TABLE ... DROP COLUMN can succeed.
+    """
+    indexes = conn.execute(sa.text(f"PRAGMA index_list('{table_name}')")).fetchall()
+
+    for idx in indexes:
+        index_name = idx[1]  # index name
+        # Get indexed columns
+        idx_info = conn.execute(
+            sa.text(f"PRAGMA index_info('{index_name}')")
+        ).fetchall()
+
+        indexed_cols = [row[2] for row in idx_info]  # col names
+        if column_name in indexed_cols:
+            conn.execute(sa.text(f"DROP INDEX IF EXISTS {index_name}"))
+
+
+def _convert_column_to_json(table: str, column: str):
+    conn = op.get_bind()
+    dialect = conn.dialect.name
+
+    # SQLite cannot ALTER COLUMN → must recreate column
+    if dialect == "sqlite":
+        # 1. Add temporary column
+        op.add_column(table, sa.Column(f"{column}_json", sa.JSON(), nullable=True))
+
+        # 2. Load old data
+        rows = conn.execute(sa.text(f'SELECT id, {column} FROM "{table}"')).fetchall()
+
+        for row in rows:
+            uid, raw = row
+            if raw is None:
+                parsed = None
+            else:
+                try:
+                    parsed = json.loads(raw)
+                except Exception:
+                    parsed = None  # fallback safe behavior
+
+            conn.execute(
+                sa.text(f'UPDATE "{table}" SET {column}_json = :val WHERE id = :id'),
+                {"val": json.dumps(parsed) if parsed else None, "id": uid},
+            )
+
+        # 3. Drop old TEXT column
+        op.drop_column(table, column)
+
+        # 4. Rename new JSON column → original name
+        op.alter_column(table, f"{column}_json", new_column_name=column)
+
+    else:
+        # PostgreSQL supports direct CAST
+        op.alter_column(
+            table,
+            column,
+            type_=sa.JSON(),
+            postgresql_using=f"{column}::json",
+        )
+
+
+def _convert_column_to_text(table: str, column: str):
+    conn = op.get_bind()
+    dialect = conn.dialect.name
+
+    if dialect == "sqlite":
+        op.add_column(table, sa.Column(f"{column}_text", sa.Text(), nullable=True))
+
+        rows = conn.execute(sa.text(f'SELECT id, {column} FROM "{table}"')).fetchall()
+
+        for uid, raw in rows:
+            conn.execute(
+                sa.text(f'UPDATE "{table}" SET {column}_text = :val WHERE id = :id'),
+                {"val": json.dumps(raw) if raw else None, "id": uid},
+            )
+
+        op.drop_column(table, column)
+        op.alter_column(table, f"{column}_text", new_column_name=column)
+
+    else:
+        op.alter_column(
+            table,
+            column,
+            type_=sa.Text(),
+            postgresql_using=f"to_json({column})::text",
+        )
+
+
+def upgrade() -> None:
+    op.add_column(
+        "user", sa.Column("profile_banner_image_url", sa.Text(), nullable=True)
+    )
+    op.add_column("user", sa.Column("timezone", sa.String(), nullable=True))
+
+    op.add_column("user", sa.Column("presence_state", sa.String(), nullable=True))
+    op.add_column("user", sa.Column("status_emoji", sa.String(), nullable=True))
+    op.add_column("user", sa.Column("status_message", sa.Text(), nullable=True))
+    op.add_column(
+        "user", sa.Column("status_expires_at", sa.BigInteger(), nullable=True)
+    )
+
+    op.add_column("user", sa.Column("oauth", sa.JSON(), nullable=True))
+
+    # Convert info (TEXT/JSONField) → JSON
+    _convert_column_to_json("user", "info")
+    # Convert settings (TEXT/JSONField) → JSON
+    _convert_column_to_json("user", "settings")
+
+    op.create_table(
+        "api_key",
+        sa.Column("id", sa.Text(), primary_key=True, unique=True),
+        sa.Column("user_id", sa.Text(), sa.ForeignKey("user.id", ondelete="CASCADE")),
+        sa.Column("key", sa.Text(), unique=True, nullable=False),
+        sa.Column("data", sa.JSON(), nullable=True),
+        sa.Column("expires_at", sa.BigInteger(), nullable=True),
+        sa.Column("last_used_at", sa.BigInteger(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+        sa.Column("updated_at", sa.BigInteger(), nullable=False),
+    )
+
+    conn = op.get_bind()
+    users = conn.execute(
+        sa.text('SELECT id, oauth_sub FROM "user" WHERE oauth_sub IS NOT NULL')
+    ).fetchall()
+
+    for uid, oauth_sub in users:
+        if oauth_sub:
+            # Example formats supported:
+            #   provider@sub
+            #   plain sub (stored as {"oidc": {"sub": sub}})
+            if "@" in oauth_sub:
+                provider, sub = oauth_sub.split("@", 1)
+            else:
+                provider, sub = "oidc", oauth_sub
+
+            oauth_json = json.dumps({provider: {"sub": sub}})
+            conn.execute(
+                sa.text('UPDATE "user" SET oauth = :oauth WHERE id = :id'),
+                {"oauth": oauth_json, "id": uid},
+            )
+
+    users_with_keys = conn.execute(
+        sa.text('SELECT id, api_key FROM "user" WHERE api_key IS NOT NULL')
+    ).fetchall()
+    now = int(time.time())
+
+    for uid, api_key in users_with_keys:
+        if api_key:
+            conn.execute(
+                sa.text("""
+                    INSERT INTO api_key (id, user_id, key, created_at, updated_at)
+                    VALUES (:id, :user_id, :key, :created_at, :updated_at)
+                """),
+                {
+                    "id": f"key_{uid}",
+                    "user_id": uid,
+                    "key": api_key,
+                    "created_at": now,
+                    "updated_at": now,
+                },
+            )
+
+    if conn.dialect.name == "sqlite":
+        _drop_sqlite_indexes_for_column("user", "api_key", conn)
+        _drop_sqlite_indexes_for_column("user", "oauth_sub", conn)
+
+    with op.batch_alter_table("user") as batch_op:
+        batch_op.drop_column("api_key")
+        batch_op.drop_column("oauth_sub")
+
+
+def downgrade() -> None:
+    # --- 1. Restore old oauth_sub column ---
+    op.add_column("user", sa.Column("oauth_sub", sa.Text(), nullable=True))
+
+    conn = op.get_bind()
+    users = conn.execute(
+        sa.text('SELECT id, oauth FROM "user" WHERE oauth IS NOT NULL')
+    ).fetchall()
+
+    for uid, oauth in users:
+        try:
+            data = json.loads(oauth)
+            provider = list(data.keys())[0]
+            sub = data[provider].get("sub")
+            oauth_sub = f"{provider}@{sub}"
+        except Exception:
+            oauth_sub = None
+
+        conn.execute(
+            sa.text('UPDATE "user" SET oauth_sub = :oauth_sub WHERE id = :id'),
+            {"oauth_sub": oauth_sub, "id": uid},
+        )
+
+    op.drop_column("user", "oauth")
+
+    # --- 2. Restore api_key field ---
+    op.add_column("user", sa.Column("api_key", sa.String(), nullable=True))
+
+    # Restore values from api_key
+    keys = conn.execute(sa.text("SELECT user_id, key FROM api_key")).fetchall()
+    for uid, key in keys:
+        conn.execute(
+            sa.text('UPDATE "user" SET api_key = :key WHERE id = :id'),
+            {"key": key, "id": uid},
+        )
+
+    # Drop new table
+    op.drop_table("api_key")
+
+    with op.batch_alter_table("user") as batch_op:
+        batch_op.drop_column("profile_banner_image_url")
+        batch_op.drop_column("timezone")
+
+        batch_op.drop_column("presence_state")
+        batch_op.drop_column("status_emoji")
+        batch_op.drop_column("status_message")
+        batch_op.drop_column("status_expires_at")
+
+    # Convert info (JSON) → TEXT
+    _convert_column_to_text("user", "info")
+    # Convert settings (JSON) → TEXT
+    _convert_column_to_text("user", "settings")

backend/open_webui/migrations/versions/b2c3d4e5f6a7_add_scim_column_to_user_table.py

@@ -0,0 +1,26 @@
+"""add scim column to user table
+
+Revision ID: b2c3d4e5f6a7
+Revises: a1b2c3d4e5f6
+Create Date: 2026-02-13 14:19:00.000000
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision: str = "b2c3d4e5f6a7"
+down_revision: Union[str, None] = "a1b2c3d4e5f6"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.add_column("user", sa.Column("scim", sa.JSON(), nullable=True))
+
+
+def downgrade() -> None:
+    op.drop_column("user", "scim")

backend/open_webui/migrations/versions/c29facfe716b_update_file_table_path.py

@@ -12,7 +12,6 @@ import json
 from sqlalchemy.sql import table, column
 from sqlalchemy import String, Text, JSON, and_
 
-
 revision = "c29facfe716b"
 down_revision = "c69f45358db4"
 branch_labels = None

backend/open_webui/migrations/versions/c440947495f3_add_chat_file_table.py

@@ -0,0 +1,56 @@
+"""Add chat_file table
+
+Revision ID: c440947495f3
+Revises: 81cc2ce44d79
+Create Date: 2025-12-21 20:27:41.694897
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision: str = "c440947495f3"
+down_revision: Union[str, None] = "81cc2ce44d79"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "chat_file",
+        sa.Column("id", sa.Text(), primary_key=True),
+        sa.Column("user_id", sa.Text(), nullable=False),
+        sa.Column(
+            "chat_id",
+            sa.Text(),
+            sa.ForeignKey("chat.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column(
+            "file_id",
+            sa.Text(),
+            sa.ForeignKey("file.id", ondelete="CASCADE"),
+            nullable=False,
+        ),
+        sa.Column("message_id", sa.Text(), nullable=True),
+        sa.Column("created_at", sa.BigInteger(), nullable=False),
+        sa.Column("updated_at", sa.BigInteger(), nullable=False),
+        # indexes
+        sa.Index("ix_chat_file_chat_id", "chat_id"),
+        sa.Index("ix_chat_file_file_id", "file_id"),
+        sa.Index("ix_chat_file_message_id", "message_id"),
+        sa.Index("ix_chat_file_user_id", "user_id"),
+        # unique constraints
+        sa.UniqueConstraint(
+            "chat_id", "file_id", name="uq_chat_file_chat_file"
+        ),  # prevent duplicate entries
+    )
+    pass
+
+
+def downgrade() -> None:
+    op.drop_table("chat_file")
+    pass

backend/open_webui/migrations/versions/f1e2d3c4b5a6_add_access_grant_table.py

@@ -0,0 +1,369 @@
+"""Add access_grant table
+
+Revision ID: f1e2d3c4b5a6
+Revises: 8452d01d26d7
+Create Date: 2026-02-05 10:00:00.000000
+
+Migrates from JSON access_control columns to normalized access_grant table.
+Access control semantics:
+- NULL: Public access (all users can read) -> insert user:* for read
+- {}: Private/owner-only (no grants) -> insert nothing
+- {read: {...}, write: {...}}: Custom permissions -> insert specific grants
+"""
+
+from typing import Sequence, Union
+import time
+import uuid
+
+from alembic import op
+import sqlalchemy as sa
+
+from open_webui.migrations.util import get_existing_tables
+
+revision: str = "f1e2d3c4b5a6"
+down_revision: Union[str, None] = "8452d01d26d7"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    existing_tables = set(get_existing_tables())
+
+    # Create access_grant table
+    if "access_grant" not in existing_tables:
+        op.create_table(
+            "access_grant",
+            sa.Column("id", sa.Text(), nullable=False, primary_key=True),
+            sa.Column("resource_type", sa.Text(), nullable=False),
+            sa.Column("resource_id", sa.Text(), nullable=False),
+            sa.Column("principal_type", sa.Text(), nullable=False),
+            sa.Column("principal_id", sa.Text(), nullable=False),
+            sa.Column("permission", sa.Text(), nullable=False),
+            sa.Column("created_at", sa.BigInteger(), nullable=False),
+            sa.UniqueConstraint(
+                "resource_type",
+                "resource_id",
+                "principal_type",
+                "principal_id",
+                "permission",
+                name="uq_access_grant_grant",
+            ),
+        )
+        op.create_index(
+            "idx_access_grant_resource",
+            "access_grant",
+            ["resource_type", "resource_id"],
+        )
+        op.create_index(
+            "idx_access_grant_principal",
+            "access_grant",
+            ["principal_type", "principal_id"],
+        )
+
+    # Backfill existing access_control JSON data
+    conn = op.get_bind()
+
+    # Tables with access_control JSON columns: (table_name, resource_type)
+    resource_tables = [
+        ("knowledge", "knowledge"),
+        ("prompt", "prompt"),
+        ("tool", "tool"),
+        ("model", "model"),
+        ("note", "note"),
+        ("channel", "channel"),
+        ("file", "file"),
+    ]
+
+    now = int(time.time())
+    inserted = set()
+
+    for table_name, resource_type in resource_tables:
+        if table_name not in existing_tables:
+            continue
+
+        # Query all rows
+        try:
+            result = conn.execute(
+                sa.text(f'SELECT id, access_control FROM "{table_name}"')
+            )
+            rows = result.fetchall()
+        except Exception:
+            continue
+
+        for row in rows:
+            resource_id = row[0]
+            access_control_json = row[1]
+
+            # Handle NULL or JSON "null" = public access (user:* for read)
+            # Could be Python None (SQL NULL) or string "null" (JSON null)
+            # EXCEPTION: files with NULL are PRIVATE (owner-only), not public
+            is_null = (
+                access_control_json is None
+                or access_control_json == "null"
+                or (
+                    isinstance(access_control_json, str)
+                    and access_control_json.strip().lower() == "null"
+                )
+            )
+            if is_null:
+                # Files: NULL = private (no entry needed, owner has implicit access)
+                # Other resources: NULL = public (insert user:* for read)
+                if resource_type == "file":
+                    continue  # Private - no entry needed
+
+                key = (resource_type, resource_id, "user", "*", "read")
+                if key not in inserted:
+                    try:
+                        conn.execute(
+                            sa.text("""
+                                INSERT INTO access_grant (id, resource_type, resource_id, principal_type, principal_id, permission, created_at)
+                                VALUES (:id, :resource_type, :resource_id, :principal_type, :principal_id, :permission, :created_at)
+                            """),
+                            {
+                                "id": str(uuid.uuid4()),
+                                "resource_type": resource_type,
+                                "resource_id": resource_id,
+                                "principal_type": "user",
+                                "principal_id": "*",
+                                "permission": "read",
+                                "created_at": now,
+                            },
+                        )
+                        inserted.add(key)
+                    except Exception:
+                        pass
+                continue
+
+            # Handle JSON parsing
+            if isinstance(access_control_json, str):
+                import json
+
+                try:
+                    access_control_json = json.loads(access_control_json)
+                except Exception:
+                    continue
+
+            # Handle {} = private/owner-only - NO entries needed
+            # Owner access is implicit, no grants to store
+            if not access_control_json or not isinstance(access_control_json, dict):
+                continue
+
+            # Check if it's effectively empty (no read/write keys with content)
+            read_data = access_control_json.get("read", {})
+            write_data = access_control_json.get("write", {})
+
+            has_read_grants = read_data.get("group_ids", []) or read_data.get(
+                "user_ids", []
+            )
+            has_write_grants = write_data.get("group_ids", []) or write_data.get(
+                "user_ids", []
+            )
+
+            if not has_read_grants and not has_write_grants:
+                # Empty permissions = private, no grants needed
+                continue
+
+            # Extract permissions and insert into access_grant table
+            for permission in ["read", "write"]:
+                perm_data = access_control_json.get(permission, {})
+                if not perm_data:
+                    continue
+
+                for group_id in perm_data.get("group_ids", []):
+                    key = (resource_type, resource_id, "group", group_id, permission)
+                    if key in inserted:
+                        continue
+                    try:
+                        conn.execute(
+                            sa.text("""
+                                INSERT INTO access_grant (id, resource_type, resource_id, principal_type, principal_id, permission, created_at)
+                                VALUES (:id, :resource_type, :resource_id, :principal_type, :principal_id, :permission, :created_at)
+                            """),
+                            {
+                                "id": str(uuid.uuid4()),
+                                "resource_type": resource_type,
+                                "resource_id": resource_id,
+                                "principal_type": "group",
+                                "principal_id": group_id,
+                                "permission": permission,
+                                "created_at": now,
+                            },
+                        )
+                        inserted.add(key)
+                    except Exception:
+                        pass
+
+                for user_id in perm_data.get("user_ids", []):
+                    key = (resource_type, resource_id, "user", user_id, permission)
+                    if key in inserted:
+                        continue
+                    try:
+                        conn.execute(
+                            sa.text("""
+                                INSERT INTO access_grant (id, resource_type, resource_id, principal_type, principal_id, permission, created_at)
+                                VALUES (:id, :resource_type, :resource_id, :principal_type, :principal_id, :permission, :created_at)
+                            """),
+                            {
+                                "id": str(uuid.uuid4()),
+                                "resource_type": resource_type,
+                                "resource_id": resource_id,
+                                "principal_type": "user",
+                                "principal_id": user_id,
+                                "permission": permission,
+                                "created_at": now,
+                            },
+                        )
+                        inserted.add(key)
+                    except Exception:
+                        pass
+
+    # Drop access_control columns from resource tables
+    for table_name, _ in resource_tables:
+        if table_name not in existing_tables:
+            continue
+        try:
+            with op.batch_alter_table(table_name) as batch:
+                batch.drop_column("access_control")
+        except Exception:
+            pass
+
+
+def downgrade() -> None:
+    import json
+
+    conn = op.get_bind()
+
+    # Resource tables mapping: (table_name, resource_type)
+    resource_tables = [
+        ("knowledge", "knowledge"),
+        ("prompt", "prompt"),
+        ("tool", "tool"),
+        ("model", "model"),
+        ("note", "note"),
+        ("channel", "channel"),
+        ("file", "file"),
+    ]
+
+    # Step 1: Re-add access_control columns to resource tables
+    for table_name, _ in resource_tables:
+        try:
+            with op.batch_alter_table(table_name) as batch:
+                batch.add_column(sa.Column("access_control", sa.JSON(), nullable=True))
+        except Exception:
+            pass
+
+    # Step 2: Query access_grant table and reconstruct JSON for each resource
+    for table_name, resource_type in resource_tables:
+        try:
+            # Get all grants for this resource type
+            result = conn.execute(
+                sa.text("""
+                    SELECT resource_id, principal_type, principal_id, permission
+                    FROM access_grant
+                    WHERE resource_type = :resource_type
+                """),
+                {"resource_type": resource_type},
+            )
+            rows = result.fetchall()
+        except Exception:
+            continue
+
+        # Group by resource_id and reconstruct JSON structure
+        resource_grants = {}
+        for row in rows:
+            resource_id = row[0]
+            principal_type = row[1]
+            principal_id = row[2]
+            permission = row[3]
+
+            if resource_id not in resource_grants:
+                resource_grants[resource_id] = {
+                    "is_public": False,
+                    "read": {"group_ids": [], "user_ids": []},
+                    "write": {"group_ids": [], "user_ids": []},
+                }
+
+            # Handle public access (user:* for read)
+            if (
+                principal_type == "user"
+                and principal_id == "*"
+                and permission == "read"
+            ):
+                resource_grants[resource_id]["is_public"] = True
+                continue
+
+            # Add to appropriate list
+            if permission in ["read", "write"]:
+                if principal_type == "group":
+                    if (
+                        principal_id
+                        not in resource_grants[resource_id][permission]["group_ids"]
+                    ):
+                        resource_grants[resource_id][permission]["group_ids"].append(
+                            principal_id
+                        )
+                elif principal_type == "user":
+                    if (
+                        principal_id
+                        not in resource_grants[resource_id][permission]["user_ids"]
+                    ):
+                        resource_grants[resource_id][permission]["user_ids"].append(
+                            principal_id
+                        )
+
+        # Step 3: Update each resource with reconstructed JSON
+        for resource_id, grants in resource_grants.items():
+            if grants["is_public"]:
+                # Public = NULL
+                access_control_value = None
+            elif (
+                not grants["read"]["group_ids"]
+                and not grants["read"]["user_ids"]
+                and not grants["write"]["group_ids"]
+                and not grants["write"]["user_ids"]
+            ):
+                # No grants = should not happen (would mean no entries), default to {}
+                access_control_value = json.dumps({})
+            else:
+                # Custom permissions
+                access_control_value = json.dumps(
+                    {
+                        "read": grants["read"],
+                        "write": grants["write"],
+                    }
+                )
+
+            try:
+                conn.execute(
+                    sa.text(
+                        f'UPDATE "{table_name}" SET access_control = :access_control WHERE id = :id'
+                    ),
+                    {"access_control": access_control_value, "id": resource_id},
+                )
+            except Exception:
+                pass
+
+        # Step 4: Set all resources WITHOUT entries to private
+        # For files: NULL means private (owner-only), so leave as NULL
+        # For other resources: {} means private, so update to {}
+        if resource_type != "file":
+            try:
+                conn.execute(
+                    sa.text(f"""
+                        UPDATE "{table_name}" 
+                        SET access_control = :private_value
+                        WHERE id NOT IN (
+                            SELECT DISTINCT resource_id FROM access_grant WHERE resource_type = :resource_type
+                        )
+                        AND access_control IS NULL
+                    """),
+                    {"private_value": json.dumps({}), "resource_type": resource_type},
+                )
+            except Exception:
+                pass
+        # For files, NULL stays NULL - no action needed
+
+    # Step 5: Drop the access_grant table
+    op.drop_index("idx_access_grant_principal", table_name="access_grant")
+    op.drop_index("idx_access_grant_resource", table_name="access_grant")
+    op.drop_table("access_grant")

backend/open_webui/models/access_grants.py

@@ -0,0 +1,838 @@
+import logging
+import time
+import uuid
+from typing import Optional
+
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db_context
+
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, Text, UniqueConstraint, or_, and_
+from sqlalchemy.dialects.postgresql import JSONB
+
+log = logging.getLogger(__name__)
+
+
+####################
+# AccessGrant DB Schema
+####################
+
+
+class AccessGrant(Base):
+    __tablename__ = "access_grant"
+
+    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_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)
+
+    __table_args__ = (
+        UniqueConstraint(
+            "resource_type",
+            "resource_id",
+            "principal_type",
+            "principal_id",
+            "permission",
+            name="uq_access_grant_grant",
+        ),
+    )
+
+
+class AccessGrantModel(BaseModel):
+    model_config = ConfigDict(from_attributes=True)
+
+    id: str
+    resource_type: str
+    resource_id: str
+    principal_type: str
+    principal_id: str
+    permission: str
+    created_at: int
+
+
+class AccessGrantResponse(BaseModel):
+    """Slim grant model for API responses — resource context is implicit from the parent."""
+
+    id: str
+    principal_type: str
+    principal_id: str
+    permission: str
+
+    @classmethod
+    def from_grant(cls, grant: "AccessGrantModel") -> "AccessGrantResponse":
+        return cls(
+            id=grant.id,
+            principal_type=grant.principal_type,
+            principal_id=grant.principal_id,
+            permission=grant.permission,
+        )
+
+
+####################
+# Conversion utilities
+####################
+
+
+def access_control_to_grants(
+    resource_type: str,
+    resource_id: str,
+    access_control: Optional[dict],
+) -> list[dict]:
+    """
+    Convert an old-style access_control JSON dict to a flat list of grant dicts.
+
+    Semantics:
+    - None  → public read (user:* read) — except files which are private
+    - {}    → private/owner-only (no grants)
+    - {read: {group_ids, user_ids}, write: {group_ids, user_ids}} → specific grants
+
+    Returns a list of dicts with keys: resource_type, resource_id, principal_type, principal_id, permission
+    """
+    grants = []
+
+    if access_control is None:
+        # NULL → public read (user:* for read)
+        # Exception: files with NULL are private (owner-only), no grants needed
+        if resource_type != "file":
+            grants.append(
+                {
+                    "resource_type": resource_type,
+                    "resource_id": resource_id,
+                    "principal_type": "user",
+                    "principal_id": "*",
+                    "permission": "read",
+                }
+            )
+        return grants
+
+    # {} → private/owner-only, no grants
+    if not access_control:
+        return grants
+
+    # Parse structured permissions
+    for permission in ["read", "write"]:
+        perm_data = access_control.get(permission, {})
+        if not perm_data:
+            continue
+
+        for group_id in perm_data.get("group_ids", []):
+            grants.append(
+                {
+                    "resource_type": resource_type,
+                    "resource_id": resource_id,
+                    "principal_type": "group",
+                    "principal_id": group_id,
+                    "permission": permission,
+                }
+            )
+
+        for user_id in perm_data.get("user_ids", []):
+            grants.append(
+                {
+                    "resource_type": resource_type,
+                    "resource_id": resource_id,
+                    "principal_type": "user",
+                    "principal_id": user_id,
+                    "permission": permission,
+                }
+            )
+
+    return grants
+
+
+def normalize_access_grants(access_grants: Optional[list]) -> list[dict]:
+    """
+    Normalize direct access_grants payloads from API forms.
+
+    Keeps only valid grants and removes duplicates by
+    (principal_type, principal_id, permission).
+    """
+    if not access_grants:
+        return []
+
+    deduped = {}
+    for grant in access_grants:
+        if isinstance(grant, BaseModel):
+            grant = grant.model_dump()
+        if not isinstance(grant, dict):
+            continue
+
+        principal_type = grant.get("principal_type")
+        principal_id = grant.get("principal_id")
+        permission = grant.get("permission")
+
+        if principal_type not in ("user", "group"):
+            continue
+        if permission not in ("read", "write"):
+            continue
+        if not isinstance(principal_id, str) or not principal_id:
+            continue
+
+        key = (principal_type, principal_id, permission)
+        deduped[key] = {
+            "id": (
+                grant.get("id")
+                if isinstance(grant.get("id"), str) and grant.get("id")
+                else str(uuid.uuid4())
+            ),
+            "principal_type": principal_type,
+            "principal_id": principal_id,
+            "permission": permission,
+        }
+
+    return list(deduped.values())
+
+
+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"
+        ):
+            return True
+    return False
+
+
+def grants_to_access_control(grants: list) -> Optional[dict]:
+    """
+    Convert a list of grant objects (AccessGrantModel or AccessGrantResponse)
+    back to the old-style access_control JSON dict for backward compatibility.
+
+    Semantics:
+    - [] (empty) → {} (private/owner-only)
+    - Contains user:*:read → None (public), but write grants are preserved
+    - Otherwise → {read: {group_ids, user_ids}, write: {group_ids, user_ids}}
+
+    Note: "public" (user:*:read) still allows additional write permissions
+    to coexist.  When the wildcard read is present the function returns None
+    for the legacy dict, so callers that need write info should inspect the
+    grants list directly.
+    """
+    if not grants:
+        return {}  # No grants = private/owner-only
+
+    result = {
+        "read": {"group_ids": [], "user_ids": []},
+        "write": {"group_ids": [], "user_ids": []},
+    }
+
+    is_public = False
+    for grant in grants:
+        if (
+            grant.principal_type == "user"
+            and grant.principal_id == "*"
+            and grant.permission == "read"
+        ):
+            is_public = True
+            continue  # Don't add wildcard to user_ids list
+
+        if grant.permission not in ("read", "write"):
+            continue
+
+        if grant.principal_type == "group":
+            if grant.principal_id not in result[grant.permission]["group_ids"]:
+                result[grant.permission]["group_ids"].append(grant.principal_id)
+        elif grant.principal_type == "user":
+            if grant.principal_id not in result[grant.permission]["user_ids"]:
+                result[grant.permission]["user_ids"].append(grant.principal_id)
+
+    if is_public:
+        return None  # Public read access
+
+    return result
+
+
+####################
+# Table Operations
+####################
+
+
+class AccessGrantsTable:
+    def grant_access(
+        self,
+        resource_type: str,
+        resource_id: str,
+        principal_type: str,
+        principal_id: str,
+        permission: str,
+        db: Optional[Session] = None,
+    ) -> Optional[AccessGrantModel]:
+        """Add a single access grant. Idempotent (ignores duplicates)."""
+        with get_db_context(db) as db:
+            # Check for existing grant
+            existing = (
+                db.query(AccessGrant)
+                .filter_by(
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                    principal_type=principal_type,
+                    principal_id=principal_id,
+                    permission=permission,
+                )
+                .first()
+            )
+            if existing:
+                return AccessGrantModel.model_validate(existing)
+
+            grant = AccessGrant(
+                id=str(uuid.uuid4()),
+                resource_type=resource_type,
+                resource_id=resource_id,
+                principal_type=principal_type,
+                principal_id=principal_id,
+                permission=permission,
+                created_at=int(time.time()),
+            )
+            db.add(grant)
+            db.commit()
+            db.refresh(grant)
+            return AccessGrantModel.model_validate(grant)
+
+    def revoke_access(
+        self,
+        resource_type: str,
+        resource_id: str,
+        principal_type: str,
+        principal_id: str,
+        permission: str,
+        db: Optional[Session] = None,
+    ) -> bool:
+        """Remove a single access grant."""
+        with get_db_context(db) as db:
+            deleted = (
+                db.query(AccessGrant)
+                .filter_by(
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                    principal_type=principal_type,
+                    principal_id=principal_id,
+                    permission=permission,
+                )
+                .delete()
+            )
+            db.commit()
+            return deleted > 0
+
+    def revoke_all_access(
+        self,
+        resource_type: str,
+        resource_id: str,
+        db: Optional[Session] = None,
+    ) -> int:
+        """Remove all access grants for a resource."""
+        with get_db_context(db) as db:
+            deleted = (
+                db.query(AccessGrant)
+                .filter_by(
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                )
+                .delete()
+            )
+            db.commit()
+            return deleted
+
+    def set_access_control(
+        self,
+        resource_type: str,
+        resource_id: str,
+        access_control: Optional[dict],
+        db: Optional[Session] = None,
+    ) -> list[AccessGrantModel]:
+        """
+        Replace all grants for a resource from an access_control JSON dict.
+        This is the primary bridge for backward compat with the frontend.
+        """
+        with get_db_context(db) as db:
+            # Delete all existing grants for this resource
+            db.query(AccessGrant).filter_by(
+                resource_type=resource_type,
+                resource_id=resource_id,
+            ).delete()
+
+            # Convert JSON to grant dicts
+            grant_dicts = access_control_to_grants(
+                resource_type, resource_id, access_control
+            )
+
+            # Insert new grants
+            results = []
+            for grant_dict in grant_dicts:
+                grant = AccessGrant(
+                    id=str(uuid.uuid4()),
+                    **grant_dict,
+                    created_at=int(time.time()),
+                )
+                db.add(grant)
+                results.append(grant)
+
+            db.commit()
+
+            return [AccessGrantModel.model_validate(g) for g in results]
+
+    def set_access_grants(
+        self,
+        resource_type: str,
+        resource_id: str,
+        access_grants: Optional[list],
+        db: Optional[Session] = None,
+    ) -> list[AccessGrantModel]:
+        """
+        Replace all grants for a resource from a direct access_grants list.
+        """
+        with get_db_context(db) as db:
+            db.query(AccessGrant).filter_by(
+                resource_type=resource_type,
+                resource_id=resource_id,
+            ).delete()
+
+            normalized_grants = normalize_access_grants(access_grants)
+
+            results = []
+            for grant_dict in normalized_grants:
+                grant = AccessGrant(
+                    id=grant_dict["id"],
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                    principal_type=grant_dict["principal_type"],
+                    principal_id=grant_dict["principal_id"],
+                    permission=grant_dict["permission"],
+                    created_at=int(time.time()),
+                )
+                db.add(grant)
+                results.append(grant)
+
+            db.commit()
+            return [AccessGrantModel.model_validate(g) for g in results]
+
+    def get_access_control(
+        self,
+        resource_type: str,
+        resource_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[dict]:
+        """
+        Reconstruct the old-style access_control JSON dict from grants.
+        For backward compat with the frontend.
+        """
+        with get_db_context(db) as db:
+            grants = (
+                db.query(AccessGrant)
+                .filter_by(
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                )
+                .all()
+            )
+            grant_models = [AccessGrantModel.model_validate(g) for g in grants]
+            return grants_to_access_control(grant_models)
+
+    def get_grants_by_resource(
+        self,
+        resource_type: str,
+        resource_id: str,
+        db: Optional[Session] = None,
+    ) -> list[AccessGrantModel]:
+        """Get all grants for a specific resource."""
+        with get_db_context(db) as db:
+            grants = (
+                db.query(AccessGrant)
+                .filter_by(
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                )
+                .all()
+            )
+            return [AccessGrantModel.model_validate(g) for g in grants]
+
+    def has_access(
+        self,
+        user_id: str,
+        resource_type: str,
+        resource_id: str,
+        permission: str = "read",
+        user_group_ids: Optional[set[str]] = None,
+        db: Optional[Session] = None,
+    ) -> bool:
+        """
+        Check if a user has the specified permission on a resource.
+
+        Access is granted if any of the following is true:
+        - There's a grant for user:* (public) with the requested permission
+        - There's a grant for the specific user with the requested permission
+        - There's a grant for any of the user's groups with the requested permission
+        """
+        with get_db_context(db) as db:
+            # Build conditions for matching grants
+            conditions = [
+                # Public access
+                and_(
+                    AccessGrant.principal_type == "user",
+                    AccessGrant.principal_id == "*",
+                ),
+                # Direct user access
+                and_(
+                    AccessGrant.principal_type == "user",
+                    AccessGrant.principal_id == user_id,
+                ),
+            ]
+
+            # Group access
+            if user_group_ids is None:
+                from open_webui.models.groups import Groups
+
+                user_groups = Groups.get_groups_by_member_id(user_id, db=db)
+                user_group_ids = {group.id for group in user_groups}
+
+            if user_group_ids:
+                conditions.append(
+                    and_(
+                        AccessGrant.principal_type == "group",
+                        AccessGrant.principal_id.in_(user_group_ids),
+                    )
+                )
+
+            exists = (
+                db.query(AccessGrant)
+                .filter(
+                    AccessGrant.resource_type == resource_type,
+                    AccessGrant.resource_id == resource_id,
+                    AccessGrant.permission == permission,
+                    or_(*conditions),
+                )
+                .first()
+            )
+            return exists is not None
+
+    def get_accessible_resource_ids(
+        self,
+        user_id: str,
+        resource_type: str,
+        resource_ids: list[str],
+        permission: str = "read",
+        user_group_ids: Optional[set[str]] = None,
+        db: Optional[Session] = None,
+    ) -> set[str]:
+        """
+        Batch check: return the subset of resource_ids that the user can access.
+
+        This replaces calling has_access() in a loop (N+1) with a single query.
+        """
+        if not resource_ids:
+            return set()
+
+        with get_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 is None:
+                from open_webui.models.groups import Groups
+
+                user_groups = Groups.get_groups_by_member_id(user_id, db=db)
+                user_group_ids = {group.id for group in user_groups}
+
+            if user_group_ids:
+                conditions.append(
+                    and_(
+                        AccessGrant.principal_type == "group",
+                        AccessGrant.principal_id.in_(user_group_ids),
+                    )
+                )
+
+            rows = (
+                db.query(AccessGrant.resource_id)
+                .filter(
+                    AccessGrant.resource_type == resource_type,
+                    AccessGrant.resource_id.in_(resource_ids),
+                    AccessGrant.permission == permission,
+                    or_(*conditions),
+                )
+                .distinct()
+                .all()
+            )
+            return {row[0] for row in rows}
+
+    def get_users_with_access(
+        self,
+        resource_type: str,
+        resource_id: str,
+        permission: str = "read",
+        db: Optional[Session] = None,
+    ) -> list:
+        """
+        Get all users who have the specified permission on a resource.
+        Returns a list of UserModel instances.
+        """
+        from open_webui.models.users import Users, UserModel
+        from open_webui.models.groups import Groups
+
+        with get_db_context(db) as db:
+            grants = (
+                db.query(AccessGrant)
+                .filter_by(
+                    resource_type=resource_type,
+                    resource_id=resource_id,
+                    permission=permission,
+                )
+                .all()
+            )
+
+            # Check for public access
+            for grant in grants:
+                if grant.principal_type == "user" and grant.principal_id == "*":
+                    result = Users.get_users(filter={"roles": ["!pending"]}, db=db)
+                    return result.get("users", [])
+
+            user_ids_with_access = set()
+
+            for grant in grants:
+                if grant.principal_type == "user":
+                    user_ids_with_access.add(grant.principal_id)
+                elif grant.principal_type == "group":
+                    group_user_ids = Groups.get_group_user_ids_by_id(
+                        grant.principal_id, db=db
+                    )
+                    if group_user_ids:
+                        user_ids_with_access.update(group_user_ids)
+
+            if not user_ids_with_access:
+                return []
+
+            return Users.get_users_by_user_ids(list(user_ids_with_access), db=db)
+
+    def has_permission_filter(
+        self,
+        db,
+        query,
+        DocumentModel,
+        filter: dict,
+        resource_type: str,
+        permission: str = "read",
+    ):
+        """
+        Apply access control filtering to a SQLAlchemy query by JOINing with access_grant.
+
+        This replaces the old JSON-column-based filtering with a proper relational JOIN.
+        """
+        group_ids = filter.get("group_ids", [])
+        user_id = filter.get("user_id")
+
+        if permission == "read_only":
+            return self._has_read_only_permission_filter(
+                db, query, DocumentModel, filter, resource_type
+            )
+
+        # Build principal conditions
+        principal_conditions = []
+
+        if group_ids or user_id:
+            # Public access: user:* read
+            principal_conditions.append(
+                and_(
+                    AccessGrant.principal_type == "user",
+                    AccessGrant.principal_id == "*",
+                )
+            )
+
+        if user_id:
+            # Owner always has access
+            principal_conditions.append(DocumentModel.user_id == user_id)
+
+            # Direct user grant
+            principal_conditions.append(
+                and_(
+                    AccessGrant.principal_type == "user",
+                    AccessGrant.principal_id == user_id,
+                )
+            )
+
+        if group_ids:
+            # Group grants
+            principal_conditions.append(
+                and_(
+                    AccessGrant.principal_type == "group",
+                    AccessGrant.principal_id.in_(group_ids),
+                )
+            )
+
+        if not principal_conditions:
+            return query
+
+        # LEFT JOIN access_grant and filter
+        # We use a subquery approach to avoid duplicates from multiple matching grants
+        from sqlalchemy import exists as sa_exists, select
+
+        grant_exists = (
+            select(AccessGrant.id)
+            .where(
+                AccessGrant.resource_type == resource_type,
+                AccessGrant.resource_id == DocumentModel.id,
+                AccessGrant.permission == permission,
+                or_(
+                    and_(
+                        AccessGrant.principal_type == "user",
+                        AccessGrant.principal_id == "*",
+                    ),
+                    *(
+                        [
+                            and_(
+                                AccessGrant.principal_type == "user",
+                                AccessGrant.principal_id == user_id,
+                            )
+                        ]
+                        if user_id
+                        else []
+                    ),
+                    *(
+                        [
+                            and_(
+                                AccessGrant.principal_type == "group",
+                                AccessGrant.principal_id.in_(group_ids),
+                            )
+                        ]
+                        if group_ids
+                        else []
+                    ),
+                ),
+            )
+            .correlate(DocumentModel)
+            .exists()
+        )
+
+        # Owner OR has a matching grant
+        owner_or_grant = [grant_exists]
+        if user_id:
+            owner_or_grant.append(DocumentModel.user_id == user_id)
+
+        query = query.filter(or_(*owner_or_grant))
+        return query
+
+    def _has_read_only_permission_filter(
+        self,
+        db,
+        query,
+        DocumentModel,
+        filter: dict,
+        resource_type: str,
+    ):
+        """
+        Filter for items where user has read BUT NOT write access.
+        Public items are NOT considered read_only.
+        """
+        group_ids = filter.get("group_ids", [])
+        user_id = filter.get("user_id")
+
+        from sqlalchemy import exists as sa_exists, select
+
+        # Has read grant (not public)
+        read_grant_exists = (
+            select(AccessGrant.id)
+            .where(
+                AccessGrant.resource_type == resource_type,
+                AccessGrant.resource_id == DocumentModel.id,
+                AccessGrant.permission == "read",
+                or_(
+                    *(
+                        [
+                            and_(
+                                AccessGrant.principal_type == "user",
+                                AccessGrant.principal_id == user_id,
+                            )
+                        ]
+                        if user_id
+                        else []
+                    ),
+                    *(
+                        [
+                            and_(
+                                AccessGrant.principal_type == "group",
+                                AccessGrant.principal_id.in_(group_ids),
+                            )
+                        ]
+                        if group_ids
+                        else []
+                    ),
+                ),
+            )
+            .correlate(DocumentModel)
+            .exists()
+        )
+
+        # Does NOT have write grant
+        write_grant_exists = (
+            select(AccessGrant.id)
+            .where(
+                AccessGrant.resource_type == resource_type,
+                AccessGrant.resource_id == DocumentModel.id,
+                AccessGrant.permission == "write",
+                or_(
+                    *(
+                        [
+                            and_(
+                                AccessGrant.principal_type == "user",
+                                AccessGrant.principal_id == user_id,
+                            )
+                        ]
+                        if user_id
+                        else []
+                    ),
+                    *(
+                        [
+                            and_(
+                                AccessGrant.principal_type == "group",
+                                AccessGrant.principal_id.in_(group_ids),
+                            )
+                        ]
+                        if group_ids
+                        else []
+                    ),
+                ),
+            )
+            .correlate(DocumentModel)
+            .exists()
+        )
+
+        # Is NOT public
+        public_grant_exists = (
+            select(AccessGrant.id)
+            .where(
+                AccessGrant.resource_type == resource_type,
+                AccessGrant.resource_id == DocumentModel.id,
+                AccessGrant.permission == "read",
+                AccessGrant.principal_type == "user",
+                AccessGrant.principal_id == "*",
+            )
+            .correlate(DocumentModel)
+            .exists()
+        )
+
+        conditions = [read_grant_exists, ~write_grant_exists, ~public_grant_exists]
+
+        # Not owner
+        if user_id:
+            conditions.append(DocumentModel.user_id != user_id)
+
+        query = query.filter(and_(*conditions))
+        return query
+
+
+AccessGrants = AccessGrantsTable()

backend/open_webui/models/auths.py

@@ -2,15 +2,14 @@ import logging
 import uuid
 from typing import Optional
 
-from open_webui.internal.db import Base, get_db
-from open_webui.models.users import UserModel, Users
-from open_webui.env import SRC_LOG_LEVELS
-from pydantic import BaseModel
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_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
-from open_webui.utils.auth import verify_password
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # DB MODEL
@@ -20,7 +19,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 class Auth(Base):
     __tablename__ = "auth"
 
-    id = Column(String, primary_key=True)
+    id = Column(String, primary_key=True, unique=True)
     email = Column(String)
     password = Column(Text)
     active = Column(Boolean)
@@ -47,15 +46,7 @@ class ApiKey(BaseModel):
     api_key: Optional[str] = None
 
 
-class UserResponse(BaseModel):
-    id: str
-    email: str
-    name: str
-    role: str
-    profile_image_url: str
-
-
-class SigninResponse(Token, UserResponse):
+class SigninResponse(Token, UserProfileImageResponse):
     pass
 
 
@@ -84,6 +75,13 @@ class SignupForm(BaseModel):
     password: str
     profile_image_url: Optional[str] = "/user.png"
 
+    @field_validator("profile_image_url")
+    @classmethod
+    def check_profile_image_url(cls, v: Optional[str]) -> Optional[str]:
+        if v is not None:
+            return validate_profile_image_url(v)
+        return v
+
 
 class AddUserForm(SignupForm):
     role: Optional[str] = "pending"
@@ -97,9 +95,10 @@ class AuthsTable:
         name: str,
         profile_image_url: str = "/user.png",
         role: str = "pending",
-        oauth_sub: Optional[str] = None,
+        oauth: Optional[dict] = None,
+        db: Optional[Session] = None,
     ) -> Optional[UserModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             log.info("insert_new_auth")
 
             id = str(uuid.uuid4())
@@ -111,7 +110,7 @@ class AuthsTable:
             db.add(result)
 
             user = Users.insert_new_user(
-                id, name, email, profile_image_url, role, oauth_sub
+                id, name, email, profile_image_url, role, oauth=oauth, db=db
             )
 
             db.commit()
@@ -122,18 +121,20 @@ class AuthsTable:
             else:
                 return None
 
-    def authenticate_user(self, email: str, password: str) -> Optional[UserModel]:
+    def authenticate_user(
+        self, email: str, verify_password: callable, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         log.info(f"authenticate_user: {email}")
 
-        user = Users.get_user_by_email(email)
+        user = Users.get_user_by_email(email, db=db)
         if not user:
             return None
 
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 auth = db.query(Auth).filter_by(id=user.id, active=True).first()
                 if auth:
-                    if verify_password(password, auth.password):
+                    if verify_password(auth.password):
                         return user
                     else:
                         return None
@@ -142,32 +143,45 @@ class AuthsTable:
         except Exception:
             return None
 
-    def authenticate_user_by_api_key(self, api_key: str) -> Optional[UserModel]:
+    def authenticate_user_by_api_key(
+        self, api_key: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         log.info(f"authenticate_user_by_api_key: {api_key}")
         # if no api_key, return None
         if not api_key:
             return None
 
         try:
-            user = Users.get_user_by_api_key(api_key)
+            user = Users.get_user_by_api_key(api_key, db=db)
             return user if user else None
         except Exception:
             return False
 
-    def authenticate_user_by_email(self, email: str) -> Optional[UserModel]:
+    def authenticate_user_by_email(
+        self, email: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         log.info(f"authenticate_user_by_email: {email}")
         try:
-            with get_db() as db:
-                auth = db.query(Auth).filter_by(email=email, active=True).first()
-                if auth:
-                    user = Users.get_user_by_id(auth.id)
-                    return user
+            with get_db_context(db) as db:
+                # Single JOIN query instead of two separate queries
+                result = (
+                    db.query(Auth, User)
+                    .join(User, Auth.id == User.id)
+                    .filter(Auth.email == email, Auth.active == True)
+                    .first()
+                )
+                if result:
+                    _, user = result
+                    return UserModel.model_validate(user)
+                return None
         except Exception:
             return None
 
-    def update_user_password_by_id(self, id: str, new_password: str) -> bool:
+    def update_user_password_by_id(
+        self, id: str, new_password: str, db: Optional[Session] = None
+    ) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 result = (
                     db.query(Auth).filter_by(id=id).update({"password": new_password})
                 )
@@ -176,20 +190,22 @@ class AuthsTable:
         except Exception:
             return False
 
-    def update_email_by_id(self, id: str, email: str) -> bool:
+    def update_email_by_id(
+        self, id: str, email: str, db: Optional[Session] = None
+    ) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 result = db.query(Auth).filter_by(id=id).update({"email": email})
                 db.commit()
                 return True if result == 1 else False
         except Exception:
             return False
 
-    def delete_auth_by_id(self, id: str) -> bool:
+    def delete_auth_by_id(self, id: str, db: Optional[Session] = None) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 # Delete User
-                result = Users.delete_user_by_id(id)
+                result = Users.delete_user_by_id(id, db=db)
 
                 if result:
                     db.query(Auth).filter_by(id=id).delete()

backend/open_webui/models/chat_messages.py

@@ -0,0 +1,644 @@
+import json
+import time
+import uuid
+from typing import Any, Optional
+
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db_context
+
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import (
+    BigInteger,
+    Boolean,
+    Column,
+    ForeignKey,
+    Text,
+    JSON,
+    Index,
+    func,
+)
+
+####################
+# Helpers
+####################
+
+
+def _normalize_timestamp(timestamp: int) -> float:
+    """Normalize and validate timestamp. Returns current time if invalid."""
+    now = time.time()
+
+    # Convert milliseconds to seconds if needed
+    if timestamp > 10_000_000_000:
+        timestamp = timestamp / 1000
+
+    # Validate: must be after 2020 and not in the future (with 1 day tolerance)
+    min_valid = 1577836800  # 2020-01-01 00:00:00 UTC
+    max_valid = now + 86400  # 1 day in the future (clock skew tolerance)
+
+    if timestamp < min_valid or timestamp > max_valid:
+        return now
+
+    return timestamp
+
+
+####################
+# ChatMessage DB Schema
+####################
+
+
+class ChatMessage(Base):
+    __tablename__ = "chat_message"
+
+    # Identity
+    id = Column(Text, primary_key=True)
+    chat_id = Column(
+        Text, ForeignKey("chat.id", ondelete="CASCADE"), nullable=False, index=True
+    )
+    user_id = Column(Text, index=True)
+
+    # Structure
+    role = Column(Text, nullable=False)  # user, assistant, system
+    parent_id = Column(Text, nullable=True)
+
+    # Content
+    content = Column(JSON, nullable=True)  # Can be str or list of blocks
+    output = Column(JSON, nullable=True)
+
+    # Model (for assistant messages)
+    model_id = Column(Text, nullable=True, index=True)
+
+    # Attachments
+    files = Column(JSON, nullable=True)
+    sources = Column(JSON, nullable=True)
+    embeds = Column(JSON, nullable=True)
+
+    # Status
+    done = Column(Boolean, default=True)
+    status_history = Column(JSON, nullable=True)
+    error = Column(JSON, nullable=True)
+
+    # Usage (tokens, timing, etc.)
+    usage = Column(JSON, nullable=True)
+
+    # Timestamps
+    created_at = Column(BigInteger, index=True)
+    updated_at = Column(BigInteger)
+
+    __table_args__ = (
+        Index("chat_message_chat_parent_idx", "chat_id", "parent_id"),
+        Index("chat_message_model_created_idx", "model_id", "created_at"),
+        Index("chat_message_user_created_idx", "user_id", "created_at"),
+    )
+
+
+####################
+# Pydantic Models
+####################
+
+
+class ChatMessageModel(BaseModel):
+    model_config = ConfigDict(from_attributes=True)
+
+    id: str
+    chat_id: str
+    user_id: str
+    role: str
+    parent_id: Optional[str] = None
+    content: Optional[Any] = None  # str or list of blocks
+    output: Optional[list] = None
+    model_id: Optional[str] = None
+    files: Optional[list] = None
+    sources: Optional[list] = None
+    embeds: Optional[list] = None
+    done: bool = True
+    status_history: Optional[list] = None
+    error: Optional[dict | str] = None
+    usage: Optional[dict] = None
+    created_at: int
+    updated_at: int
+
+
+####################
+# Table Operations
+####################
+
+
+class ChatMessageTable:
+    def upsert_message(
+        self,
+        message_id: str,
+        chat_id: str,
+        user_id: str,
+        data: dict,
+        db: Optional[Session] = None,
+    ) -> Optional[ChatMessageModel]:
+        """Insert or update a chat message."""
+        with get_db_context(db) as db:
+            now = int(time.time())
+            timestamp = data.get("timestamp", now)
+
+            # Use composite ID: {chat_id}-{message_id}
+            composite_id = f"{chat_id}-{message_id}"
+
+            existing = db.get(ChatMessage, composite_id)
+            if existing:
+                # Update existing
+                if "role" in data:
+                    existing.role = data["role"]
+                if "parent_id" in data:
+                    existing.parent_id = data.get("parent_id") or data.get("parentId")
+                if "content" in data:
+                    existing.content = data.get("content")
+                if "output" in data:
+                    existing.output = data.get("output")
+                if "model_id" in data or "model" in data:
+                    existing.model_id = data.get("model_id") or data.get("model")
+                if "files" in data:
+                    existing.files = data.get("files")
+                if "sources" in data:
+                    existing.sources = data.get("sources")
+                if "embeds" in data:
+                    existing.embeds = data.get("embeds")
+                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")
+                # Extract usage - check direct field first, then info.usage
+                usage = data.get("usage")
+                if not usage:
+                    info = data.get("info", {})
+                    usage = info.get("usage") if info else None
+                if usage:
+                    existing.usage = usage
+                existing.updated_at = now
+                db.commit()
+                db.refresh(existing)
+                return ChatMessageModel.model_validate(existing)
+            else:
+                # Insert new
+                # Extract usage - check direct field first, then info.usage
+                usage = data.get("usage")
+                if not usage:
+                    info = data.get("info", {})
+                    usage = info.get("usage") if info else None
+                message = ChatMessage(
+                    id=composite_id,
+                    chat_id=chat_id,
+                    user_id=user_id,
+                    role=data.get("role", "user"),
+                    parent_id=data.get("parent_id") or data.get("parentId"),
+                    content=data.get("content"),
+                    output=data.get("output"),
+                    model_id=data.get("model_id") or data.get("model"),
+                    files=data.get("files"),
+                    sources=data.get("sources"),
+                    embeds=data.get("embeds"),
+                    done=data.get("done", True),
+                    status_history=data.get("status_history")
+                    or data.get("statusHistory"),
+                    error=data.get("error"),
+                    usage=usage,
+                    created_at=timestamp,
+                    updated_at=now,
+                )
+                db.add(message)
+                db.commit()
+                db.refresh(message)
+                return ChatMessageModel.model_validate(message)
+
+    def get_message_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[ChatMessageModel]:
+        with get_db_context(db) as db:
+            message = db.get(ChatMessage, id)
+            return ChatMessageModel.model_validate(message) if message else None
+
+    def get_messages_by_chat_id(
+        self, chat_id: str, db: Optional[Session] = None
+    ) -> list[ChatMessageModel]:
+        with get_db_context(db) as db:
+            messages = (
+                db.query(ChatMessage)
+                .filter_by(chat_id=chat_id)
+                .order_by(ChatMessage.created_at.asc())
+                .all()
+            )
+            return [ChatMessageModel.model_validate(message) for message in messages]
+
+    def get_messages_by_user_id(
+        self,
+        user_id: str,
+        skip: int = 0,
+        limit: int = 50,
+        db: Optional[Session] = None,
+    ) -> list[ChatMessageModel]:
+        with get_db_context(db) as db:
+            messages = (
+                db.query(ChatMessage)
+                .filter_by(user_id=user_id)
+                .order_by(ChatMessage.created_at.desc())
+                .offset(skip)
+                .limit(limit)
+                .all()
+            )
+            return [ChatMessageModel.model_validate(message) for message in messages]
+
+    def get_messages_by_model_id(
+        self,
+        model_id: str,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        skip: int = 0,
+        limit: int = 100,
+        db: Optional[Session] = None,
+    ) -> list[ChatMessageModel]:
+        with get_db_context(db) as db:
+            query = db.query(ChatMessage).filter_by(model_id=model_id)
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+            messages = (
+                query.order_by(ChatMessage.created_at.desc())
+                .offset(skip)
+                .limit(limit)
+                .all()
+            )
+            return [ChatMessageModel.model_validate(message) for message in messages]
+
+    def get_chat_ids_by_model_id(
+        self,
+        model_id: str,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        skip: int = 0,
+        limit: int = 50,
+        db: Optional[Session] = None,
+    ) -> list[str]:
+        """Get distinct chat_ids that used a specific model."""
+
+        with get_db_context(db) as db:
+            query = db.query(
+                ChatMessage.chat_id,
+                func.max(ChatMessage.created_at).label("last_message_at"),
+            ).filter(ChatMessage.model_id == model_id)
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+
+            # Group by chat_id and order by most recent message in each chat
+            chat_ids = (
+                query.group_by(ChatMessage.chat_id)
+                .order_by(func.max(ChatMessage.created_at).desc())
+                .offset(skip)
+                .limit(limit)
+                .all()
+            )
+            return [chat_id for chat_id, _ in chat_ids]
+
+    def delete_messages_by_chat_id(
+        self, chat_id: str, db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
+            db.query(ChatMessage).filter_by(chat_id=chat_id).delete()
+            db.commit()
+            return True
+
+    # Analytics methods
+    def get_message_count_by_model(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        group_id: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, int]:
+        with get_db_context(db) as db:
+            from sqlalchemy import func
+            from open_webui.models.groups import GroupMember
+
+            query = db.query(
+                ChatMessage.model_id, func.count(ChatMessage.id).label("count")
+            ).filter(
+                ChatMessage.role == "assistant",
+                ChatMessage.model_id.isnot(None),
+                ~ChatMessage.user_id.like("shared-%"),
+            )
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+            if group_id:
+                group_users = (
+                    db.query(GroupMember.user_id)
+                    .filter(GroupMember.group_id == group_id)
+                    .subquery()
+                )
+                query = query.filter(ChatMessage.user_id.in_(group_users))
+
+            results = query.group_by(ChatMessage.model_id).all()
+            return {row.model_id: row.count for row in results}
+
+    def get_token_usage_by_model(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        group_id: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, dict]:
+        """Aggregate token usage by model using database-level aggregation."""
+        with get_db_context(db) as db:
+            from sqlalchemy import func, cast, Integer
+            from open_webui.models.groups import GroupMember
+
+            dialect = db.bind.dialect.name
+
+            if dialect == "sqlite":
+                input_tokens = cast(
+                    func.json_extract(ChatMessage.usage, "$.input_tokens"), Integer
+                )
+                output_tokens = cast(
+                    func.json_extract(ChatMessage.usage, "$.output_tokens"), Integer
+                )
+            elif dialect == "postgresql":
+                # Use json_extract_path_text for PostgreSQL JSON columns
+                input_tokens = cast(
+                    func.json_extract_path_text(ChatMessage.usage, "input_tokens"),
+                    Integer,
+                )
+                output_tokens = cast(
+                    func.json_extract_path_text(ChatMessage.usage, "output_tokens"),
+                    Integer,
+                )
+            else:
+                raise NotImplementedError(f"Unsupported dialect: {dialect}")
+
+            query = db.query(
+                ChatMessage.model_id,
+                func.coalesce(func.sum(input_tokens), 0).label("input_tokens"),
+                func.coalesce(func.sum(output_tokens), 0).label("output_tokens"),
+                func.count(ChatMessage.id).label("message_count"),
+            ).filter(
+                ChatMessage.role == "assistant",
+                ChatMessage.model_id.isnot(None),
+                ChatMessage.usage.isnot(None),
+                ~ChatMessage.user_id.like("shared-%"),
+            )
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+            if group_id:
+                group_users = (
+                    db.query(GroupMember.user_id)
+                    .filter(GroupMember.group_id == group_id)
+                    .subquery()
+                )
+                query = query.filter(ChatMessage.user_id.in_(group_users))
+
+            results = query.group_by(ChatMessage.model_id).all()
+
+            return {
+                row.model_id: {
+                    "input_tokens": row.input_tokens,
+                    "output_tokens": row.output_tokens,
+                    "total_tokens": row.input_tokens + row.output_tokens,
+                    "message_count": row.message_count,
+                }
+                for row in results
+            }
+
+    def get_token_usage_by_user(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, dict]:
+        """Aggregate token usage by user using database-level aggregation."""
+        with get_db_context(db) as db:
+            from sqlalchemy import func, cast, Integer
+
+            dialect = db.bind.dialect.name
+
+            if dialect == "sqlite":
+                input_tokens = cast(
+                    func.json_extract(ChatMessage.usage, "$.input_tokens"), Integer
+                )
+                output_tokens = cast(
+                    func.json_extract(ChatMessage.usage, "$.output_tokens"), Integer
+                )
+            elif dialect == "postgresql":
+                # Use json_extract_path_text for PostgreSQL JSON columns
+                input_tokens = cast(
+                    func.json_extract_path_text(ChatMessage.usage, "input_tokens"),
+                    Integer,
+                )
+                output_tokens = cast(
+                    func.json_extract_path_text(ChatMessage.usage, "output_tokens"),
+                    Integer,
+                )
+            else:
+                raise NotImplementedError(f"Unsupported dialect: {dialect}")
+
+            query = db.query(
+                ChatMessage.user_id,
+                func.coalesce(func.sum(input_tokens), 0).label("input_tokens"),
+                func.coalesce(func.sum(output_tokens), 0).label("output_tokens"),
+                func.count(ChatMessage.id).label("message_count"),
+            ).filter(
+                ChatMessage.role == "assistant",
+                ChatMessage.user_id.isnot(None),
+                ChatMessage.usage.isnot(None),
+                ~ChatMessage.user_id.like("shared-%"),
+            )
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+
+            results = query.group_by(ChatMessage.user_id).all()
+
+            return {
+                row.user_id: {
+                    "input_tokens": row.input_tokens,
+                    "output_tokens": row.output_tokens,
+                    "total_tokens": row.input_tokens + row.output_tokens,
+                    "message_count": row.message_count,
+                }
+                for row in results
+            }
+
+    def get_message_count_by_user(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        group_id: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, int]:
+        with get_db_context(db) as db:
+            from sqlalchemy import func
+            from open_webui.models.groups import GroupMember
+
+            query = db.query(
+                ChatMessage.user_id, func.count(ChatMessage.id).label("count")
+            ).filter(~ChatMessage.user_id.like("shared-%"))
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+            if group_id:
+                group_users = (
+                    db.query(GroupMember.user_id)
+                    .filter(GroupMember.group_id == group_id)
+                    .subquery()
+                )
+                query = query.filter(ChatMessage.user_id.in_(group_users))
+
+            results = query.group_by(ChatMessage.user_id).all()
+            return {row.user_id: row.count for row in results}
+
+    def get_message_count_by_chat(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        group_id: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, int]:
+        with get_db_context(db) as db:
+            from sqlalchemy import func
+            from open_webui.models.groups import GroupMember
+
+            query = db.query(
+                ChatMessage.chat_id, func.count(ChatMessage.id).label("count")
+            ).filter(~ChatMessage.user_id.like("shared-%"))
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+            if group_id:
+                group_users = (
+                    db.query(GroupMember.user_id)
+                    .filter(GroupMember.group_id == group_id)
+                    .subquery()
+                )
+                query = query.filter(ChatMessage.user_id.in_(group_users))
+
+            results = query.group_by(ChatMessage.chat_id).all()
+            return {row.chat_id: row.count for row in results}
+
+    def get_daily_message_counts_by_model(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        group_id: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, dict[str, int]]:
+        """Get message counts grouped by day and model."""
+        with get_db_context(db) as db:
+            from datetime import datetime, timedelta
+            from open_webui.models.groups import GroupMember
+
+            query = db.query(ChatMessage.created_at, ChatMessage.model_id).filter(
+                ChatMessage.role == "assistant",
+                ChatMessage.model_id.isnot(None),
+                ~ChatMessage.user_id.like("shared-%"),
+            )
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+            if group_id:
+                group_users = (
+                    db.query(GroupMember.user_id)
+                    .filter(GroupMember.group_id == group_id)
+                    .subquery()
+                )
+                query = query.filter(ChatMessage.user_id.in_(group_users))
+
+            results = query.all()
+
+            # Group by date -> model -> count
+            daily_counts: dict[str, dict[str, int]] = {}
+            for timestamp, model_id in results:
+                date_str = datetime.fromtimestamp(
+                    _normalize_timestamp(timestamp)
+                ).strftime("%Y-%m-%d")
+                if date_str not in daily_counts:
+                    daily_counts[date_str] = {}
+                daily_counts[date_str][model_id] = (
+                    daily_counts[date_str].get(model_id, 0) + 1
+                )
+
+            # Fill in missing days
+            if start_date and end_date:
+                current = datetime.fromtimestamp(_normalize_timestamp(start_date))
+                end_dt = datetime.fromtimestamp(_normalize_timestamp(end_date))
+                while current <= end_dt:
+                    date_str = current.strftime("%Y-%m-%d")
+                    if date_str not in daily_counts:
+                        daily_counts[date_str] = {}
+                    current += timedelta(days=1)
+
+            return daily_counts
+
+    def get_hourly_message_counts_by_model(
+        self,
+        start_date: Optional[int] = None,
+        end_date: Optional[int] = None,
+        db: Optional[Session] = None,
+    ) -> dict[str, dict[str, int]]:
+        """Get message counts grouped by hour and model."""
+        with get_db_context(db) as db:
+            from datetime import datetime, timedelta
+
+            query = db.query(ChatMessage.created_at, ChatMessage.model_id).filter(
+                ChatMessage.role == "assistant",
+                ChatMessage.model_id.isnot(None),
+                ~ChatMessage.user_id.like("shared-%"),
+            )
+
+            if start_date:
+                query = query.filter(ChatMessage.created_at >= start_date)
+            if end_date:
+                query = query.filter(ChatMessage.created_at <= end_date)
+
+            results = query.all()
+
+            # Group by hour -> model -> count
+            hourly_counts: dict[str, dict[str, int]] = {}
+            for timestamp, model_id in results:
+                hour_str = datetime.fromtimestamp(
+                    _normalize_timestamp(timestamp)
+                ).strftime("%Y-%m-%d %H:00")
+                if hour_str not in hourly_counts:
+                    hourly_counts[hour_str] = {}
+                hourly_counts[hour_str][model_id] = (
+                    hourly_counts[hour_str].get(model_id, 0) + 1
+                )
+
+            # Fill in missing hours
+            if start_date and end_date:
+                current = datetime.fromtimestamp(
+                    _normalize_timestamp(start_date)
+                ).replace(minute=0, second=0, microsecond=0)
+                end_dt = datetime.fromtimestamp(_normalize_timestamp(end_date))
+                while current <= end_dt:
+                    hour_str = current.strftime("%Y-%m-%d %H:00")
+                    if hour_str not in hourly_counts:
+                        hourly_counts[hour_str] = {}
+                    current += timedelta(hours=1)
+
+            return hourly_counts
+
+
+ChatMessages = ChatMessageTable()

backend/open_webui/models/feedbacks.py

@@ -3,15 +3,14 @@ import time
 import uuid
 from typing import Optional
 
-from open_webui.internal.db import Base, get_db
-from open_webui.models.chats import Chats
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
+from open_webui.models.users import User
 
-from open_webui.env import SRC_LOG_LEVELS
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Column, Text, JSON, Boolean
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 
 ####################
@@ -21,7 +20,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 class Feedback(Base):
     __tablename__ = "feedback"
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
     user_id = Column(Text)
     version = Column(BigInteger, default=0)
     type = Column(Text)
@@ -62,6 +61,20 @@ class FeedbackResponse(BaseModel):
     updated_at: int
 
 
+class FeedbackIdResponse(BaseModel):
+    id: str
+    user_id: str
+    created_at: int
+    updated_at: int
+
+
+class LeaderboardFeedbackData(BaseModel):
+    """Minimal feedback data for leaderboard computation (excludes snapshot/meta)."""
+
+    id: str
+    data: Optional[dict] = None
+
+
 class RatingData(BaseModel):
     rating: Optional[str | int] = None
     model_id: Optional[str] = None
@@ -92,11 +105,44 @@ class FeedbackForm(BaseModel):
     model_config = ConfigDict(extra="allow")
 
 
+class UserResponse(BaseModel):
+    id: str
+    name: str
+    email: str
+    role: str = "pending"
+
+    last_active_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
+    created_at: int  # timestamp in epoch
+
+    model_config = ConfigDict(from_attributes=True)
+
+
+class FeedbackUserResponse(FeedbackResponse):
+    user: Optional[UserResponse] = None
+
+
+class FeedbackListResponse(BaseModel):
+    items: list[FeedbackUserResponse]
+    total: int
+
+
+class ModelHistoryEntry(BaseModel):
+    date: str
+    won: int
+    lost: int
+
+
+class ModelHistoryResponse(BaseModel):
+    model_id: str
+    history: list[ModelHistoryEntry]
+
+
 class FeedbackTable:
     def insert_new_feedback(
-        self, user_id: str, form_data: FeedbackForm
+        self, user_id: str, form_data: FeedbackForm, db: Optional[Session] = None
     ) -> Optional[FeedbackModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             id = str(uuid.uuid4())
             feedback = FeedbackModel(
                 **{
@@ -121,9 +167,11 @@ class FeedbackTable:
                 log.exception(f"Error creating a new feedback: {e}")
                 return None
 
-    def get_feedback_by_id(self, id: str) -> Optional[FeedbackModel]:
+    def get_feedback_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[FeedbackModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 feedback = db.query(Feedback).filter_by(id=id).first()
                 if not feedback:
                     return None
@@ -132,10 +180,10 @@ class FeedbackTable:
             return None
 
     def get_feedback_by_id_and_user_id(
-        self, id: str, user_id: str
+        self, id: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[FeedbackModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 feedback = db.query(Feedback).filter_by(id=id, user_id=user_id).first()
                 if not feedback:
                     return None
@@ -143,8 +191,93 @@ class FeedbackTable:
         except Exception:
             return None
 
-    def get_all_feedbacks(self) -> list[FeedbackModel]:
-        with get_db() as db:
+    def get_feedbacks_by_chat_id(
+        self, chat_id: str, db: Optional[Session] = None
+    ) -> list[FeedbackModel]:
+        """Get all feedbacks for a specific chat."""
+        try:
+            with get_db_context(db) as db:
+                # meta.chat_id stores the chat reference
+                feedbacks = (
+                    db.query(Feedback)
+                    .filter(Feedback.meta["chat_id"].as_string() == chat_id)
+                    .order_by(Feedback.created_at.desc())
+                    .all()
+                )
+                return [FeedbackModel.model_validate(fb) for fb in feedbacks]
+        except Exception:
+            return []
+
+    def get_feedback_items(
+        self,
+        filter: dict = {},
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> FeedbackListResponse:
+        with get_db_context(db) as db:
+            query = db.query(Feedback, User).join(User, Feedback.user_id == User.id)
+
+            if filter:
+                order_by = filter.get("order_by")
+                direction = filter.get("direction")
+
+                if order_by == "username":
+                    if direction == "asc":
+                        query = query.order_by(User.name.asc())
+                    else:
+                        query = query.order_by(User.name.desc())
+                elif order_by == "model_id":
+                    # it's stored in feedback.data['model_id']
+                    if direction == "asc":
+                        query = query.order_by(
+                            Feedback.data["model_id"].as_string().asc()
+                        )
+                    else:
+                        query = query.order_by(
+                            Feedback.data["model_id"].as_string().desc()
+                        )
+                elif order_by == "rating":
+                    # it's stored in feedback.data['rating']
+                    if direction == "asc":
+                        query = query.order_by(
+                            Feedback.data["rating"].as_string().asc()
+                        )
+                    else:
+                        query = query.order_by(
+                            Feedback.data["rating"].as_string().desc()
+                        )
+                elif order_by == "updated_at":
+                    if direction == "asc":
+                        query = query.order_by(Feedback.updated_at.asc())
+                    else:
+                        query = query.order_by(Feedback.updated_at.desc())
+
+            else:
+                query = query.order_by(Feedback.created_at.desc())
+
+            # Count BEFORE pagination
+            total = query.count()
+
+            if skip:
+                query = query.offset(skip)
+            if limit:
+                query = query.limit(limit)
+
+            items = query.all()
+
+            feedbacks = []
+            for feedback, user in items:
+                feedback_model = FeedbackModel.model_validate(feedback)
+                user_model = UserResponse.model_validate(user)
+                feedbacks.append(
+                    FeedbackUserResponse(**feedback_model.model_dump(), user=user_model)
+                )
+
+            return FeedbackListResponse(items=feedbacks, total=total)
+
+    def get_all_feedbacks(self, db: Optional[Session] = None) -> list[FeedbackModel]:
+        with get_db_context(db) as db:
             return [
                 FeedbackModel.model_validate(feedback)
                 for feedback in db.query(Feedback)
@@ -152,8 +285,110 @@ class FeedbackTable:
                 .all()
             ]
 
-    def get_feedbacks_by_type(self, type: str) -> list[FeedbackModel]:
-        with get_db() as db:
+    def get_all_feedback_ids(
+        self, db: Optional[Session] = None
+    ) -> list[FeedbackIdResponse]:
+        with get_db_context(db) as db:
+            return [
+                FeedbackIdResponse(
+                    id=row.id,
+                    user_id=row.user_id,
+                    created_at=row.created_at,
+                    updated_at=row.updated_at,
+                )
+                for row in db.query(
+                    Feedback.id,
+                    Feedback.user_id,
+                    Feedback.created_at,
+                    Feedback.updated_at,
+                )
+                .order_by(Feedback.updated_at.desc())
+                .all()
+            ]
+
+    def get_feedbacks_for_leaderboard(
+        self, db: Optional[Session] = None
+    ) -> list[LeaderboardFeedbackData]:
+        """Fetch only id and data for leaderboard computation (excludes snapshot/meta)."""
+        with get_db_context(db) as db:
+            return [
+                LeaderboardFeedbackData(id=row.id, data=row.data)
+                for row in db.query(Feedback.id, Feedback.data).all()
+            ]
+
+    def get_model_evaluation_history(
+        self, model_id: str, days: int = 30, db: Optional[Session] = None
+    ) -> list[ModelHistoryEntry]:
+        """
+        Get daily wins/losses for a specific model over the past N days.
+        If days=0, returns all time data starting from first feedback.
+        Returns: [{"date": "2026-01-08", "won": 5, "lost": 2}, ...]
+        """
+        from datetime import datetime, timedelta
+        from collections import defaultdict
+
+        with get_db_context(db) as db:
+            if days == 0:
+                # All time - no cutoff
+                rows = db.query(Feedback.created_at, Feedback.data).all()
+            else:
+                cutoff = int(time.time()) - (days * 86400)
+                rows = (
+                    db.query(Feedback.created_at, Feedback.data)
+                    .filter(Feedback.created_at >= cutoff)
+                    .all()
+                )
+
+        daily_counts = defaultdict(lambda: {"won": 0, "lost": 0})
+        first_date = None
+
+        for created_at, data in rows:
+            if not data:
+                continue
+            if data.get("model_id") != model_id:
+                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
+
+            # Track first date for this model
+            if first_date is None or date_str < first_date:
+                first_date = date_str
+
+        # Generate date range
+        result = []
+        today = datetime.now().date()
+
+        if days == 0 and first_date:
+            # All time: start from first feedback date
+            start_date = datetime.strptime(first_date, "%Y-%m-%d").date()
+            num_days = (today - start_date).days + 1
+        else:
+            # Fixed range
+            num_days = days
+            start_date = today - timedelta(days=days - 1)
+
+        for i in range(num_days):
+            d = start_date + timedelta(days=i)
+            date_str = d.strftime("%Y-%m-%d")
+            counts = daily_counts.get(date_str, {"won": 0, "lost": 0})
+            result.append(
+                ModelHistoryEntry(date=date_str, won=counts["won"], lost=counts["lost"])
+            )
+
+        return result
+
+    def get_feedbacks_by_type(
+        self, type: str, db: Optional[Session] = None
+    ) -> list[FeedbackModel]:
+        with get_db_context(db) as db:
             return [
                 FeedbackModel.model_validate(feedback)
                 for feedback in db.query(Feedback)
@@ -162,8 +397,10 @@ class FeedbackTable:
                 .all()
             ]
 
-    def get_feedbacks_by_user_id(self, user_id: str) -> list[FeedbackModel]:
-        with get_db() as db:
+    def get_feedbacks_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> list[FeedbackModel]:
+        with get_db_context(db) as db:
             return [
                 FeedbackModel.model_validate(feedback)
                 for feedback in db.query(Feedback)
@@ -173,9 +410,9 @@ class FeedbackTable:
             ]
 
     def update_feedback_by_id(
-        self, id: str, form_data: FeedbackForm
+        self, id: str, form_data: FeedbackForm, db: Optional[Session] = None
     ) -> Optional[FeedbackModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             feedback = db.query(Feedback).filter_by(id=id).first()
             if not feedback:
                 return None
@@ -193,9 +430,13 @@ class FeedbackTable:
             return FeedbackModel.model_validate(feedback)
 
     def update_feedback_by_id_and_user_id(
-        self, id: str, user_id: str, form_data: FeedbackForm
+        self,
+        id: str,
+        user_id: str,
+        form_data: FeedbackForm,
+        db: Optional[Session] = None,
     ) -> Optional[FeedbackModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             feedback = db.query(Feedback).filter_by(id=id, user_id=user_id).first()
             if not feedback:
                 return None
@@ -212,8 +453,8 @@ class FeedbackTable:
             db.commit()
             return FeedbackModel.model_validate(feedback)
 
-    def delete_feedback_by_id(self, id: str) -> bool:
-        with get_db() as db:
+    def delete_feedback_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             feedback = db.query(Feedback).filter_by(id=id).first()
             if not feedback:
                 return False
@@ -221,8 +462,10 @@ class FeedbackTable:
             db.commit()
             return True
 
-    def delete_feedback_by_id_and_user_id(self, id: str, user_id: str) -> bool:
-        with get_db() as db:
+    def delete_feedback_by_id_and_user_id(
+        self, id: str, user_id: str, db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
             feedback = db.query(Feedback).filter_by(id=id, user_id=user_id).first()
             if not feedback:
                 return False
@@ -230,25 +473,19 @@ class FeedbackTable:
             db.commit()
             return True
 
-    def delete_feedbacks_by_user_id(self, user_id: str) -> bool:
-        with get_db() as db:
-            feedbacks = db.query(Feedback).filter_by(user_id=user_id).all()
-            if not feedbacks:
-                return False
-            for feedback in feedbacks:
-                db.delete(feedback)
+    def delete_feedbacks_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
+            result = db.query(Feedback).filter_by(user_id=user_id).delete()
             db.commit()
-            return True
+            return result > 0
 
-    def delete_all_feedbacks(self) -> bool:
-        with get_db() as db:
-            feedbacks = db.query(Feedback).all()
-            if not feedbacks:
-                return False
-            for feedback in feedbacks:
-                db.delete(feedback)
+    def delete_all_feedbacks(self, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
+            result = db.query(Feedback).delete()
             db.commit()
-            return True
+            return result > 0
 
 
 Feedbacks = FeedbackTable()

backend/open_webui/models/files.py

@@ -2,13 +2,12 @@ import logging
 import time
 from typing import Optional
 
-from open_webui.internal.db import Base, JSONField, get_db
-from open_webui.env import SRC_LOG_LEVELS
-from pydantic import BaseModel, ConfigDict
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
+from pydantic import BaseModel, ConfigDict, model_validator
 from sqlalchemy import BigInteger, Column, String, Text, JSON
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # Files DB Schema
@@ -17,7 +16,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 class File(Base):
     __tablename__ = "file"
-    id = Column(String, primary_key=True)
+    id = Column(String, primary_key=True, unique=True)
     user_id = Column(String)
     hash = Column(Text, nullable=True)
 
@@ -27,8 +26,6 @@ class File(Base):
     data = Column(JSON, nullable=True)
     meta = Column(JSON, nullable=True)
 
-    access_control = Column(JSON, nullable=True)
-
     created_at = Column(BigInteger)
     updated_at = Column(BigInteger)
 
@@ -46,8 +43,6 @@ class FileModel(BaseModel):
     data: Optional[dict] = None
     meta: Optional[dict] = None
 
-    access_control: Optional[dict] = None
-
     created_at: Optional[int]  # timestamp in epoch
     updated_at: Optional[int]  # timestamp in epoch
 
@@ -64,6 +59,25 @@ class FileMeta(BaseModel):
 
     model_config = ConfigDict(extra="allow")
 
+    @model_validator(mode="before")
+    @classmethod
+    def sanitize_meta(cls, data):
+        """Sanitize metadata fields to handle malformed legacy data."""
+        if not isinstance(data, dict):
+            return data
+
+        # Handle content_type that may be a list like ['application/pdf', None]
+        content_type = data.get("content_type")
+        if isinstance(content_type, list):
+            # Extract first non-None string value
+            data["content_type"] = next(
+                (item for item in content_type if isinstance(item, str)), None
+            )
+        elif content_type is not None and not isinstance(content_type, str):
+            data["content_type"] = None
+
+        return data
+
 
 class FileModelResponse(BaseModel):
     id: str
@@ -75,14 +89,15 @@ class FileModelResponse(BaseModel):
     meta: FileMeta
 
     created_at: int  # timestamp in epoch
-    updated_at: int  # timestamp in epoch
+    updated_at: Optional[int] = None  # timestamp in epoch, optional for legacy files
 
     model_config = ConfigDict(extra="allow")
 
 
 class FileMetadataResponse(BaseModel):
     id: str
-    meta: dict
+    hash: Optional[str] = None
+    meta: Optional[dict] = None
     created_at: int  # timestamp in epoch
     updated_at: int  # timestamp in epoch
 
@@ -94,12 +109,24 @@ class FileForm(BaseModel):
     path: str
     data: dict = {}
     meta: dict = {}
-    access_control: Optional[dict] = None
+
+
+class FileUpdateForm(BaseModel):
+    hash: Optional[str] = None
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+
+
+class FileListResponse(BaseModel):
+    items: list[FileModel]
+    total: int
 
 
 class FilesTable:
-    def insert_new_file(self, user_id: str, form_data: FileForm) -> Optional[FileModel]:
-        with get_db() as db:
+    def insert_new_file(
+        self, user_id: str, form_data: FileForm, db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        with get_db_context(db) as db:
             file = FileModel(
                 **{
                     **form_data.model_dump(),
@@ -122,16 +149,23 @@ class FilesTable:
                 log.exception(f"Error inserting a new file: {e}")
                 return None
 
-    def get_file_by_id(self, id: str) -> Optional[FileModel]:
-        with get_db() as db:
-            try:
-                file = db.get(File, id)
-                return FileModel.model_validate(file)
-            except Exception:
-                return None
+    def get_file_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        try:
+            with get_db_context(db) as db:
+                try:
+                    file = db.get(File, id)
+                    return FileModel.model_validate(file)
+                except Exception:
+                    return None
+        except Exception:
+            return None
 
-    def get_file_by_id_and_user_id(self, id: str, user_id: str) -> Optional[FileModel]:
-        with get_db() as db:
+    def get_file_by_id_and_user_id(
+        self, id: str, user_id: str, db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        with get_db_context(db) as db:
             try:
                 file = db.query(File).filter_by(id=id, user_id=user_id).first()
                 if file:
@@ -141,12 +175,15 @@ class FilesTable:
             except Exception:
                 return None
 
-    def get_file_metadata_by_id(self, id: str) -> Optional[FileMetadataResponse]:
-        with get_db() as db:
+    def get_file_metadata_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[FileMetadataResponse]:
+        with get_db_context(db) as db:
             try:
                 file = db.get(File, id)
                 return FileMetadataResponse(
                     id=file.id,
+                    hash=file.hash,
                     meta=file.meta,
                     created_at=file.created_at,
                     updated_at=file.updated_at,
@@ -154,12 +191,14 @@ class FilesTable:
             except Exception:
                 return None
 
-    def get_files(self) -> list[FileModel]:
-        with get_db() as db:
+    def get_files(self, db: Optional[Session] = None) -> list[FileModel]:
+        with get_db_context(db) as db:
             return [FileModel.model_validate(file) for file in db.query(File).all()]
 
-    def check_access_by_user_id(self, id, user_id, permission="write") -> bool:
-        file = self.get_file_by_id(id)
+    def check_access_by_user_id(
+        self, id, user_id, permission="write", db: Optional[Session] = None
+    ) -> bool:
+        file = self.get_file_by_id(id, db=db)
         if not file:
             return False
         if file.user_id == user_id:
@@ -167,8 +206,10 @@ class FilesTable:
         # Implement additional access control logic here as needed
         return False
 
-    def get_files_by_ids(self, ids: list[str]) -> list[FileModel]:
-        with get_db() as db:
+    def get_files_by_ids(
+        self, ids: list[str], db: Optional[Session] = None
+    ) -> list[FileModel]:
+        with get_db_context(db) as db:
             return [
                 FileModel.model_validate(file)
                 for file in db.query(File)
@@ -177,64 +218,165 @@ class FilesTable:
                 .all()
             ]
 
-    def get_file_metadatas_by_ids(self, ids: list[str]) -> list[FileMetadataResponse]:
-        with get_db() as db:
+    def get_file_metadatas_by_ids(
+        self, ids: list[str], db: Optional[Session] = None
+    ) -> list[FileMetadataResponse]:
+        with get_db_context(db) as db:
             return [
                 FileMetadataResponse(
                     id=file.id,
+                    hash=file.hash,
                     meta=file.meta,
                     created_at=file.created_at,
                     updated_at=file.updated_at,
                 )
                 for file in db.query(
-                    File.id, File.meta, File.created_at, File.updated_at
+                    File.id, File.hash, File.meta, File.created_at, File.updated_at
                 )
                 .filter(File.id.in_(ids))
                 .order_by(File.updated_at.desc())
                 .all()
             ]
 
-    def get_files_by_user_id(self, user_id: str) -> list[FileModel]:
-        with get_db() as db:
+    def get_files_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> list[FileModel]:
+        with get_db_context(db) as db:
             return [
                 FileModel.model_validate(file)
                 for file in db.query(File).filter_by(user_id=user_id).all()
             ]
 
-    def update_file_hash_by_id(self, id: str, hash: str) -> Optional[FileModel]:
-        with get_db() as db:
+    @staticmethod
+    def _glob_to_like_pattern(glob: str) -> str:
+        """
+        Convert a glob/fnmatch pattern to a SQL LIKE pattern.
+
+        Escapes SQL special characters and converts glob wildcards:
+        - `*` becomes `%` (match any sequence of characters)
+        - `?` becomes `_` (match exactly one character)
+
+        Args:
+            glob: A glob pattern (e.g., "*.txt", "file?.doc")
+
+        Returns:
+            A SQL LIKE compatible pattern with proper escaping.
+        """
+        # Escape SQL special characters first, then convert glob wildcards
+        pattern = glob.replace("\\", "\\\\")
+        pattern = pattern.replace("%", "\\%")
+        pattern = pattern.replace("_", "\\_")
+        pattern = pattern.replace("*", "%")
+        pattern = pattern.replace("?", "_")
+        return pattern
+
+    def search_files(
+        self,
+        user_id: Optional[str] = None,
+        filename: str = "*",
+        skip: int = 0,
+        limit: int = 100,
+        db: Optional[Session] = None,
+    ) -> list[FileModel]:
+        """
+        Search files with glob pattern matching, optional user filter, and pagination.
+
+        Args:
+            user_id: Filter by user ID. If None, returns files for all users.
+            filename: Glob pattern to match filenames (e.g., "*.txt"). Default "*" matches all.
+            skip: Number of results to skip for pagination.
+            limit: Maximum number of results to return.
+            db: Optional database session.
+
+        Returns:
+            List of matching FileModel objects, ordered by updated_at descending.
+        """
+        with get_db_context(db) as db:
+            query = db.query(File)
+
+            if user_id:
+                query = query.filter_by(user_id=user_id)
+
+            pattern = self._glob_to_like_pattern(filename)
+            if pattern != "%":
+                query = query.filter(File.filename.ilike(pattern, escape="\\"))
+
+            return [
+                FileModel.model_validate(file)
+                for file in query.order_by(File.updated_at.desc())
+                .offset(skip)
+                .limit(limit)
+                .all()
+            ]
+
+    def update_file_by_id(
+        self, id: str, form_data: FileUpdateForm, db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        with get_db_context(db) as db:
+            try:
+                file = db.query(File).filter_by(id=id).first()
+
+                if form_data.hash is not None:
+                    file.hash = form_data.hash
+
+                if form_data.data is not None:
+                    file.data = {**(file.data if file.data else {}), **form_data.data}
+
+                if form_data.meta is not None:
+                    file.meta = {**(file.meta if file.meta else {}), **form_data.meta}
+
+                file.updated_at = int(time.time())
+                db.commit()
+                return FileModel.model_validate(file)
+            except Exception as e:
+                log.exception(f"Error updating file completely by id: {e}")
+                return None
+
+    def update_file_hash_by_id(
+        self, id: str, hash: Optional[str], db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        with get_db_context(db) as db:
             try:
                 file = db.query(File).filter_by(id=id).first()
                 file.hash = hash
+                file.updated_at = int(time.time())
                 db.commit()
 
                 return FileModel.model_validate(file)
             except Exception:
                 return None
 
-    def update_file_data_by_id(self, id: str, data: dict) -> Optional[FileModel]:
-        with get_db() as db:
+    def update_file_data_by_id(
+        self, id: str, data: dict, db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        with get_db_context(db) as db:
             try:
                 file = db.query(File).filter_by(id=id).first()
                 file.data = {**(file.data if file.data else {}), **data}
+                file.updated_at = int(time.time())
                 db.commit()
                 return FileModel.model_validate(file)
             except Exception as e:
 
                 return None
 
-    def update_file_metadata_by_id(self, id: str, meta: dict) -> Optional[FileModel]:
-        with get_db() as db:
+    def update_file_metadata_by_id(
+        self, id: str, meta: dict, db: Optional[Session] = None
+    ) -> Optional[FileModel]:
+        with get_db_context(db) as db:
             try:
                 file = db.query(File).filter_by(id=id).first()
                 file.meta = {**(file.meta if file.meta else {}), **meta}
+                file.updated_at = int(time.time())
                 db.commit()
                 return FileModel.model_validate(file)
             except Exception:
                 return None
 
-    def delete_file_by_id(self, id: str) -> bool:
-        with get_db() as db:
+                return False
+
+    def delete_file_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             try:
                 db.query(File).filter_by(id=id).delete()
                 db.commit()
@@ -243,8 +385,8 @@ class FilesTable:
             except Exception:
                 return False
 
-    def delete_all_files(self) -> bool:
-        with get_db() as db:
+    def delete_all_files(self, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             try:
                 db.query(File).delete()
                 db.commit()

backend/open_webui/models/folders.py

@@ -7,13 +7,11 @@ import re
 
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Column, Text, JSON, Boolean, func
+from sqlalchemy.orm import Session
 
-from open_webui.internal.db import Base, get_db
-from open_webui.env import SRC_LOG_LEVELS
-
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 
 ####################
@@ -23,7 +21,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 class Folder(Base):
     __tablename__ = "folder"
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
     parent_id = Column(Text, nullable=True)
     user_id = Column(Text)
     name = Column(Text)
@@ -85,9 +83,13 @@ class FolderUpdateForm(BaseModel):
 
 class FolderTable:
     def insert_new_folder(
-        self, user_id: str, form_data: FolderForm, parent_id: Optional[str] = None
+        self,
+        user_id: str,
+        form_data: FolderForm,
+        parent_id: Optional[str] = None,
+        db: Optional[Session] = None,
     ) -> Optional[FolderModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             id = str(uuid.uuid4())
             folder = FolderModel(
                 **{
@@ -113,10 +115,10 @@ class FolderTable:
                 return None
 
     def get_folder_by_id_and_user_id(
-        self, id: str, user_id: str
+        self, id: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[FolderModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 folder = db.query(Folder).filter_by(id=id, user_id=user_id).first()
 
                 if not folder:
@@ -127,15 +129,15 @@ class FolderTable:
             return None
 
     def get_children_folders_by_id_and_user_id(
-        self, id: str, user_id: str
+        self, id: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[list[FolderModel]]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 folders = []
 
                 def get_children(folder):
                     children = self.get_folders_by_parent_id_and_user_id(
-                        folder.id, user_id
+                        folder.id, user_id, db=db
                     )
                     for child in children:
                         get_children(child)
@@ -150,18 +152,24 @@ class FolderTable:
         except Exception:
             return None
 
-    def get_folders_by_user_id(self, user_id: str) -> list[FolderModel]:
-        with get_db() as db:
+    def get_folders_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> list[FolderModel]:
+        with get_db_context(db) as db:
             return [
                 FolderModel.model_validate(folder)
                 for folder in db.query(Folder).filter_by(user_id=user_id).all()
             ]
 
     def get_folder_by_parent_id_and_user_id_and_name(
-        self, parent_id: Optional[str], user_id: str, name: str
+        self,
+        parent_id: Optional[str],
+        user_id: str,
+        name: str,
+        db: Optional[Session] = None,
     ) -> Optional[FolderModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 # Check if folder exists
                 folder = (
                     db.query(Folder)
@@ -179,9 +187,9 @@ class FolderTable:
             return None
 
     def get_folders_by_parent_id_and_user_id(
-        self, parent_id: Optional[str], user_id: str
+        self, parent_id: Optional[str], user_id: str, db: Optional[Session] = None
     ) -> list[FolderModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             return [
                 FolderModel.model_validate(folder)
                 for folder in db.query(Folder)
@@ -194,9 +202,10 @@ class FolderTable:
         id: str,
         user_id: str,
         parent_id: str,
+        db: Optional[Session] = None,
     ) -> Optional[FolderModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 folder = db.query(Folder).filter_by(id=id, user_id=user_id).first()
 
                 if not folder:
@@ -213,10 +222,14 @@ class FolderTable:
             return
 
     def update_folder_by_id_and_user_id(
-        self, id: str, user_id: str, form_data: FolderUpdateForm
+        self,
+        id: str,
+        user_id: str,
+        form_data: FolderUpdateForm,
+        db: Optional[Session] = None,
     ) -> Optional[FolderModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 folder = db.query(Folder).filter_by(id=id, user_id=user_id).first()
 
                 if not folder:
@@ -259,10 +272,10 @@ class FolderTable:
             return
 
     def update_folder_is_expanded_by_id_and_user_id(
-        self, id: str, user_id: str, is_expanded: bool
+        self, id: str, user_id: str, is_expanded: bool, db: Optional[Session] = None
     ) -> Optional[FolderModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 folder = db.query(Folder).filter_by(id=id, user_id=user_id).first()
 
                 if not folder:
@@ -278,10 +291,12 @@ class FolderTable:
             log.error(f"update_folder: {e}")
             return
 
-    def delete_folder_by_id_and_user_id(self, id: str, user_id: str) -> list[str]:
+    def delete_folder_by_id_and_user_id(
+        self, id: str, user_id: str, db: Optional[Session] = None
+    ) -> list[str]:
         try:
             folder_ids = []
-            with get_db() as db:
+            with get_db_context(db) as db:
                 folder = db.query(Folder).filter_by(id=id, user_id=user_id).first()
                 if not folder:
                     return folder_ids
@@ -291,7 +306,7 @@ class FolderTable:
                 # Delete all children folders
                 def delete_children(folder):
                     folder_children = self.get_folders_by_parent_id_and_user_id(
-                        folder.id, user_id
+                        folder.id, user_id, db=db
                     )
                     for folder_child in folder_children:
 
@@ -316,7 +331,7 @@ class FolderTable:
         return name.strip().lower()
 
     def search_folders_by_names(
-        self, user_id: str, queries: list[str]
+        self, user_id: str, queries: list[str], db: Optional[Session] = None
     ) -> list[FolderModel]:
         """
         Search for folders for a user where the name matches any of the queries, treating _ and space as equivalent, case-insensitive.
@@ -326,7 +341,7 @@ class FolderTable:
             return []
 
         results = {}
-        with get_db() as db:
+        with get_db_context(db) as db:
             folders = db.query(Folder).filter_by(user_id=user_id).all()
             for folder in folders:
                 if self.normalize_folder_name(folder.name) in normalized_queries:
@@ -334,7 +349,7 @@ class FolderTable:
 
                     # get children folders
                     children = self.get_children_folders_by_id_and_user_id(
-                        folder.id, user_id
+                        folder.id, user_id, db=db
                     )
                     for child in children:
                         results[child.id] = child
@@ -347,14 +362,14 @@ class FolderTable:
             return results
 
     def search_folders_by_name_contains(
-        self, user_id: str, query: str
+        self, user_id: str, query: str, db: Optional[Session] = None
     ) -> list[FolderModel]:
         """
         Partial match: normalized name contains (as substring) the normalized query.
         """
         normalized_query = self.normalize_folder_name(query)
         results = []
-        with get_db() as db:
+        with get_db_context(db) as db:
             folders = db.query(Folder).filter_by(user_id=user_id).all()
             for folder in folders:
                 norm_name = self.normalize_folder_name(folder.name)

backend/open_webui/models/functions.py

@@ -2,14 +2,13 @@ import logging
 import time
 from typing import Optional
 
-from open_webui.internal.db import Base, JSONField, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 from open_webui.models.users import Users, UserModel
-from open_webui.env import SRC_LOG_LEVELS
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Boolean, Column, String, Text, Index
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # Functions DB Schema
@@ -19,7 +18,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 class Function(Base):
     __tablename__ = "function"
 
-    id = Column(String, primary_key=True)
+    id = Column(String, primary_key=True, unique=True)
     user_id = Column(String)
     name = Column(Text)
     type = Column(Text)
@@ -105,7 +104,11 @@ class FunctionValves(BaseModel):
 
 class FunctionsTable:
     def insert_new_function(
-        self, user_id: str, type: str, form_data: FunctionForm
+        self,
+        user_id: str,
+        type: str,
+        form_data: FunctionForm,
+        db: Optional[Session] = None,
     ) -> Optional[FunctionModel]:
         function = FunctionModel(
             **{
@@ -118,7 +121,7 @@ class FunctionsTable:
         )
 
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 result = Function(**function.model_dump())
                 db.add(result)
                 db.commit()
@@ -132,11 +135,14 @@ class FunctionsTable:
             return None
 
     def sync_functions(
-        self, user_id: str, functions: list[FunctionWithValvesModel]
+        self,
+        user_id: str,
+        functions: list[FunctionWithValvesModel],
+        db: Optional[Session] = None,
     ) -> list[FunctionWithValvesModel]:
         # Synchronize functions for a user by updating existing ones, inserting new ones, and removing those that are no longer present.
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 # Get existing functions
                 existing_functions = db.query(Function).all()
                 existing_ids = {func.id for func in existing_functions}
@@ -179,18 +185,39 @@ class FunctionsTable:
             log.exception(f"Error syncing functions for user {user_id}: {e}")
             return []
 
-    def get_function_by_id(self, id: str) -> Optional[FunctionModel]:
+    def get_function_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[FunctionModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 function = db.get(Function, id)
                 return FunctionModel.model_validate(function)
         except Exception:
             return None
 
+    def get_functions_by_ids(
+        self, ids: list[str], db: Optional[Session] = None
+    ) -> list[FunctionModel]:
+        """
+        Batch fetch multiple functions by their IDs in a single query.
+        Returns functions in the same order as the input IDs (None entries filtered out).
+        """
+        if not ids:
+            return []
+        try:
+            with get_db_context(db) as db:
+                functions = db.query(Function).filter(Function.id.in_(ids)).all()
+                # Create a dict for O(1) lookup
+                func_dict = {f.id: FunctionModel.model_validate(f) for f in functions}
+                # Return in original order, filtering out any not found
+                return [func_dict[id] for id in ids if id in func_dict]
+        except Exception:
+            return []
+
     def get_functions(
-        self, active_only=False, include_valves=False
+        self, active_only=False, include_valves=False, db: Optional[Session] = None
     ) -> list[FunctionModel | FunctionWithValvesModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             if active_only:
                 functions = db.query(Function).filter_by(is_active=True).all()
 
@@ -207,12 +234,14 @@ class FunctionsTable:
                     FunctionModel.model_validate(function) for function in functions
                 ]
 
-    def get_function_list(self) -> list[FunctionUserResponse]:
-        with get_db() as db:
+    def get_function_list(
+        self, db: Optional[Session] = None
+    ) -> list[FunctionUserResponse]:
+        with get_db_context(db) as db:
             functions = db.query(Function).order_by(Function.updated_at.desc()).all()
             user_ids = list(set(func.user_id for func in functions))
 
-            users = Users.get_users_by_user_ids(user_ids) if user_ids else []
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
             users_dict = {user.id: user for user in users}
 
             return [
@@ -230,9 +259,9 @@ class FunctionsTable:
             ]
 
     def get_functions_by_type(
-        self, type: str, active_only=False
+        self, type: str, active_only=False, db: Optional[Session] = None
     ) -> list[FunctionModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             if active_only:
                 return [
                     FunctionModel.model_validate(function)
@@ -246,8 +275,10 @@ class FunctionsTable:
                     for function in db.query(Function).filter_by(type=type).all()
                 ]
 
-    def get_global_filter_functions(self) -> list[FunctionModel]:
-        with get_db() as db:
+    def get_global_filter_functions(
+        self, db: Optional[Session] = None
+    ) -> list[FunctionModel]:
+        with get_db_context(db) as db:
             return [
                 FunctionModel.model_validate(function)
                 for function in db.query(Function)
@@ -255,8 +286,10 @@ class FunctionsTable:
                 .all()
             ]
 
-    def get_global_action_functions(self) -> list[FunctionModel]:
-        with get_db() as db:
+    def get_global_action_functions(
+        self, db: Optional[Session] = None
+    ) -> list[FunctionModel]:
+        with get_db_context(db) as db:
             return [
                 FunctionModel.model_validate(function)
                 for function in db.query(Function)
@@ -264,8 +297,10 @@ class FunctionsTable:
                 .all()
             ]
 
-    def get_function_valves_by_id(self, id: str) -> Optional[dict]:
-        with get_db() as db:
+    def get_function_valves_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[dict]:
+        with get_db_context(db) as db:
             try:
                 function = db.get(Function, id)
                 return function.valves if function.valves else {}
@@ -274,23 +309,23 @@ class FunctionsTable:
                 return None
 
     def update_function_valves_by_id(
-        self, id: str, valves: dict
+        self, id: str, valves: dict, db: Optional[Session] = None
     ) -> Optional[FunctionValves]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             try:
                 function = db.get(Function, id)
                 function.valves = valves
                 function.updated_at = int(time.time())
                 db.commit()
                 db.refresh(function)
-                return self.get_function_by_id(id)
+                return FunctionModel.model_validate(function)
             except Exception:
                 return None
 
     def update_function_metadata_by_id(
-        self, id: str, metadata: dict
+        self, id: str, metadata: dict, db: Optional[Session] = None
     ) -> Optional[FunctionModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             try:
                 function = db.get(Function, id)
 
@@ -303,7 +338,7 @@ class FunctionsTable:
                     function.updated_at = int(time.time())
                     db.commit()
                     db.refresh(function)
-                    return self.get_function_by_id(id)
+                    return FunctionModel.model_validate(function)
                 else:
                     return None
             except Exception as e:
@@ -311,10 +346,10 @@ class FunctionsTable:
                 return None
 
     def get_user_valves_by_id_and_user_id(
-        self, id: str, user_id: str
+        self, id: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[dict]:
         try:
-            user = Users.get_user_by_id(user_id)
+            user = Users.get_user_by_id(user_id, db=db)
             user_settings = user.settings.model_dump() if user.settings else {}
 
             # Check if user has "functions" and "valves" settings
@@ -329,10 +364,10 @@ class FunctionsTable:
             return None
 
     def update_user_valves_by_id_and_user_id(
-        self, id: str, user_id: str, valves: dict
+        self, id: str, user_id: str, valves: dict, db: Optional[Session] = None
     ) -> Optional[dict]:
         try:
-            user = Users.get_user_by_id(user_id)
+            user = Users.get_user_by_id(user_id, db=db)
             user_settings = user.settings.model_dump() if user.settings else {}
 
             # Check if user has "functions" and "valves" settings
@@ -344,7 +379,7 @@ class FunctionsTable:
             user_settings["functions"]["valves"][id] = valves
 
             # Update the user settings in the database
-            Users.update_user_by_id(user_id, {"settings": user_settings})
+            Users.update_user_by_id(user_id, {"settings": user_settings}, db=db)
 
             return user_settings["functions"]["valves"][id]
         except Exception as e:
@@ -353,8 +388,10 @@ class FunctionsTable:
             )
             return None
 
-    def update_function_by_id(self, id: str, updated: dict) -> Optional[FunctionModel]:
-        with get_db() as db:
+    def update_function_by_id(
+        self, id: str, updated: dict, db: Optional[Session] = None
+    ) -> Optional[FunctionModel]:
+        with get_db_context(db) as db:
             try:
                 db.query(Function).filter_by(id=id).update(
                     {
@@ -363,12 +400,13 @@ class FunctionsTable:
                     }
                 )
                 db.commit()
-                return self.get_function_by_id(id)
+                function = db.get(Function, id)
+                return FunctionModel.model_validate(function) if function else None
             except Exception:
                 return None
 
-    def deactivate_all_functions(self) -> Optional[bool]:
-        with get_db() as db:
+    def deactivate_all_functions(self, db: Optional[Session] = None) -> Optional[bool]:
+        with get_db_context(db) as db:
             try:
                 db.query(Function).update(
                     {
@@ -381,8 +419,8 @@ class FunctionsTable:
             except Exception:
                 return None
 
-    def delete_function_by_id(self, id: str) -> bool:
-        with get_db() as db:
+    def delete_function_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             try:
                 db.query(Function).filter_by(id=id).delete()
                 db.commit()

backend/open_webui/models/groups.py

@@ -4,18 +4,28 @@ import time
 from typing import Optional
 import uuid
 
-from open_webui.internal.db import Base, get_db
-from open_webui.env import SRC_LOG_LEVELS
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 
 from open_webui.models.files import FileMetadataResponse
 
 
 from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Column, String, Text, JSON, func
-
+from sqlalchemy import (
+    BigInteger,
+    Column,
+    String,
+    Text,
+    JSON,
+    and_,
+    func,
+    ForeignKey,
+    cast,
+    or_,
+    select,
+)
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # UserGroup DB Schema
@@ -35,14 +45,12 @@ class Group(Base):
     meta = Column(JSON, nullable=True)
 
     permissions = Column(JSON, nullable=True)
-    user_ids = Column(JSON, nullable=True)
 
     created_at = Column(BigInteger)
     updated_at = Column(BigInteger)
 
 
 class GroupModel(BaseModel):
-    model_config = ConfigDict(from_attributes=True)
     id: str
     user_id: str
 
@@ -53,49 +61,79 @@ class GroupModel(BaseModel):
     meta: Optional[dict] = None
 
     permissions: Optional[dict] = None
-    user_ids: list[str] = []
 
     created_at: int  # timestamp in epoch
     updated_at: int  # timestamp in epoch
 
+    model_config = ConfigDict(from_attributes=True)
+
+
+class GroupMember(Base):
+    __tablename__ = "group_member"
+
+    id = Column(Text, unique=True, primary_key=True)
+    group_id = Column(
+        Text,
+        ForeignKey("group.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    user_id = Column(Text, nullable=False)
+    created_at = Column(BigInteger, nullable=True)
+    updated_at = Column(BigInteger, nullable=True)
+
+
+class GroupMemberModel(BaseModel):
+    id: str
+    group_id: str
+    user_id: str
+    created_at: Optional[int] = None  # timestamp in epoch
+    updated_at: Optional[int] = None  # timestamp in epoch
+
 
 ####################
 # Forms
 ####################
 
 
-class GroupResponse(BaseModel):
+class GroupResponse(GroupModel):
+    member_count: Optional[int] = None
+
+
+class GroupInfoResponse(BaseModel):
     id: str
     user_id: str
     name: str
     description: str
-    permissions: Optional[dict] = None
-    data: Optional[dict] = None
-    meta: Optional[dict] = None
-    user_ids: list[str] = []
-    created_at: int  # timestamp in epoch
-    updated_at: int  # timestamp in epoch
+    member_count: Optional[int] = None
+    created_at: int
+    updated_at: int
 
 
 class GroupForm(BaseModel):
     name: str
     description: str
     permissions: Optional[dict] = None
+    data: Optional[dict] = None
 
 
 class UserIdsForm(BaseModel):
     user_ids: Optional[list[str]] = None
 
 
-class GroupUpdateForm(GroupForm, UserIdsForm):
+class GroupUpdateForm(GroupForm):
     pass
 
 
+class GroupListResponse(BaseModel):
+    items: list[GroupResponse] = []
+    total: int = 0
+
+
 class GroupTable:
     def insert_new_group(
-        self, user_id: str, form_data: GroupForm
+        self, user_id: str, form_data: GroupForm, db: Optional[Session] = None
     ) -> Optional[GroupModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             group = GroupModel(
                 **{
                     **form_data.model_dump(exclude_none=True),
@@ -119,48 +157,266 @@ class GroupTable:
             except Exception:
                 return None
 
-    def get_groups(self) -> list[GroupModel]:
-        with get_db() as db:
+    def get_all_groups(self, db: Optional[Session] = None) -> list[GroupModel]:
+        with get_db_context(db) as db:
+            groups = db.query(Group).order_by(Group.updated_at.desc()).all()
+            return [GroupModel.model_validate(group) for group in groups]
+
+    def get_groups(self, filter, db: Optional[Session] = None) -> list[GroupResponse]:
+        with get_db_context(db) as db:
+            member_count = func.count(GroupMember.user_id).label("member_count")
+            query = db.query(Group, member_count).outerjoin(
+                GroupMember, GroupMember.group_id == Group.id
+            )
+
+            if filter:
+                if "query" in filter:
+                    query = query.filter(Group.name.ilike(f"%{filter['query']}%"))
+
+                # When share filter is present, member check is handled in the share logic
+                if "share" in filter:
+                    share_value = filter["share"]
+                    member_id = filter.get("member_id")
+                    json_share = Group.data["config"]["share"]
+                    json_share_str = json_share.as_string()
+                    json_share_lower = func.lower(json_share_str)
+
+                    if share_value:
+                        anyone_can_share = or_(
+                            Group.data.is_(None),
+                            json_share_str.is_(None),
+                            json_share_lower == "true",
+                            json_share_lower == "1",  # Handle SQLite boolean true
+                        )
+
+                        if member_id:
+                            member_groups_select = select(GroupMember.group_id).where(
+                                GroupMember.user_id == member_id
+                            )
+                            members_only_and_is_member = and_(
+                                json_share_lower == "members",
+                                Group.id.in_(member_groups_select),
+                            )
+                            query = query.filter(
+                                or_(anyone_can_share, members_only_and_is_member)
+                            )
+                        else:
+                            query = query.filter(anyone_can_share)
+                    else:
+                        query = query.filter(
+                            and_(Group.data.isnot(None), json_share_lower == "false")
+                        )
+
+                else:
+                    # Only apply member_id filter when share filter is NOT present
+                    if "member_id" in filter:
+                        query = query.filter(
+                            Group.id.in_(
+                                select(GroupMember.group_id).where(
+                                    GroupMember.user_id == filter["member_id"]
+                                )
+                            )
+                        )
+
+            results = query.group_by(Group.id).order_by(Group.updated_at.desc()).all()
+
             return [
-                GroupModel.model_validate(group)
-                for group in db.query(Group).order_by(Group.updated_at.desc()).all()
+                GroupResponse.model_validate(
+                    {
+                        **GroupModel.model_validate(group).model_dump(),
+                        "member_count": count or 0,
+                    }
+                )
+                for group, count in results
             ]
 
-    def get_groups_by_member_id(self, user_id: str) -> list[GroupModel]:
-        with get_db() as db:
+    def search_groups(
+        self,
+        filter: Optional[dict] = None,
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> GroupListResponse:
+        with get_db_context(db) as db:
+            query = db.query(Group)
+
+            if filter:
+                if "query" in filter:
+                    query = query.filter(Group.name.ilike(f"%{filter['query']}%"))
+                if "member_id" in filter:
+                    query = query.filter(
+                        Group.id.in_(
+                            select(GroupMember.group_id).where(
+                                GroupMember.user_id == filter["member_id"]
+                            )
+                        )
+                    )
+
+                if "share" in filter:
+                    share_value = filter["share"]
+                    query = query.filter(
+                        Group.data.op("->>")("share") == str(share_value)
+                    )
+
+            total = query.count()
+
+            member_count = func.count(GroupMember.user_id).label("member_count")
+            results = (
+                query.add_columns(member_count)
+                .outerjoin(GroupMember, GroupMember.group_id == Group.id)
+                .group_by(Group.id)
+                .order_by(Group.updated_at.desc())
+                .offset(skip)
+                .limit(limit)
+                .all()
+            )
+
+            return {
+                "items": [
+                    GroupResponse.model_validate(
+                        {
+                            **GroupModel.model_validate(group).model_dump(),
+                            "member_count": count or 0,
+                        }
+                    )
+                    for group, count in results
+                ],
+                "total": total,
+            }
+
+    def get_groups_by_member_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> list[GroupModel]:
+        with get_db_context(db) as db:
             return [
                 GroupModel.model_validate(group)
                 for group in db.query(Group)
-                .filter(
-                    func.json_array_length(Group.user_ids) > 0
-                )  # Ensure array exists
-                .filter(
-                    Group.user_ids.cast(String).like(f'%"{user_id}"%')
-                )  # String-based check
+                .join(GroupMember, GroupMember.group_id == Group.id)
+                .filter(GroupMember.user_id == user_id)
                 .order_by(Group.updated_at.desc())
                 .all()
             ]
 
-    def get_group_by_id(self, id: str) -> Optional[GroupModel]:
+    def get_groups_by_member_ids(
+        self, user_ids: list[str], db: Optional[Session] = None
+    ) -> dict[str, list[GroupModel]]:
+        """Fetch groups for multiple users in a single query to avoid N+1."""
+        with get_db_context(db) as db:
+            # Query GroupMember joined with Group, filtering by user_ids
+            results = (
+                db.query(GroupMember.user_id, Group)
+                .join(Group, Group.id == GroupMember.group_id)
+                .filter(GroupMember.user_id.in_(user_ids))
+                .order_by(Group.updated_at.desc())
+                .all()
+            )
+
+            # Group groups by user_id
+            user_groups: dict[str, list[GroupModel]] = {uid: [] for uid in user_ids}
+            for user_id, group in results:
+                user_groups[user_id].append(GroupModel.model_validate(group))
+
+            return user_groups
+
+    def get_group_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[GroupModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 group = db.query(Group).filter_by(id=id).first()
                 return GroupModel.model_validate(group) if group else None
         except Exception:
             return None
 
-    def get_group_user_ids_by_id(self, id: str) -> Optional[str]:
-        group = self.get_group_by_id(id)
-        if group:
-            return group.user_ids
-        else:
-            return None
+    def get_group_user_ids_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> list[str]:
+        with get_db_context(db) as db:
+            members = (
+                db.query(GroupMember.user_id).filter(GroupMember.group_id == id).all()
+            )
+
+            if not members:
+                return []
+
+            return [m[0] for m in members]
+
+    def get_group_user_ids_by_ids(
+        self, group_ids: list[str], db: Optional[Session] = None
+    ) -> dict[str, list[str]]:
+        with get_db_context(db) as db:
+            members = (
+                db.query(GroupMember.group_id, GroupMember.user_id)
+                .filter(GroupMember.group_id.in_(group_ids))
+                .all()
+            )
+
+            group_user_ids: dict[str, list[str]] = {
+                group_id: [] for group_id in group_ids
+            }
+
+            for group_id, user_id in members:
+                group_user_ids[group_id].append(user_id)
+
+            return group_user_ids
+
+    def set_group_user_ids_by_id(
+        self, group_id: str, user_ids: list[str], db: Optional[Session] = None
+    ) -> None:
+        with get_db_context(db) as db:
+            # Delete existing members
+            db.query(GroupMember).filter(GroupMember.group_id == group_id).delete()
+
+            # Insert new members
+            now = int(time.time())
+            new_members = [
+                GroupMember(
+                    id=str(uuid.uuid4()),
+                    group_id=group_id,
+                    user_id=user_id,
+                    created_at=now,
+                    updated_at=now,
+                )
+                for user_id in user_ids
+            ]
+
+            db.add_all(new_members)
+            db.commit()
+
+    def get_group_member_count_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> int:
+        with get_db_context(db) as db:
+            count = (
+                db.query(func.count(GroupMember.user_id))
+                .filter(GroupMember.group_id == id)
+                .scalar()
+            )
+            return count if count else 0
+
+    def get_group_member_counts_by_ids(
+        self, ids: list[str], db: Optional[Session] = None
+    ) -> dict[str, int]:
+        if not ids:
+            return {}
+        with get_db_context(db) as db:
+            rows = (
+                db.query(GroupMember.group_id, func.count(GroupMember.user_id))
+                .filter(GroupMember.group_id.in_(ids))
+                .group_by(GroupMember.group_id)
+                .all()
+            )
+            return {group_id: count for group_id, count in rows}
 
     def update_group_by_id(
-        self, id: str, form_data: GroupUpdateForm, overwrite: bool = False
+        self,
+        id: str,
+        form_data: GroupUpdateForm,
+        overwrite: bool = False,
+        db: Optional[Session] = None,
     ) -> Optional[GroupModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 db.query(Group).filter_by(id=id).update(
                     {
                         **form_data.model_dump(exclude_none=True),
@@ -168,22 +424,22 @@ class GroupTable:
                     }
                 )
                 db.commit()
-                return self.get_group_by_id(id=id)
+                return self.get_group_by_id(id=id, db=db)
         except Exception as e:
             log.exception(e)
             return None
 
-    def delete_group_by_id(self, id: str) -> bool:
+    def delete_group_by_id(self, id: str, db: Optional[Session] = None) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 db.query(Group).filter_by(id=id).delete()
                 db.commit()
                 return True
         except Exception:
             return False
 
-    def delete_all_groups(self) -> bool:
-        with get_db() as db:
+    def delete_all_groups(self, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             try:
                 db.query(Group).delete()
                 db.commit()
@@ -192,36 +448,47 @@ class GroupTable:
             except Exception:
                 return False
 
-    def remove_user_from_all_groups(self, user_id: str) -> bool:
-        with get_db() as db:
+    def remove_user_from_all_groups(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
             try:
-                groups = self.get_groups_by_member_id(user_id)
+                # Find all groups the user belongs to
+                groups = (
+                    db.query(Group)
+                    .join(GroupMember, GroupMember.group_id == Group.id)
+                    .filter(GroupMember.user_id == user_id)
+                    .all()
+                )
 
+                # Remove the user from each group
                 for group in groups:
-                    group.user_ids.remove(user_id)
-                    db.query(Group).filter_by(id=group.id).update(
-                        {
-                            "user_ids": group.user_ids,
-                            "updated_at": int(time.time()),
-                        }
-                    )
-                    db.commit()
+                    db.query(GroupMember).filter(
+                        GroupMember.group_id == group.id, GroupMember.user_id == user_id
+                    ).delete()
 
+                    db.query(Group).filter_by(id=group.id).update(
+                        {"updated_at": int(time.time())}
+                    )
+
+                db.commit()
                 return True
+
             except Exception:
+                db.rollback()
                 return False
 
     def create_groups_by_group_names(
-        self, user_id: str, group_names: list[str]
+        self, user_id: str, group_names: list[str], db: Optional[Session] = None
     ) -> list[GroupModel]:
 
         # check for existing groups
-        existing_groups = self.get_groups()
+        existing_groups = self.get_all_groups(db=db)
         existing_group_names = {group.name for group in existing_groups}
 
         new_groups = []
 
-        with get_db() as db:
+        with get_db_context(db) as db:
             for group_name in group_names:
                 if group_name not in existing_group_names:
                     new_group = GroupModel(
@@ -243,96 +510,136 @@ class GroupTable:
                         continue
             return new_groups
 
-    def sync_groups_by_group_names(self, user_id: str, group_names: list[str]) -> bool:
-        with get_db() as db:
+    def sync_groups_by_group_names(
+        self, user_id: str, group_names: list[str], db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
             try:
-                groups = db.query(Group).filter(Group.name.in_(group_names)).all()
-                group_ids = [group.id for group in groups]
+                now = int(time.time())
 
-                # Remove user from groups not in the new list
-                existing_groups = self.get_groups_by_member_id(user_id)
+                # 1. Groups that SHOULD contain the user
+                target_groups = (
+                    db.query(Group).filter(Group.name.in_(group_names)).all()
+                )
+                target_group_ids = {g.id for g in target_groups}
 
-                for group in existing_groups:
-                    if group.id not in group_ids:
-                        group.user_ids.remove(user_id)
-                        db.query(Group).filter_by(id=group.id).update(
-                            {
-                                "user_ids": group.user_ids,
-                                "updated_at": int(time.time()),
-                            }
+                # 2. Groups the user is CURRENTLY in
+                existing_group_ids = {
+                    g.id
+                    for g in db.query(Group)
+                    .join(GroupMember, GroupMember.group_id == Group.id)
+                    .filter(GroupMember.user_id == user_id)
+                    .all()
+                }
+
+                # 3. Determine adds + removals
+                groups_to_add = target_group_ids - existing_group_ids
+                groups_to_remove = existing_group_ids - target_group_ids
+
+                # 4. Remove in one bulk delete
+                if groups_to_remove:
+                    db.query(GroupMember).filter(
+                        GroupMember.user_id == user_id,
+                        GroupMember.group_id.in_(groups_to_remove),
+                    ).delete(synchronize_session=False)
+
+                    db.query(Group).filter(Group.id.in_(groups_to_remove)).update(
+                        {"updated_at": now}, synchronize_session=False
+                    )
+
+                # 5. Bulk insert missing memberships
+                for group_id in groups_to_add:
+                    db.add(
+                        GroupMember(
+                            id=str(uuid.uuid4()),
+                            group_id=group_id,
+                            user_id=user_id,
+                            created_at=now,
+                            updated_at=now,
                         )
+                    )
 
-                # Add user to new groups
-                for group in groups:
-                    if user_id not in group.user_ids:
-                        group.user_ids.append(user_id)
-                        db.query(Group).filter_by(id=group.id).update(
-                            {
-                                "user_ids": group.user_ids,
-                                "updated_at": int(time.time()),
-                            }
-                        )
+                if groups_to_add:
+                    db.query(Group).filter(Group.id.in_(groups_to_add)).update(
+                        {"updated_at": now}, synchronize_session=False
+                    )
 
                 db.commit()
                 return True
+
             except Exception as e:
                 log.exception(e)
+                db.rollback()
                 return False
 
     def add_users_to_group(
-        self, id: str, user_ids: Optional[list[str]] = None
+        self,
+        id: str,
+        user_ids: Optional[list[str]] = None,
+        db: Optional[Session] = None,
     ) -> Optional[GroupModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 group = db.query(Group).filter_by(id=id).first()
                 if not group:
                     return None
 
-                group_user_ids = group.user_ids
-                if not group_user_ids or not isinstance(group_user_ids, list):
-                    group_user_ids = []
+                now = int(time.time())
 
-                group_user_ids = list(set(group_user_ids))  # Deduplicate
+                for user_id in user_ids or []:
+                    try:
+                        db.add(
+                            GroupMember(
+                                id=str(uuid.uuid4()),
+                                group_id=id,
+                                user_id=user_id,
+                                created_at=now,
+                                updated_at=now,
+                            )
+                        )
+                        db.flush()  # Detect unique constraint violation early
+                    except Exception:
+                        db.rollback()  # Clear failed INSERT
+                        db.begin()  # Start a new transaction
+                        continue  # Duplicate → ignore
 
-                for user_id in user_ids:
-                    if user_id not in group_user_ids:
-                        group_user_ids.append(user_id)
-
-                group.user_ids = group_user_ids
-                group.updated_at = int(time.time())
+                group.updated_at = now
                 db.commit()
                 db.refresh(group)
+
                 return GroupModel.model_validate(group)
+
         except Exception as e:
             log.exception(e)
             return None
 
     def remove_users_from_group(
-        self, id: str, user_ids: Optional[list[str]] = None
+        self,
+        id: str,
+        user_ids: Optional[list[str]] = None,
+        db: Optional[Session] = None,
     ) -> Optional[GroupModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 group = db.query(Group).filter_by(id=id).first()
                 if not group:
                     return None
 
-                group_user_ids = group.user_ids
-
-                if not group_user_ids or not isinstance(group_user_ids, list):
+                if not user_ids:
                     return GroupModel.model_validate(group)
 
-                group_user_ids = list(set(group_user_ids))  # Deduplicate
+                # Remove users from group_member in batch
+                db.query(GroupMember).filter(
+                    GroupMember.group_id == id, GroupMember.user_id.in_(user_ids)
+                ).delete(synchronize_session=False)
 
-                for user_id in user_ids:
-                    if user_id in group_user_ids:
-                        group_user_ids.remove(user_id)
-
-                group.user_ids = group_user_ids
+                # Update group timestamp
                 group.updated_at = int(time.time())
 
                 db.commit()
                 db.refresh(group)
                 return GroupModel.model_validate(group)
+
         except Exception as e:
             log.exception(e)
             return None

backend/open_webui/models/knowledge.py

@@ -4,21 +4,33 @@ import time
 from typing import Optional
 import uuid
 
-from open_webui.internal.db import Base, get_db
-from open_webui.env import SRC_LOG_LEVELS
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 
-from open_webui.models.files import FileMetadataResponse
+from open_webui.models.files import (
+    File,
+    FileModel,
+    FileMetadataResponse,
+    FileModelResponse,
+)
 from open_webui.models.groups import Groups
-from open_webui.models.users import Users, UserResponse
+from open_webui.models.users import User, UserModel, Users, UserResponse
+from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 
 
-from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Column, String, Text, JSON
-
-from open_webui.utils.access_control import has_access
+from pydantic import BaseModel, ConfigDict, Field
+from sqlalchemy import (
+    BigInteger,
+    Column,
+    ForeignKey,
+    String,
+    Text,
+    JSON,
+    UniqueConstraint,
+    or_,
+)
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # Knowledge DB Schema
@@ -34,26 +46,8 @@ class Knowledge(Base):
     name = Column(Text)
     description = Column(Text)
 
-    data = Column(JSON, nullable=True)
     meta = Column(JSON, nullable=True)
 
-    access_control = Column(JSON, nullable=True)  # Controls data access levels.
-    # Defines access control rules for this entry.
-    # - `None`: Public access, available to all users with the "user" role.
-    # - `{}`: Private access, restricted exclusively to the owner.
-    # - Custom permissions: Specific access control for reading and writing;
-    #   Can specify group or user-level restrictions:
-    #   {
-    #      "read": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      },
-    #      "write": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      }
-    #   }
-
     created_at = Column(BigInteger)
     updated_at = Column(BigInteger)
 
@@ -67,20 +61,50 @@ class KnowledgeModel(BaseModel):
     name: str
     description: str
 
-    data: Optional[dict] = None
     meta: Optional[dict] = None
 
-    access_control: Optional[dict] = None
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
 
     created_at: int  # timestamp in epoch
     updated_at: int  # timestamp in epoch
 
 
+class KnowledgeFile(Base):
+    __tablename__ = "knowledge_file"
+
+    id = Column(Text, unique=True, primary_key=True)
+
+    knowledge_id = Column(
+        Text, ForeignKey("knowledge.id", ondelete="CASCADE"), nullable=False
+    )
+    file_id = Column(Text, ForeignKey("file.id", ondelete="CASCADE"), nullable=False)
+    user_id = Column(Text, nullable=False)
+
+    created_at = Column(BigInteger, nullable=False)
+    updated_at = Column(BigInteger, nullable=False)
+
+    __table_args__ = (
+        UniqueConstraint(
+            "knowledge_id", "file_id", name="uq_knowledge_file_knowledge_file"
+        ),
+    )
+
+
+class KnowledgeFileModel(BaseModel):
+    id: str
+    knowledge_id: str
+    file_id: str
+    user_id: str
+
+    created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
+
+    model_config = ConfigDict(from_attributes=True)
+
+
 ####################
 # Forms
 ####################
-
-
 class KnowledgeUserModel(KnowledgeModel):
     user: Optional[UserResponse] = None
 
@@ -90,52 +114,86 @@ class KnowledgeResponse(KnowledgeModel):
 
 
 class KnowledgeUserResponse(KnowledgeUserModel):
-    files: Optional[list[FileMetadataResponse | dict]] = None
+    pass
 
 
 class KnowledgeForm(BaseModel):
     name: str
     description: str
-    data: Optional[dict] = None
-    access_control: Optional[dict] = None
+    access_grants: Optional[list[dict]] = None
+
+
+class FileUserResponse(FileModelResponse):
+    user: Optional[UserResponse] = None
+
+
+class KnowledgeListResponse(BaseModel):
+    items: list[KnowledgeUserModel]
+    total: int
+
+
+class KnowledgeFileListResponse(BaseModel):
+    items: list[FileUserResponse]
+    total: int
 
 
 class KnowledgeTable:
+    def _get_access_grants(
+        self, knowledge_id: str, db: Optional[Session] = None
+    ) -> list[AccessGrantModel]:
+        return AccessGrants.get_grants_by_resource("knowledge", knowledge_id, db=db)
+
+    def _to_knowledge_model(
+        self, knowledge: Knowledge, db: Optional[Session] = None
+    ) -> KnowledgeModel:
+        knowledge_data = KnowledgeModel.model_validate(knowledge).model_dump(
+            exclude={"access_grants"}
+        )
+        knowledge_data["access_grants"] = self._get_access_grants(
+            knowledge_data["id"], db=db
+        )
+        return KnowledgeModel.model_validate(knowledge_data)
+
     def insert_new_knowledge(
-        self, user_id: str, form_data: KnowledgeForm
+        self, user_id: str, form_data: KnowledgeForm, db: Optional[Session] = None
     ) -> Optional[KnowledgeModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             knowledge = KnowledgeModel(
                 **{
-                    **form_data.model_dump(),
+                    **form_data.model_dump(exclude={"access_grants"}),
                     "id": str(uuid.uuid4()),
                     "user_id": user_id,
                     "created_at": int(time.time()),
                     "updated_at": int(time.time()),
+                    "access_grants": [],
                 }
             )
 
             try:
-                result = Knowledge(**knowledge.model_dump())
+                result = Knowledge(**knowledge.model_dump(exclude={"access_grants"}))
                 db.add(result)
                 db.commit()
                 db.refresh(result)
+                AccessGrants.set_access_grants(
+                    "knowledge", result.id, form_data.access_grants, db=db
+                )
                 if result:
-                    return KnowledgeModel.model_validate(result)
+                    return self._to_knowledge_model(result, db=db)
                 else:
                     return None
             except Exception:
                 return None
 
-    def get_knowledge_bases(self) -> list[KnowledgeUserModel]:
-        with get_db() as db:
+    def get_knowledge_bases(
+        self, skip: int = 0, limit: int = 30, db: Optional[Session] = None
+    ) -> list[KnowledgeUserModel]:
+        with get_db_context(db) as db:
             all_knowledge = (
                 db.query(Knowledge).order_by(Knowledge.updated_at.desc()).all()
             )
-
             user_ids = list(set(knowledge.user_id for knowledge in all_knowledge))
 
-            users = Users.get_users_by_user_ids(user_ids) if user_ids else []
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
             users_dict = {user.id: user for user in users}
 
             knowledge_bases = []
@@ -144,68 +202,456 @@ class KnowledgeTable:
                 knowledge_bases.append(
                     KnowledgeUserModel.model_validate(
                         {
-                            **KnowledgeModel.model_validate(knowledge).model_dump(),
+                            **self._to_knowledge_model(knowledge, db=db).model_dump(),
                             "user": user.model_dump() if user else None,
                         }
                     )
                 )
             return knowledge_bases
 
-    def check_access_by_user_id(self, id, user_id, permission="write") -> bool:
-        knowledge = self.get_knowledge_by_id(id)
+    def search_knowledge_bases(
+        self,
+        user_id: str,
+        filter: dict,
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> KnowledgeListResponse:
+        try:
+            with get_db_context(db) as db:
+                query = db.query(Knowledge, User).outerjoin(
+                    User, User.id == Knowledge.user_id
+                )
+
+                if filter:
+                    query_key = filter.get("query")
+                    if query_key:
+                        query = query.filter(
+                            or_(
+                                Knowledge.name.ilike(f"%{query_key}%"),
+                                Knowledge.description.ilike(f"%{query_key}%"),
+                                User.name.ilike(f"%{query_key}%"),
+                                User.email.ilike(f"%{query_key}%"),
+                                User.username.ilike(f"%{query_key}%"),
+                            )
+                        )
+
+                    view_option = filter.get("view_option")
+                    if view_option == "created":
+                        query = query.filter(Knowledge.user_id == user_id)
+                    elif view_option == "shared":
+                        query = query.filter(Knowledge.user_id != user_id)
+
+                    query = AccessGrants.has_permission_filter(
+                        db=db,
+                        query=query,
+                        DocumentModel=Knowledge,
+                        filter=filter,
+                        resource_type="knowledge",
+                        permission="read",
+                    )
+
+                query = query.order_by(Knowledge.updated_at.desc(), Knowledge.id.asc())
+
+                total = query.count()
+                if skip:
+                    query = query.offset(skip)
+                if limit:
+                    query = query.limit(limit)
+
+                items = query.all()
+
+                knowledge_bases = []
+                for knowledge_base, user in items:
+                    knowledge_bases.append(
+                        KnowledgeUserModel.model_validate(
+                            {
+                                **self._to_knowledge_model(
+                                    knowledge_base, db=db
+                                ).model_dump(),
+                                "user": (
+                                    UserModel.model_validate(user).model_dump()
+                                    if user
+                                    else None
+                                ),
+                            }
+                        )
+                    )
+
+                return KnowledgeListResponse(items=knowledge_bases, total=total)
+        except Exception as e:
+            print(e)
+            return KnowledgeListResponse(items=[], total=0)
+
+    def search_knowledge_files(
+        self, filter: dict, skip: int = 0, limit: int = 30, db: Optional[Session] = None
+    ) -> KnowledgeFileListResponse:
+        """
+        Scalable version: search files across all knowledge bases the user has
+        READ access to, without loading all KBs or using large IN() lists.
+        """
+        try:
+            with get_db_context(db) as db:
+                # Base query: join Knowledge → KnowledgeFile → File
+                query = (
+                    db.query(File, User, Knowledge)
+                    .join(KnowledgeFile, File.id == KnowledgeFile.file_id)
+                    .join(Knowledge, KnowledgeFile.knowledge_id == Knowledge.id)
+                    .outerjoin(User, User.id == KnowledgeFile.user_id)
+                )
+
+                # Apply access-control directly to the joined query
+                # This makes the database handle filtering, even with 10k+ KBs
+                query = AccessGrants.has_permission_filter(
+                    db=db,
+                    query=query,
+                    DocumentModel=Knowledge,
+                    filter=filter,
+                    resource_type="knowledge",
+                    permission="read",
+                )
+
+                # Apply filename search
+                if filter:
+                    q = filter.get("query")
+                    if q:
+                        query = query.filter(File.filename.ilike(f"%{q}%"))
+
+                # Order by file changes
+                query = query.order_by(File.updated_at.desc(), File.id.asc())
+
+                # Count before pagination
+                total = query.count()
+
+                if skip:
+                    query = query.offset(skip)
+                if limit:
+                    query = query.limit(limit)
+
+                rows = query.all()
+
+                items = []
+                for file, user, knowledge in rows:
+                    items.append(
+                        FileUserResponse(
+                            **FileModel.model_validate(file).model_dump(),
+                            user=(
+                                UserResponse(
+                                    **UserModel.model_validate(user).model_dump()
+                                )
+                                if user
+                                else None
+                            ),
+                            collection=self._to_knowledge_model(
+                                knowledge, db=db
+                            ).model_dump(),
+                        )
+                    )
+
+                return KnowledgeFileListResponse(items=items, total=total)
+
+        except Exception as e:
+            print("search_knowledge_files error:", e)
+            return KnowledgeFileListResponse(items=[], total=0)
+
+    def check_access_by_user_id(
+        self, id, user_id, permission="write", db: Optional[Session] = None
+    ) -> bool:
+        knowledge = self.get_knowledge_by_id(id, db=db)
         if not knowledge:
             return False
         if knowledge.user_id == user_id:
             return True
-        user_group_ids = {group.id for group in Groups.get_groups_by_member_id(user_id)}
-        return has_access(user_id, permission, knowledge.access_control, user_group_ids)
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
+        return AccessGrants.has_access(
+            user_id=user_id,
+            resource_type="knowledge",
+            resource_id=knowledge.id,
+            permission=permission,
+            user_group_ids=user_group_ids,
+            db=db,
+        )
 
     def get_knowledge_bases_by_user_id(
-        self, user_id: str, permission: str = "write"
+        self, user_id: str, permission: str = "write", db: Optional[Session] = None
     ) -> list[KnowledgeUserModel]:
-        knowledge_bases = self.get_knowledge_bases()
-        user_group_ids = {group.id for group in Groups.get_groups_by_member_id(user_id)}
+        knowledge_bases = self.get_knowledge_bases(db=db)
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
         return [
             knowledge_base
             for knowledge_base in knowledge_bases
             if knowledge_base.user_id == user_id
-            or has_access(
-                user_id, permission, knowledge_base.access_control, user_group_ids
+            or 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,
             )
         ]
 
-    def get_knowledge_by_id(self, id: str) -> Optional[KnowledgeModel]:
+    def get_knowledge_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[KnowledgeModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 knowledge = db.query(Knowledge).filter_by(id=id).first()
-                return KnowledgeModel.model_validate(knowledge) if knowledge else None
+                return self._to_knowledge_model(knowledge, db=db) if knowledge else None
         except Exception:
             return None
 
-    def update_knowledge_by_id(
-        self, id: str, form_data: KnowledgeForm, overwrite: bool = False
+    def get_knowledge_by_id_and_user_id(
+        self, id: str, user_id: str, db: Optional[Session] = None
+    ) -> Optional[KnowledgeModel]:
+        knowledge = self.get_knowledge_by_id(id, db=db)
+        if not knowledge:
+            return None
+
+        if knowledge.user_id == user_id:
+            return knowledge
+
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
+        if AccessGrants.has_access(
+            user_id=user_id,
+            resource_type="knowledge",
+            resource_id=knowledge.id,
+            permission="write",
+            user_group_ids=user_group_ids,
+            db=db,
+        ):
+            return knowledge
+        return None
+
+    def get_knowledges_by_file_id(
+        self, file_id: str, db: Optional[Session] = None
+    ) -> list[KnowledgeModel]:
+        try:
+            with get_db_context(db) as db:
+                knowledges = (
+                    db.query(Knowledge)
+                    .join(KnowledgeFile, Knowledge.id == KnowledgeFile.knowledge_id)
+                    .filter(KnowledgeFile.file_id == file_id)
+                    .all()
+                )
+                return [
+                    self._to_knowledge_model(knowledge, db=db)
+                    for knowledge in knowledges
+                ]
+        except Exception:
+            return []
+
+    def search_files_by_id(
+        self,
+        knowledge_id: str,
+        user_id: str,
+        filter: dict,
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> KnowledgeFileListResponse:
+        try:
+            with get_db_context(db) as db:
+                query = (
+                    db.query(File, User)
+                    .join(KnowledgeFile, File.id == KnowledgeFile.file_id)
+                    .outerjoin(User, User.id == KnowledgeFile.user_id)
+                    .filter(KnowledgeFile.knowledge_id == knowledge_id)
+                )
+
+                # Default sort: updated_at descending
+                primary_sort = File.updated_at.desc()
+
+                if filter:
+                    query_key = filter.get("query")
+                    if query_key:
+                        query = query.filter(or_(File.filename.ilike(f"%{query_key}%")))
+
+                    view_option = filter.get("view_option")
+                    if view_option == "created":
+                        query = query.filter(KnowledgeFile.user_id == user_id)
+                    elif view_option == "shared":
+                        query = query.filter(KnowledgeFile.user_id != user_id)
+
+                    order_by = filter.get("order_by")
+                    direction = filter.get("direction")
+                    is_asc = direction == "asc"
+
+                    if order_by == "name":
+                        primary_sort = (
+                            File.filename.asc() if is_asc else File.filename.desc()
+                        )
+                    elif order_by == "created_at":
+                        primary_sort = (
+                            File.created_at.asc() if is_asc else File.created_at.desc()
+                        )
+                    elif order_by == "updated_at":
+                        primary_sort = (
+                            File.updated_at.asc() if is_asc else File.updated_at.desc()
+                        )
+
+                # Apply sort with secondary key for deterministic pagination
+                query = query.order_by(primary_sort, File.id.asc())
+
+                # Count BEFORE pagination
+                total = query.count()
+
+                if skip:
+                    query = query.offset(skip)
+                if limit:
+                    query = query.limit(limit)
+
+                items = query.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
+                            ),
+                        )
+                    )
+
+                return KnowledgeFileListResponse(items=files, total=total)
+        except Exception as e:
+            print(e)
+            return KnowledgeFileListResponse(items=[], total=0)
+
+    def get_files_by_id(
+        self, knowledge_id: str, db: Optional[Session] = None
+    ) -> list[FileModel]:
+        try:
+            with get_db_context(db) as db:
+                files = (
+                    db.query(File)
+                    .join(KnowledgeFile, File.id == KnowledgeFile.file_id)
+                    .filter(KnowledgeFile.knowledge_id == knowledge_id)
+                    .all()
+                )
+                return [FileModel.model_validate(file) for file in files]
+        except Exception:
+            return []
+
+    def get_file_metadatas_by_id(
+        self, knowledge_id: str, db: Optional[Session] = None
+    ) -> list[FileMetadataResponse]:
+        try:
+            with get_db_context(db) as db:
+                files = self.get_files_by_id(knowledge_id, db=db)
+                return [FileMetadataResponse(**file.model_dump()) for file in files]
+        except Exception:
+            return []
+
+    def add_file_to_knowledge_by_id(
+        self,
+        knowledge_id: str,
+        file_id: str,
+        user_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[KnowledgeFileModel]:
+        with get_db_context(db) as db:
+            knowledge_file = KnowledgeFileModel(
+                **{
+                    "id": str(uuid.uuid4()),
+                    "knowledge_id": knowledge_id,
+                    "file_id": file_id,
+                    "user_id": user_id,
+                    "created_at": int(time.time()),
+                    "updated_at": int(time.time()),
+                }
+            )
+
+            try:
+                result = KnowledgeFile(**knowledge_file.model_dump())
+                db.add(result)
+                db.commit()
+                db.refresh(result)
+                if result:
+                    return KnowledgeFileModel.model_validate(result)
+                else:
+                    return None
+            except Exception:
+                return None
+
+    def remove_file_from_knowledge_by_id(
+        self, knowledge_id: str, file_id: str, db: Optional[Session] = None
+    ) -> bool:
+        try:
+            with get_db_context(db) as db:
+                db.query(KnowledgeFile).filter_by(
+                    knowledge_id=knowledge_id, file_id=file_id
+                ).delete()
+                db.commit()
+                return True
+        except Exception:
+            return False
+
+    def reset_knowledge_by_id(
+        self, id: str, db: Optional[Session] = None
     ) -> Optional[KnowledgeModel]:
         try:
-            with get_db() as db:
-                knowledge = self.get_knowledge_by_id(id=id)
+            with get_db_context(db) as db:
+                # Delete all knowledge_file entries for this knowledge_id
+                db.query(KnowledgeFile).filter_by(knowledge_id=id).delete()
+                db.commit()
+
+                # Update the knowledge entry's updated_at timestamp
                 db.query(Knowledge).filter_by(id=id).update(
                     {
-                        **form_data.model_dump(),
                         "updated_at": int(time.time()),
                     }
                 )
                 db.commit()
-                return self.get_knowledge_by_id(id=id)
+
+                return self.get_knowledge_by_id(id=id, db=db)
+        except Exception as e:
+            log.exception(e)
+            return None
+
+    def update_knowledge_by_id(
+        self,
+        id: str,
+        form_data: KnowledgeForm,
+        overwrite: bool = False,
+        db: Optional[Session] = None,
+    ) -> Optional[KnowledgeModel]:
+        try:
+            with get_db_context(db) as db:
+                knowledge = self.get_knowledge_by_id(id=id, db=db)
+                db.query(Knowledge).filter_by(id=id).update(
+                    {
+                        **form_data.model_dump(exclude={"access_grants"}),
+                        "updated_at": int(time.time()),
+                    }
+                )
+                db.commit()
+                if form_data.access_grants is not None:
+                    AccessGrants.set_access_grants(
+                        "knowledge", id, form_data.access_grants, db=db
+                    )
+                return self.get_knowledge_by_id(id=id, db=db)
         except Exception as e:
             log.exception(e)
             return None
 
     def update_knowledge_data_by_id(
-        self, id: str, data: dict
+        self, id: str, data: dict, db: Optional[Session] = None
     ) -> Optional[KnowledgeModel]:
         try:
-            with get_db() as db:
-                knowledge = self.get_knowledge_by_id(id=id)
+            with get_db_context(db) as db:
+                knowledge = self.get_knowledge_by_id(id=id, db=db)
                 db.query(Knowledge).filter_by(id=id).update(
                     {
                         "data": data,
@@ -213,23 +659,27 @@ class KnowledgeTable:
                     }
                 )
                 db.commit()
-                return self.get_knowledge_by_id(id=id)
+                return self.get_knowledge_by_id(id=id, db=db)
         except Exception as e:
             log.exception(e)
             return None
 
-    def delete_knowledge_by_id(self, id: str) -> bool:
+    def delete_knowledge_by_id(self, id: str, db: Optional[Session] = None) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
+                AccessGrants.revoke_all_access("knowledge", id, db=db)
                 db.query(Knowledge).filter_by(id=id).delete()
                 db.commit()
                 return True
         except Exception:
             return False
 
-    def delete_all_knowledge(self) -> bool:
-        with get_db() as db:
+    def delete_all_knowledge(self, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             try:
+                knowledge_ids = [row[0] for row in db.query(Knowledge.id).all()]
+                for knowledge_id in knowledge_ids:
+                    AccessGrants.revoke_all_access("knowledge", knowledge_id, db=db)
                 db.query(Knowledge).delete()
                 db.commit()
 

backend/open_webui/models/memories.py

@@ -2,7 +2,8 @@ import time
 import uuid
 from typing import Optional
 
-from open_webui.internal.db import Base, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db, get_db_context
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Column, String, Text
 
@@ -14,7 +15,7 @@ from sqlalchemy import BigInteger, Column, String, Text
 class Memory(Base):
     __tablename__ = "memory"
 
-    id = Column(String, primary_key=True)
+    id = Column(String, primary_key=True, unique=True)
     user_id = Column(String)
     content = Column(Text)
     updated_at = Column(BigInteger)
@@ -41,8 +42,9 @@ class MemoriesTable:
         self,
         user_id: str,
         content: str,
+        db: Optional[Session] = None,
     ) -> Optional[MemoryModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             id = str(uuid.uuid4())
 
             memory = MemoryModel(
@@ -68,8 +70,9 @@ class MemoriesTable:
         id: str,
         user_id: str,
         content: str,
+        db: Optional[Session] = None,
     ) -> Optional[MemoryModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             try:
                 memory = db.get(Memory, id)
                 if not memory or memory.user_id != user_id:
@@ -79,36 +82,41 @@ class MemoriesTable:
                 memory.updated_at = int(time.time())
 
                 db.commit()
-                return self.get_memory_by_id(id)
+                db.refresh(memory)
+                return MemoryModel.model_validate(memory)
             except Exception:
                 return None
 
-    def get_memories(self) -> list[MemoryModel]:
-        with get_db() as db:
+    def get_memories(self, db: Optional[Session] = None) -> list[MemoryModel]:
+        with get_db_context(db) as db:
             try:
                 memories = db.query(Memory).all()
                 return [MemoryModel.model_validate(memory) for memory in memories]
             except Exception:
                 return None
 
-    def get_memories_by_user_id(self, user_id: str) -> list[MemoryModel]:
-        with get_db() as db:
+    def get_memories_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> list[MemoryModel]:
+        with get_db_context(db) as db:
             try:
                 memories = db.query(Memory).filter_by(user_id=user_id).all()
                 return [MemoryModel.model_validate(memory) for memory in memories]
             except Exception:
                 return None
 
-    def get_memory_by_id(self, id: str) -> Optional[MemoryModel]:
-        with get_db() as db:
+    def get_memory_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[MemoryModel]:
+        with get_db_context(db) as db:
             try:
                 memory = db.get(Memory, id)
                 return MemoryModel.model_validate(memory)
             except Exception:
                 return None
 
-    def delete_memory_by_id(self, id: str) -> bool:
-        with get_db() as db:
+    def delete_memory_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             try:
                 db.query(Memory).filter_by(id=id).delete()
                 db.commit()
@@ -118,8 +126,10 @@ class MemoriesTable:
             except Exception:
                 return False
 
-    def delete_memories_by_user_id(self, user_id: str) -> bool:
-        with get_db() as db:
+    def delete_memories_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
             try:
                 db.query(Memory).filter_by(user_id=user_id).delete()
                 db.commit()
@@ -128,8 +138,10 @@ class MemoriesTable:
             except Exception:
                 return False
 
-    def delete_memory_by_id_and_user_id(self, id: str, user_id: str) -> bool:
-        with get_db() as db:
+    def delete_memory_by_id_and_user_id(
+        self, id: str, user_id: str, db: Optional[Session] = None
+    ) -> bool:
+        with get_db_context(db) as db:
             try:
                 memory = db.get(Memory, id)
                 if not memory or memory.user_id != user_id:

backend/open_webui/models/messages.py

@@ -3,12 +3,14 @@ import time
 import uuid
 from typing import Optional
 
-from open_webui.internal.db import Base, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 from open_webui.models.tags import TagModel, Tag, Tags
-from open_webui.models.users import Users, UserNameResponse
+from open_webui.models.users import Users, User, UserNameResponse
+from open_webui.models.channels import Channels, ChannelMember
 
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, ConfigDict, field_validator
 from sqlalchemy import BigInteger, Boolean, Column, String, Text, JSON
 from sqlalchemy import or_, func, select, and_, text
 from sqlalchemy.sql import exists
@@ -20,7 +22,7 @@ from sqlalchemy.sql import exists
 
 class MessageReaction(Base):
     __tablename__ = "message_reaction"
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
     user_id = Column(Text)
     message_id = Column(Text)
     name = Column(Text)
@@ -39,7 +41,7 @@ class MessageReactionModel(BaseModel):
 
 class Message(Base):
     __tablename__ = "message"
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
 
     user_id = Column(Text)
     channel_id = Column(Text, nullable=True)
@@ -47,6 +49,11 @@ class Message(Base):
     reply_to_id = Column(Text, nullable=True)
     parent_id = Column(Text, nullable=True)
 
+    # Pins
+    is_pinned = Column(Boolean, nullable=False, default=False)
+    pinned_at = Column(BigInteger, nullable=True)
+    pinned_by = Column(Text, nullable=True)
+
     content = Column(Text)
     data = Column(JSON, nullable=True)
     meta = Column(JSON, nullable=True)
@@ -65,12 +72,17 @@ class MessageModel(BaseModel):
     reply_to_id: Optional[str] = None
     parent_id: Optional[str] = None
 
+    # Pins
+    is_pinned: bool = False
+    pinned_by: Optional[str] = None
+    pinned_at: Optional[int] = None  # timestamp in epoch (time_ns)
+
     content: str
     data: Optional[dict] = None
     meta: Optional[dict] = None
 
-    created_at: int  # timestamp in epoch
-    updated_at: int  # timestamp in epoch
+    created_at: int  # timestamp in epoch (time_ns)
+    updated_at: int  # timestamp in epoch (time_ns)
 
 
 ####################
@@ -79,6 +91,7 @@ class MessageModel(BaseModel):
 
 
 class MessageForm(BaseModel):
+    temp_id: Optional[str] = None
     content: str
     reply_to_id: Optional[str] = None
     parent_id: Optional[str] = None
@@ -88,7 +101,7 @@ class MessageForm(BaseModel):
 
 class Reactions(BaseModel):
     name: str
-    user_ids: list[str]
+    users: list[dict]
     count: int
 
 
@@ -96,8 +109,25 @@ class MessageUserResponse(MessageModel):
     user: Optional[UserNameResponse] = None
 
 
+class MessageUserSlimResponse(MessageUserResponse):
+    data: bool | None = None
+
+    @field_validator("data", mode="before")
+    def convert_data_to_bool(cls, v):
+        # No data or not a dict → False
+        if not isinstance(v, dict):
+            return False
+
+        # True if ANY value in the dict is non-empty
+        return any(bool(val) for val in v.values())
+
+
 class MessageReplyToResponse(MessageUserResponse):
-    reply_to_message: Optional[MessageUserResponse] = None
+    reply_to_message: Optional[MessageUserSlimResponse] = None
+
+
+class MessageWithReactionsResponse(MessageUserSlimResponse):
+    reactions: list[Reactions]
 
 
 class MessageResponse(MessageReplyToResponse):
@@ -108,12 +138,18 @@ class MessageResponse(MessageReplyToResponse):
 
 class MessageTable:
     def insert_new_message(
-        self, form_data: MessageForm, channel_id: str, user_id: str
+        self,
+        form_data: MessageForm,
+        channel_id: str,
+        user_id: str,
+        db: Optional[Session] = None,
     ) -> Optional[MessageModel]:
-        with get_db() as db:
-            id = str(uuid.uuid4())
+        with get_db_context(db) as db:
+            channel_member = Channels.join_channel(channel_id, user_id)
 
+            id = str(uuid.uuid4())
             ts = int(time.time_ns())
+
             message = MessageModel(
                 **{
                     "id": id,
@@ -121,6 +157,9 @@ class MessageTable:
                     "channel_id": channel_id,
                     "reply_to_id": form_data.reply_to_id,
                     "parent_id": form_data.parent_id,
+                    "is_pinned": False,
+                    "pinned_at": None,
+                    "pinned_by": None,
                     "content": form_data.content,
                     "data": form_data.data,
                     "meta": form_data.meta,
@@ -128,33 +167,64 @@ class MessageTable:
                     "updated_at": ts,
                 }
             )
-
             result = Message(**message.model_dump())
+
             db.add(result)
             db.commit()
             db.refresh(result)
             return MessageModel.model_validate(result) if result else None
 
-    def get_message_by_id(self, id: str) -> Optional[MessageResponse]:
-        with get_db() as db:
+    def get_message_by_id(
+        self,
+        id: str,
+        include_thread_replies: Optional[bool] = True,
+        db: Optional[Session] = None,
+    ) -> Optional[MessageResponse]:
+        with get_db_context(db) as db:
             message = db.get(Message, id)
             if not message:
                 return None
 
             reply_to_message = (
-                self.get_message_by_id(message.reply_to_id)
+                self.get_message_by_id(
+                    message.reply_to_id, include_thread_replies=False, db=db
+                )
                 if message.reply_to_id
                 else None
             )
 
-            reactions = self.get_reactions_by_message_id(id)
-            thread_replies = self.get_thread_replies_by_message_id(id)
+            reactions = self.get_reactions_by_message_id(id, db=db)
+
+            thread_replies = []
+            if include_thread_replies:
+                thread_replies = self.get_thread_replies_by_message_id(id, db=db)
+
+            # Check if message was sent by webhook (webhook info in meta takes precedence)
+            webhook_info = message.meta.get("webhook") if message.meta else None
+            if webhook_info and webhook_info.get("id"):
+                # Look up webhook by ID to get current name
+                webhook = Channels.get_webhook_by_id(webhook_info.get("id"), db=db)
+                if webhook:
+                    user_info = {
+                        "id": webhook.id,
+                        "name": webhook.name,
+                        "role": "webhook",
+                    }
+                else:
+                    # Webhook was deleted, use placeholder
+                    user_info = {
+                        "id": webhook_info.get("id"),
+                        "name": "Deleted Webhook",
+                        "role": "webhook",
+                    }
+            else:
+                user = Users.get_user_by_id(message.user_id, db=db)
+                user_info = user.model_dump() if user else None
 
-            user = Users.get_user_by_id(message.user_id)
             return MessageResponse.model_validate(
                 {
                     **MessageModel.model_validate(message).model_dump(),
-                    "user": user.model_dump() if user else None,
+                    "user": user_info,
                     "reply_to_message": (
                         reply_to_message.model_dump() if reply_to_message else None
                     ),
@@ -166,8 +236,10 @@ class MessageTable:
                 }
             )
 
-    def get_thread_replies_by_message_id(self, id: str) -> list[MessageReplyToResponse]:
-        with get_db() as db:
+    def get_thread_replies_by_message_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> list[MessageReplyToResponse]:
+        with get_db_context(db) as db:
             all_messages = (
                 db.query(Message)
                 .filter_by(parent_id=id)
@@ -178,14 +250,35 @@ class MessageTable:
             messages = []
             for message in all_messages:
                 reply_to_message = (
-                    self.get_message_by_id(message.reply_to_id)
+                    self.get_message_by_id(
+                        message.reply_to_id, include_thread_replies=False, db=db
+                    )
                     if message.reply_to_id
                     else None
                 )
+
+                webhook_info = message.meta.get("webhook") if message.meta else None
+                user_info = None
+                if webhook_info and webhook_info.get("id"):
+                    webhook = Channels.get_webhook_by_id(webhook_info.get("id"), db=db)
+                    if webhook:
+                        user_info = {
+                            "id": webhook.id,
+                            "name": webhook.name,
+                            "role": "webhook",
+                        }
+                    else:
+                        user_info = {
+                            "id": webhook_info.get("id"),
+                            "name": "Deleted Webhook",
+                            "role": "webhook",
+                        }
+
                 messages.append(
                     MessageReplyToResponse.model_validate(
                         {
                             **MessageModel.model_validate(message).model_dump(),
+                            "user": user_info,
                             "reply_to_message": (
                                 reply_to_message.model_dump()
                                 if reply_to_message
@@ -196,17 +289,23 @@ class MessageTable:
                 )
             return messages
 
-    def get_reply_user_ids_by_message_id(self, id: str) -> list[str]:
-        with get_db() as db:
+    def get_reply_user_ids_by_message_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> list[str]:
+        with get_db_context(db) as db:
             return [
                 message.user_id
                 for message in db.query(Message).filter_by(parent_id=id).all()
             ]
 
     def get_messages_by_channel_id(
-        self, channel_id: str, skip: int = 0, limit: int = 50
+        self,
+        channel_id: str,
+        skip: int = 0,
+        limit: int = 50,
+        db: Optional[Session] = None,
     ) -> list[MessageReplyToResponse]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             all_messages = (
                 db.query(Message)
                 .filter_by(channel_id=channel_id, parent_id=None)
@@ -219,14 +318,35 @@ class MessageTable:
             messages = []
             for message in all_messages:
                 reply_to_message = (
-                    self.get_message_by_id(message.reply_to_id)
+                    self.get_message_by_id(
+                        message.reply_to_id, include_thread_replies=False, db=db
+                    )
                     if message.reply_to_id
                     else None
                 )
+
+                webhook_info = message.meta.get("webhook") if message.meta else None
+                user_info = None
+                if webhook_info and webhook_info.get("id"):
+                    webhook = Channels.get_webhook_by_id(webhook_info.get("id"), db=db)
+                    if webhook:
+                        user_info = {
+                            "id": webhook.id,
+                            "name": webhook.name,
+                            "role": "webhook",
+                        }
+                    else:
+                        user_info = {
+                            "id": webhook_info.get("id"),
+                            "name": "Deleted Webhook",
+                            "role": "webhook",
+                        }
+
                 messages.append(
                     MessageReplyToResponse.model_validate(
                         {
                             **MessageModel.model_validate(message).model_dump(),
+                            "user": user_info,
                             "reply_to_message": (
                                 reply_to_message.model_dump()
                                 if reply_to_message
@@ -238,9 +358,14 @@ class MessageTable:
             return messages
 
     def get_messages_by_parent_id(
-        self, channel_id: str, parent_id: str, skip: int = 0, limit: int = 50
+        self,
+        channel_id: str,
+        parent_id: str,
+        skip: int = 0,
+        limit: int = 50,
+        db: Optional[Session] = None,
     ) -> list[MessageReplyToResponse]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             message = db.get(Message, parent_id)
 
             if not message:
@@ -262,14 +387,35 @@ class MessageTable:
             messages = []
             for message in all_messages:
                 reply_to_message = (
-                    self.get_message_by_id(message.reply_to_id)
+                    self.get_message_by_id(
+                        message.reply_to_id, include_thread_replies=False, db=db
+                    )
                     if message.reply_to_id
                     else None
                 )
+
+                webhook_info = message.meta.get("webhook") if message.meta else None
+                user_info = None
+                if webhook_info and webhook_info.get("id"):
+                    webhook = Channels.get_webhook_by_id(webhook_info.get("id"), db=db)
+                    if webhook:
+                        user_info = {
+                            "id": webhook.id,
+                            "name": webhook.name,
+                            "role": "webhook",
+                        }
+                    else:
+                        user_info = {
+                            "id": webhook_info.get("id"),
+                            "name": "Deleted Webhook",
+                            "role": "webhook",
+                        }
+
                 messages.append(
                     MessageReplyToResponse.model_validate(
                         {
                             **MessageModel.model_validate(message).model_dump(),
+                            "user": user_info,
                             "reply_to_message": (
                                 reply_to_message.model_dump()
                                 if reply_to_message
@@ -280,10 +426,40 @@ class MessageTable:
                 )
             return messages
 
-    def update_message_by_id(
-        self, id: str, form_data: MessageForm
+    def get_last_message_by_channel_id(
+        self, channel_id: str, db: Optional[Session] = None
     ) -> Optional[MessageModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
+            message = (
+                db.query(Message)
+                .filter_by(channel_id=channel_id)
+                .order_by(Message.created_at.desc())
+                .first()
+            )
+            return MessageModel.model_validate(message) if message else None
+
+    def get_pinned_messages_by_channel_id(
+        self,
+        channel_id: str,
+        skip: int = 0,
+        limit: int = 50,
+        db: Optional[Session] = None,
+    ) -> list[MessageModel]:
+        with get_db_context(db) as db:
+            all_messages = (
+                db.query(Message)
+                .filter_by(channel_id=channel_id, is_pinned=True)
+                .order_by(Message.pinned_at.desc())
+                .offset(skip)
+                .limit(limit)
+                .all()
+            )
+            return [MessageModel.model_validate(message) for message in all_messages]
+
+    def update_message_by_id(
+        self, id: str, form_data: MessageForm, db: Optional[Session] = None
+    ) -> Optional[MessageModel]:
+        with get_db_context(db) as db:
             message = db.get(Message, id)
             message.content = form_data.content
             message.data = {
@@ -299,10 +475,52 @@ class MessageTable:
             db.refresh(message)
             return MessageModel.model_validate(message) if message else None
 
+    def update_is_pinned_by_id(
+        self,
+        id: str,
+        is_pinned: bool,
+        pinned_by: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> Optional[MessageModel]:
+        with get_db_context(db) as db:
+            message = db.get(Message, id)
+            message.is_pinned = is_pinned
+            message.pinned_at = int(time.time_ns()) if is_pinned else None
+            message.pinned_by = pinned_by if is_pinned else None
+            db.commit()
+            db.refresh(message)
+            return MessageModel.model_validate(message) if message else None
+
+    def get_unread_message_count(
+        self,
+        channel_id: str,
+        user_id: str,
+        last_read_at: Optional[int] = None,
+        db: Optional[Session] = None,
+    ) -> int:
+        with get_db_context(db) as db:
+            query = db.query(Message).filter(
+                Message.channel_id == channel_id,
+                Message.parent_id == None,  # only count top-level messages
+                Message.created_at > (last_read_at if last_read_at else 0),
+            )
+            if user_id:
+                query = query.filter(Message.user_id != user_id)
+            return query.count()
+
     def add_reaction_to_message(
-        self, id: str, user_id: str, name: str
+        self, id: str, user_id: str, name: str, db: Optional[Session] = None
     ) -> Optional[MessageReactionModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
+            # check for existing reaction
+            existing_reaction = (
+                db.query(MessageReaction)
+                .filter_by(message_id=id, user_id=user_id, name=name)
+                .first()
+            )
+            if existing_reaction:
+                return MessageReactionModel.model_validate(existing_reaction)
+
             reaction_id = str(uuid.uuid4())
             reaction = MessageReactionModel(
                 id=reaction_id,
@@ -317,47 +535,62 @@ class MessageTable:
             db.refresh(result)
             return MessageReactionModel.model_validate(result) if result else None
 
-    def get_reactions_by_message_id(self, id: str) -> list[Reactions]:
-        with get_db() as db:
-            all_reactions = db.query(MessageReaction).filter_by(message_id=id).all()
+    def get_reactions_by_message_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> list[Reactions]:
+        with get_db_context(db) as db:
+            # JOIN User so all user info is fetched in one query
+            results = (
+                db.query(MessageReaction, User)
+                .join(User, MessageReaction.user_id == User.id)
+                .filter(MessageReaction.message_id == id)
+                .all()
+            )
 
             reactions = {}
-            for reaction in all_reactions:
+
+            for reaction, user in results:
                 if reaction.name not in reactions:
                     reactions[reaction.name] = {
                         "name": reaction.name,
-                        "user_ids": [],
+                        "users": [],
                         "count": 0,
                     }
-                reactions[reaction.name]["user_ids"].append(reaction.user_id)
+
+                reactions[reaction.name]["users"].append(
+                    {
+                        "id": user.id,
+                        "name": user.name,
+                    }
+                )
                 reactions[reaction.name]["count"] += 1
 
             return [Reactions(**reaction) for reaction in reactions.values()]
 
     def remove_reaction_by_id_and_user_id_and_name(
-        self, id: str, user_id: str, name: str
+        self, id: str, user_id: str, name: str, db: Optional[Session] = None
     ) -> bool:
-        with get_db() as db:
+        with get_db_context(db) as db:
             db.query(MessageReaction).filter_by(
                 message_id=id, user_id=user_id, name=name
             ).delete()
             db.commit()
             return True
 
-    def delete_reactions_by_id(self, id: str) -> bool:
-        with get_db() as db:
+    def delete_reactions_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             db.query(MessageReaction).filter_by(message_id=id).delete()
             db.commit()
             return True
 
-    def delete_replies_by_id(self, id: str) -> bool:
-        with get_db() as db:
+    def delete_replies_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             db.query(Message).filter_by(parent_id=id).delete()
             db.commit()
             return True
 
-    def delete_message_by_id(self, id: str) -> bool:
-        with get_db() as db:
+    def delete_message_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             db.query(Message).filter_by(id=id).delete()
 
             # Delete all reactions to this message
@@ -366,5 +599,35 @@ class MessageTable:
             db.commit()
             return True
 
+    def search_messages_by_channel_ids(
+        self,
+        channel_ids: list[str],
+        query: str,
+        start_timestamp: Optional[int] = None,
+        end_timestamp: Optional[int] = None,
+        limit: int = 10,
+        db: Optional[Session] = None,
+    ) -> list[MessageModel]:
+        """Search messages in specified channels by content."""
+        with get_db_context(db) as db:
+            query_builder = db.query(Message).filter(
+                Message.channel_id.in_(channel_ids),
+                Message.content.ilike(f"%{query}%"),
+            )
+
+            if start_timestamp:
+                query_builder = query_builder.filter(
+                    Message.created_at >= start_timestamp
+                )
+            if end_timestamp:
+                query_builder = query_builder.filter(
+                    Message.created_at <= end_timestamp
+                )
+
+            messages = (
+                query_builder.order_by(Message.created_at.desc()).limit(limit).all()
+            )
+            return [MessageModel.model_validate(msg) for msg in messages]
+
 
 Messages = MessageTable()

backend/open_webui/models/models.py

@@ -2,25 +2,23 @@ import logging
 import time
 from typing import Optional
 
-from open_webui.internal.db import Base, JSONField, get_db
-from open_webui.env import SRC_LOG_LEVELS
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 
 from open_webui.models.groups import Groups
-from open_webui.models.users import Users, UserResponse
+from open_webui.models.users import User, UserModel, Users, UserResponse
+from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, ConfigDict, Field
 
-from sqlalchemy import or_, and_, func
+from sqlalchemy import String, cast, or_, and_, func
 from sqlalchemy.dialects import postgresql, sqlite
-from sqlalchemy import BigInteger, Column, Text, JSON, Boolean
-
-
-from open_webui.utils.access_control import has_access
 
+from sqlalchemy.dialects.postgresql import JSONB
+from sqlalchemy import BigInteger, Column, Text, Boolean
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 
 ####################
@@ -53,7 +51,7 @@ class ModelMeta(BaseModel):
 class Model(Base):
     __tablename__ = "model"
 
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
     """
         The model's id as used in the API. If set to an existing model, it will override the model.
     """
@@ -79,23 +77,6 @@ class Model(Base):
         Holds a JSON encoded blob of metadata, see `ModelMeta`.
     """
 
-    access_control = Column(JSON, nullable=True)  # Controls data access levels.
-    # Defines access control rules for this entry.
-    # - `None`: Public access, available to all users with the "user" role.
-    # - `{}`: Private access, restricted exclusively to the owner.
-    # - Custom permissions: Specific access control for reading and writing;
-    #   Can specify group or user-level restrictions:
-    #   {
-    #      "read": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      },
-    #      "write": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      }
-    #   }
-
     is_active = Column(Boolean, default=True)
 
     updated_at = Column(BigInteger)
@@ -111,7 +92,7 @@ class ModelModel(BaseModel):
     params: ModelParams
     meta: ModelMeta
 
-    access_control: Optional[dict] = None
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
 
     is_active: bool
     updated_at: int  # timestamp in epoch
@@ -129,58 +110,88 @@ class ModelUserResponse(ModelModel):
     user: Optional[UserResponse] = None
 
 
+class ModelAccessResponse(ModelUserResponse):
+    write_access: Optional[bool] = False
+
+
 class ModelResponse(ModelModel):
     pass
 
 
+class ModelListResponse(BaseModel):
+    items: list[ModelUserResponse]
+    total: int
+
+
+class ModelAccessListResponse(BaseModel):
+    items: list[ModelAccessResponse]
+    total: int
+
+
 class ModelForm(BaseModel):
     id: str
     base_model_id: Optional[str] = None
     name: str
     meta: ModelMeta
     params: ModelParams
-    access_control: Optional[dict] = None
+    access_grants: Optional[list[dict]] = None
     is_active: bool = True
 
 
 class ModelsTable:
-    def insert_new_model(
-        self, form_data: ModelForm, user_id: str
-    ) -> Optional[ModelModel]:
-        model = ModelModel(
-            **{
-                **form_data.model_dump(),
-                "user_id": user_id,
-                "created_at": int(time.time()),
-                "updated_at": int(time.time()),
-            }
+    def _get_access_grants(
+        self, model_id: str, db: Optional[Session] = None
+    ) -> list[AccessGrantModel]:
+        return AccessGrants.get_grants_by_resource("model", model_id, db=db)
+
+    def _to_model_model(self, model: Model, db: Optional[Session] = None) -> ModelModel:
+        model_data = ModelModel.model_validate(model).model_dump(
+            exclude={"access_grants"}
         )
+        model_data["access_grants"] = self._get_access_grants(model_data["id"], db=db)
+        return ModelModel.model_validate(model_data)
+
+    def insert_new_model(
+        self, form_data: ModelForm, user_id: str, db: Optional[Session] = None
+    ) -> Optional[ModelModel]:
         try:
-            with get_db() as db:
-                result = Model(**model.model_dump())
+            with get_db_context(db) as db:
+                result = Model(
+                    **{
+                        **form_data.model_dump(exclude={"access_grants"}),
+                        "user_id": user_id,
+                        "created_at": int(time.time()),
+                        "updated_at": int(time.time()),
+                    }
+                )
                 db.add(result)
                 db.commit()
                 db.refresh(result)
+                AccessGrants.set_access_grants(
+                    "model", result.id, form_data.access_grants, db=db
+                )
 
                 if result:
-                    return ModelModel.model_validate(result)
+                    return self._to_model_model(result, db=db)
                 else:
                     return None
         except Exception as e:
             log.exception(f"Failed to insert a new model: {e}")
             return None
 
-    def get_all_models(self) -> list[ModelModel]:
-        with get_db() as db:
-            return [ModelModel.model_validate(model) for model in db.query(Model).all()]
+    def get_all_models(self, db: Optional[Session] = None) -> list[ModelModel]:
+        with get_db_context(db) as db:
+            return [
+                self._to_model_model(model, db=db) for model in db.query(Model).all()
+            ]
 
-    def get_models(self) -> list[ModelUserResponse]:
-        with get_db() as db:
+    def get_models(self, db: Optional[Session] = None) -> list[ModelUserResponse]:
+        with get_db_context(db) as db:
             all_models = db.query(Model).filter(Model.base_model_id != None).all()
 
             user_ids = list(set(model.user_id for model in all_models))
 
-            users = Users.get_users_by_user_ids(user_ids) if user_ids else []
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
             users_dict = {user.id: user for user in users}
 
             models = []
@@ -189,78 +200,208 @@ class ModelsTable:
                 models.append(
                     ModelUserResponse.model_validate(
                         {
-                            **ModelModel.model_validate(model).model_dump(),
+                            **self._to_model_model(model, db=db).model_dump(),
                             "user": user.model_dump() if user else None,
                         }
                     )
                 )
             return models
 
-    def get_base_models(self) -> list[ModelModel]:
-        with get_db() as db:
+    def get_base_models(self, db: Optional[Session] = None) -> list[ModelModel]:
+        with get_db_context(db) as db:
             return [
-                ModelModel.model_validate(model)
+                self._to_model_model(model, db=db)
                 for model in db.query(Model).filter(Model.base_model_id == None).all()
             ]
 
     def get_models_by_user_id(
-        self, user_id: str, permission: str = "write"
+        self, user_id: str, permission: str = "write", db: Optional[Session] = None
     ) -> list[ModelUserResponse]:
-        models = self.get_models()
-        user_group_ids = {group.id for group in Groups.get_groups_by_member_id(user_id)}
+        models = self.get_models(db=db)
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
         return [
             model
             for model in models
             if model.user_id == user_id
-            or has_access(user_id, permission, model.access_control, user_group_ids)
+            or AccessGrants.has_access(
+                user_id=user_id,
+                resource_type="model",
+                resource_id=model.id,
+                permission=permission,
+                user_group_ids=user_group_ids,
+                db=db,
+            )
         ]
 
-    def get_model_by_id(self, id: str) -> Optional[ModelModel]:
+    def _has_permission(self, db, query, filter: dict, permission: str = "read"):
+        return AccessGrants.has_permission_filter(
+            db=db,
+            query=query,
+            DocumentModel=Model,
+            filter=filter,
+            resource_type="model",
+            permission=permission,
+        )
+
+    def search_models(
+        self,
+        user_id: str,
+        filter: dict = {},
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> ModelListResponse:
+        with get_db_context(db) as db:
+            # Join GroupMember so we can order by group_id when requested
+            query = db.query(Model, User).outerjoin(User, User.id == Model.user_id)
+            query = query.filter(Model.base_model_id != None)
+
+            if filter:
+                query_key = filter.get("query")
+                if query_key:
+                    query = query.filter(
+                        or_(
+                            Model.name.ilike(f"%{query_key}%"),
+                            Model.base_model_id.ilike(f"%{query_key}%"),
+                            User.name.ilike(f"%{query_key}%"),
+                            User.email.ilike(f"%{query_key}%"),
+                            User.username.ilike(f"%{query_key}%"),
+                        )
+                    )
+
+                view_option = filter.get("view_option")
+                if view_option == "created":
+                    query = query.filter(Model.user_id == user_id)
+                elif view_option == "shared":
+                    query = query.filter(Model.user_id != user_id)
+
+                # Apply access control filtering
+                query = self._has_permission(
+                    db,
+                    query,
+                    filter,
+                    permission="read",
+                )
+
+                tag = filter.get("tag")
+                if tag:
+                    # TODO: This is a simple implementation and should be improved for performance
+                    like_pattern = f'%"{tag.lower()}"%'  # `"tag"` inside JSON array
+                    meta_text = func.lower(cast(Model.meta, String))
+
+                    query = query.filter(meta_text.like(like_pattern))
+
+                order_by = filter.get("order_by")
+                direction = filter.get("direction")
+
+                if order_by == "name":
+                    if direction == "asc":
+                        query = query.order_by(Model.name.asc())
+                    else:
+                        query = query.order_by(Model.name.desc())
+                elif order_by == "created_at":
+                    if direction == "asc":
+                        query = query.order_by(Model.created_at.asc())
+                    else:
+                        query = query.order_by(Model.created_at.desc())
+                elif order_by == "updated_at":
+                    if direction == "asc":
+                        query = query.order_by(Model.updated_at.asc())
+                    else:
+                        query = query.order_by(Model.updated_at.desc())
+
+            else:
+                query = query.order_by(Model.created_at.desc())
+
+            # Count BEFORE pagination
+            total = query.count()
+
+            if skip:
+                query = query.offset(skip)
+            if limit:
+                query = query.limit(limit)
+
+            items = query.all()
+
+            models = []
+            for model, user in items:
+                models.append(
+                    ModelUserResponse(
+                        **self._to_model_model(model, db=db).model_dump(),
+                        user=(
+                            UserResponse(**UserModel.model_validate(user).model_dump())
+                            if user
+                            else None
+                        ),
+                    )
+                )
+
+            return ModelListResponse(items=models, total=total)
+
+    def get_model_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[ModelModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 model = db.get(Model, id)
-                return ModelModel.model_validate(model)
+                return self._to_model_model(model, db=db) if model else None
         except Exception:
             return None
 
-    def toggle_model_by_id(self, id: str) -> Optional[ModelModel]:
-        with get_db() as db:
+    def get_models_by_ids(
+        self, ids: list[str], db: Optional[Session] = None
+    ) -> list[ModelModel]:
+        try:
+            with get_db_context(db) as db:
+                models = db.query(Model).filter(Model.id.in_(ids)).all()
+                return [self._to_model_model(model, db=db) for model in models]
+        except Exception:
+            return []
+
+    def toggle_model_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[ModelModel]:
+        with get_db_context(db) as db:
             try:
-                is_active = db.query(Model).filter_by(id=id).first().is_active
+                model = db.query(Model).filter_by(id=id).first()
+                if not model:
+                    return None
 
-                db.query(Model).filter_by(id=id).update(
-                    {
-                        "is_active": not is_active,
-                        "updated_at": int(time.time()),
-                    }
-                )
+                model.is_active = not model.is_active
+                model.updated_at = int(time.time())
                 db.commit()
+                db.refresh(model)
 
-                return self.get_model_by_id(id)
+                return self._to_model_model(model, db=db)
             except Exception:
                 return None
 
-    def update_model_by_id(self, id: str, model: ModelForm) -> Optional[ModelModel]:
+    def update_model_by_id(
+        self, id: str, model: ModelForm, db: Optional[Session] = None
+    ) -> Optional[ModelModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 # update only the fields that are present in the model
-                result = (
-                    db.query(Model)
-                    .filter_by(id=id)
-                    .update(model.model_dump(exclude={"id"}))
-                )
-                db.commit()
+                data = model.model_dump(exclude={"id", "access_grants"})
+                result = db.query(Model).filter_by(id=id).update(data)
 
-                model = db.get(Model, id)
-                db.refresh(model)
-                return ModelModel.model_validate(model)
+                db.commit()
+                if model.access_grants is not None:
+                    AccessGrants.set_access_grants(
+                        "model", id, model.access_grants, db=db
+                    )
+
+                return self.get_model_by_id(id, db=db)
         except Exception as e:
             log.exception(f"Failed to update the model by id {id}: {e}")
             return None
 
-    def delete_model_by_id(self, id: str) -> bool:
+    def delete_model_by_id(self, id: str, db: Optional[Session] = None) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
+                AccessGrants.revoke_all_access("model", id, db=db)
                 db.query(Model).filter_by(id=id).delete()
                 db.commit()
 
@@ -268,9 +409,12 @@ class ModelsTable:
         except Exception:
             return False
 
-    def delete_all_models(self) -> bool:
+    def delete_all_models(self, db: Optional[Session] = None) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
+                model_ids = [row[0] for row in db.query(Model.id).all()]
+                for model_id in model_ids:
+                    AccessGrants.revoke_all_access("model", model_id, db=db)
                 db.query(Model).delete()
                 db.commit()
 
@@ -278,9 +422,11 @@ class ModelsTable:
         except Exception:
             return False
 
-    def sync_models(self, user_id: str, models: list[ModelModel]) -> list[ModelModel]:
+    def sync_models(
+        self, user_id: str, models: list[ModelModel], db: Optional[Session] = None
+    ) -> list[ModelModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 # Get existing models
                 existing_models = db.query(Model).all()
                 existing_ids = {model.id for model in existing_models}
@@ -293,7 +439,7 @@ class ModelsTable:
                     if model.id in existing_ids:
                         db.query(Model).filter_by(id=model.id).update(
                             {
-                                **model.model_dump(),
+                                **model.model_dump(exclude={"access_grants"}),
                                 "user_id": user_id,
                                 "updated_at": int(time.time()),
                             }
@@ -301,22 +447,27 @@ class ModelsTable:
                     else:
                         new_model = Model(
                             **{
-                                **model.model_dump(),
+                                **model.model_dump(exclude={"access_grants"}),
                                 "user_id": user_id,
                                 "updated_at": int(time.time()),
                             }
                         )
                         db.add(new_model)
+                    AccessGrants.set_access_grants(
+                        "model", model.id, model.access_grants, db=db
+                    )
 
                 # Remove models that are no longer present
                 for model in existing_models:
                     if model.id not in new_model_ids:
+                        AccessGrants.revoke_all_access("model", model.id, db=db)
                         db.delete(model)
 
                 db.commit()
 
                 return [
-                    ModelModel.model_validate(model) for model in db.query(Model).all()
+                    self._to_model_model(model, db=db)
+                    for model in db.query(Model).all()
                 ]
         except Exception as e:
             log.exception(f"Error syncing models for user {user_id}: {e}")

backend/open_webui/models/notes.py

@@ -4,16 +4,16 @@ import uuid
 from typing import Optional
 from functools import lru_cache
 
-from open_webui.internal.db import Base, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db, get_db_context
 from open_webui.models.groups import Groups
-from open_webui.utils.access_control import has_access
-from open_webui.models.users import Users, UserResponse
+from open_webui.models.users import User, UserModel, Users, UserResponse
+from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 
 
-from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Boolean, Column, String, Text, JSON
-from sqlalchemy import or_, func, select, and_, text
-from sqlalchemy.sql import exists
+from pydantic import BaseModel, ConfigDict, Field
+from sqlalchemy import BigInteger, Column, Text, JSON
+from sqlalchemy import or_, func, cast
 
 ####################
 # Note DB Schema
@@ -23,15 +23,13 @@ from sqlalchemy.sql import exists
 class Note(Base):
     __tablename__ = "note"
 
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
     user_id = Column(Text)
 
     title = Column(Text)
     data = Column(JSON, nullable=True)
     meta = Column(JSON, nullable=True)
 
-    access_control = Column(JSON, nullable=True)
-
     created_at = Column(BigInteger)
     updated_at = Column(BigInteger)
 
@@ -46,7 +44,7 @@ class NoteModel(BaseModel):
     data: Optional[dict] = None
     meta: Optional[dict] = None
 
-    access_control: Optional[dict] = None
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
 
     created_at: int  # timestamp in epoch
     updated_at: int  # timestamp in epoch
@@ -61,64 +59,206 @@ class NoteForm(BaseModel):
     title: str
     data: Optional[dict] = None
     meta: Optional[dict] = None
-    access_control: Optional[dict] = None
+    access_grants: Optional[list[dict]] = None
 
 
 class NoteUpdateForm(BaseModel):
     title: Optional[str] = None
     data: Optional[dict] = None
     meta: Optional[dict] = None
-    access_control: Optional[dict] = None
+    access_grants: Optional[list[dict]] = None
 
 
 class NoteUserResponse(NoteModel):
     user: Optional[UserResponse] = None
 
 
+class NoteItemResponse(BaseModel):
+    id: str
+    title: str
+    data: Optional[dict]
+    updated_at: int
+    created_at: int
+    user: Optional[UserResponse] = None
+
+
+class NoteListResponse(BaseModel):
+    items: list[NoteUserResponse]
+    total: int
+
+
 class NoteTable:
+    def _get_access_grants(
+        self, note_id: str, db: Optional[Session] = None
+    ) -> list[AccessGrantModel]:
+        return AccessGrants.get_grants_by_resource("note", note_id, db=db)
+
+    def _to_note_model(self, note: Note, db: Optional[Session] = None) -> NoteModel:
+        note_data = NoteModel.model_validate(note).model_dump(exclude={"access_grants"})
+        note_data["access_grants"] = self._get_access_grants(note_data["id"], db=db)
+        return NoteModel.model_validate(note_data)
+
+    def _has_permission(self, db, query, filter: dict, permission: str = "read"):
+        return AccessGrants.has_permission_filter(
+            db=db,
+            query=query,
+            DocumentModel=Note,
+            filter=filter,
+            resource_type="note",
+            permission=permission,
+        )
+
     def insert_new_note(
-        self,
-        form_data: NoteForm,
-        user_id: str,
+        self, user_id: str, form_data: NoteForm, db: Optional[Session] = None
     ) -> Optional[NoteModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             note = NoteModel(
                 **{
                     "id": str(uuid.uuid4()),
                     "user_id": user_id,
-                    **form_data.model_dump(),
+                    **form_data.model_dump(exclude={"access_grants"}),
                     "created_at": int(time.time_ns()),
                     "updated_at": int(time.time_ns()),
+                    "access_grants": [],
                 }
             )
 
-            new_note = Note(**note.model_dump())
+            new_note = Note(**note.model_dump(exclude={"access_grants"}))
 
             db.add(new_note)
             db.commit()
-            return note
+            AccessGrants.set_access_grants(
+                "note", note.id, form_data.access_grants, db=db
+            )
+            return self._to_note_model(new_note, db=db)
 
     def get_notes(
-        self, skip: Optional[int] = None, limit: Optional[int] = None
+        self, skip: int = 0, limit: int = 50, db: Optional[Session] = None
     ) -> list[NoteModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             query = db.query(Note).order_by(Note.updated_at.desc())
             if skip is not None:
                 query = query.offset(skip)
             if limit is not None:
                 query = query.limit(limit)
             notes = query.all()
-            return [NoteModel.model_validate(note) for note in notes]
+            return [self._to_note_model(note, db=db) for note in notes]
+
+    def search_notes(
+        self,
+        user_id: str,
+        filter: dict = {},
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> NoteListResponse:
+        with get_db_context(db) as db:
+            query = db.query(Note, User).outerjoin(User, User.id == Note.user_id)
+            if filter:
+                query_key = filter.get("query")
+                if query_key:
+                    # Normalize search by removing hyphens and spaces (e.g., "todo" matches "to-do" and "to do")
+                    normalized_query = query_key.replace("-", "").replace(" ", "")
+                    query = query.filter(
+                        or_(
+                            func.replace(
+                                func.replace(Note.title, "-", ""), " ", ""
+                            ).ilike(f"%{normalized_query}%"),
+                            func.replace(
+                                func.replace(
+                                    cast(Note.data["content"]["md"], Text), "-", ""
+                                ),
+                                " ",
+                                "",
+                            ).ilike(f"%{normalized_query}%"),
+                        )
+                    )
+
+                view_option = filter.get("view_option")
+                if view_option == "created":
+                    query = query.filter(Note.user_id == user_id)
+                elif view_option == "shared":
+                    query = query.filter(Note.user_id != user_id)
+
+                # Apply access control filtering
+                if "permission" in filter:
+                    permission = filter["permission"]
+                else:
+                    permission = "write"
+
+                query = self._has_permission(
+                    db,
+                    query,
+                    filter,
+                    permission=permission,
+                )
+
+                order_by = filter.get("order_by")
+                direction = filter.get("direction")
+
+                if order_by == "name":
+                    if direction == "asc":
+                        query = query.order_by(Note.title.asc())
+                    else:
+                        query = query.order_by(Note.title.desc())
+                elif order_by == "created_at":
+                    if direction == "asc":
+                        query = query.order_by(Note.created_at.asc())
+                    else:
+                        query = query.order_by(Note.created_at.desc())
+                elif order_by == "updated_at":
+                    if direction == "asc":
+                        query = query.order_by(Note.updated_at.asc())
+                    else:
+                        query = query.order_by(Note.updated_at.desc())
+                else:
+                    query = query.order_by(Note.updated_at.desc())
+
+            else:
+                query = query.order_by(Note.updated_at.desc())
+
+            # Count BEFORE pagination
+            total = query.count()
+
+            if skip:
+                query = query.offset(skip)
+            if limit:
+                query = query.limit(limit)
+
+            items = query.all()
+
+            notes = []
+            for note, user in items:
+                notes.append(
+                    NoteUserResponse(
+                        **self._to_note_model(note, db=db).model_dump(),
+                        user=(
+                            UserResponse(**UserModel.model_validate(user).model_dump())
+                            if user
+                            else None
+                        ),
+                    )
+                )
+
+            return NoteListResponse(items=notes, total=total)
 
     def get_notes_by_user_id(
         self,
         user_id: str,
-        skip: Optional[int] = None,
-        limit: Optional[int] = None,
+        permission: str = "read",
+        skip: int = 0,
+        limit: int = 50,
+        db: Optional[Session] = None,
     ) -> list[NoteModel]:
-        with get_db() as db:
-            query = db.query(Note).filter(Note.user_id == user_id)
-            query = query.order_by(Note.updated_at.desc())
+        with get_db_context(db) as db:
+            user_group_ids = [
+                group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+            ]
+
+            query = db.query(Note).order_by(Note.updated_at.desc())
+            query = self._has_permission(
+                db, query, {"user_id": user_id, "group_ids": user_group_ids}, permission
+            )
 
             if skip is not None:
                 query = query.offset(skip)
@@ -126,67 +266,19 @@ class NoteTable:
                 query = query.limit(limit)
 
             notes = query.all()
-            return [NoteModel.model_validate(note) for note in notes]
+            return [self._to_note_model(note, db=db) for note in notes]
 
-    def get_notes_by_permission(
-        self,
-        user_id: str,
-        permission: str = "write",
-        skip: Optional[int] = None,
-        limit: Optional[int] = None,
-    ) -> list[NoteModel]:
-        with get_db() as db:
-            user_groups = Groups.get_groups_by_member_id(user_id)
-            user_group_ids = {group.id for group in user_groups}
-
-            # Order newest-first. We stream to keep memory usage low.
-            query = (
-                db.query(Note)
-                .order_by(Note.updated_at.desc())
-                .execution_options(stream_results=True)
-                .yield_per(256)
-            )
-
-            results: list[NoteModel] = []
-            n_skipped = 0
-
-            for note in query:
-                # Fast-pass #1: owner
-                if note.user_id == user_id:
-                    permitted = True
-                # Fast-pass #2: public/open
-                elif note.access_control is None:
-                    # Technically this should mean public access for both read and write, but we'll only do read for now
-                    # We might want to change this behavior later
-                    permitted = permission == "read"
-                else:
-                    permitted = has_access(
-                        user_id, permission, note.access_control, user_group_ids
-                    )
-
-                if not permitted:
-                    continue
-
-                # Apply skip AFTER permission filtering so it counts only accessible notes
-                if skip and n_skipped < skip:
-                    n_skipped += 1
-                    continue
-
-                results.append(NoteModel.model_validate(note))
-                if limit is not None and len(results) >= limit:
-                    break
-
-            return results
-
-    def get_note_by_id(self, id: str) -> Optional[NoteModel]:
-        with get_db() as db:
+    def get_note_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[NoteModel]:
+        with get_db_context(db) as db:
             note = db.query(Note).filter(Note.id == id).first()
-            return NoteModel.model_validate(note) if note else None
+            return self._to_note_model(note, db=db) if note else None
 
     def update_note_by_id(
-        self, id: str, form_data: NoteUpdateForm
+        self, id: str, form_data: NoteUpdateForm, db: Optional[Session] = None
     ) -> Optional[NoteModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             note = db.query(Note).filter(Note.id == id).first()
             if not note:
                 return None
@@ -200,19 +292,25 @@ class NoteTable:
             if "meta" in form_data:
                 note.meta = {**note.meta, **form_data["meta"]}
 
-            if "access_control" in form_data:
-                note.access_control = form_data["access_control"]
+            if "access_grants" in form_data:
+                AccessGrants.set_access_grants(
+                    "note", id, form_data["access_grants"], db=db
+                )
 
             note.updated_at = int(time.time_ns())
 
             db.commit()
-            return NoteModel.model_validate(note) if note else None
+            return self._to_note_model(note, db=db) if note else None
 
-    def delete_note_by_id(self, id: str):
-        with get_db() as db:
-            db.query(Note).filter(Note.id == id).delete()
-            db.commit()
-            return True
+    def delete_note_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        try:
+            with get_db_context(db) as db:
+                AccessGrants.revoke_all_access("note", id, db=db)
+                db.query(Note).filter(Note.id == id).delete()
+                db.commit()
+                return True
+        except Exception:
+            return False
 
 
 Notes = NoteTable()

backend/open_webui/models/oauth_sessions.py

@@ -8,14 +8,14 @@ import json
 
 from cryptography.fernet import Fernet
 
-from open_webui.internal.db import Base, get_db
-from open_webui.env import SRC_LOG_LEVELS, OAUTH_SESSION_TOKEN_ENCRYPTION_KEY
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db, get_db_context
+from open_webui.env import OAUTH_SESSION_TOKEN_ENCRYPTION_KEY
 
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Column, String, Text, Index
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # DB MODEL
@@ -25,7 +25,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 class OAuthSession(Base):
     __tablename__ = "oauth_session"
 
-    id = Column(Text, primary_key=True)
+    id = Column(Text, primary_key=True, unique=True)
     user_id = Column(Text, nullable=False)
     provider = Column(Text, nullable=False)
     token = Column(
@@ -102,7 +102,7 @@ class OAuthSessionTable:
             decrypted = self.fernet.decrypt(token.encode()).decode()
             return json.loads(decrypted)
         except Exception as e:
-            log.error(f"Error decrypting tokens: {e}")
+            log.error(f"Error decrypting tokens: {type(e).__name__}: {e}")
             raise
 
     def create_session(
@@ -110,10 +110,11 @@ class OAuthSessionTable:
         user_id: str,
         provider: str,
         token: dict,
+        db: Optional[Session] = None,
     ) -> Optional[OAuthSessionModel]:
         """Create a new OAuth session"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 current_time = int(time.time())
                 id = str(uuid.uuid4())
 
@@ -142,10 +143,12 @@ class OAuthSessionTable:
             log.error(f"Error creating OAuth session: {e}")
             return None
 
-    def get_session_by_id(self, session_id: str) -> Optional[OAuthSessionModel]:
+    def get_session_by_id(
+        self, session_id: str, db: Optional[Session] = None
+    ) -> Optional[OAuthSessionModel]:
         """Get OAuth session by ID"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 session = db.query(OAuthSession).filter_by(id=session_id).first()
                 if session:
                     session.token = self._decrypt_token(session.token)
@@ -157,11 +160,11 @@ class OAuthSessionTable:
             return None
 
     def get_session_by_id_and_user_id(
-        self, session_id: str, user_id: str
+        self, session_id: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[OAuthSessionModel]:
         """Get OAuth session by ID and user ID"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 session = (
                     db.query(OAuthSession)
                     .filter_by(id=session_id, user_id=user_id)
@@ -177,11 +180,11 @@ class OAuthSessionTable:
             return None
 
     def get_session_by_provider_and_user_id(
-        self, provider: str, user_id: str
+        self, provider: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[OAuthSessionModel]:
         """Get OAuth session by provider and user ID"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 session = (
                     db.query(OAuthSession)
                     .filter_by(provider=provider, user_id=user_id)
@@ -196,16 +199,25 @@ class OAuthSessionTable:
             log.error(f"Error getting OAuth session by provider and user ID: {e}")
             return None
 
-    def get_sessions_by_user_id(self, user_id: str) -> List[OAuthSessionModel]:
+    def get_sessions_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> List[OAuthSessionModel]:
         """Get all OAuth sessions for a user"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 sessions = db.query(OAuthSession).filter_by(user_id=user_id).all()
 
                 results = []
                 for session in sessions:
-                    session.token = self._decrypt_token(session.token)
-                    results.append(OAuthSessionModel.model_validate(session))
+                    try:
+                        session.token = self._decrypt_token(session.token)
+                        results.append(OAuthSessionModel.model_validate(session))
+                    except Exception as e:
+                        log.warning(
+                            f"Skipping OAuth session {session.id} due to decryption failure, deleting corrupted session: {type(e).__name__}: {e}"
+                        )
+                        db.query(OAuthSession).filter_by(id=session.id).delete()
+                        db.commit()
 
                 return results
 
@@ -214,11 +226,11 @@ class OAuthSessionTable:
             return []
 
     def update_session_by_id(
-        self, session_id: str, token: dict
+        self, session_id: str, token: dict, db: Optional[Session] = None
     ) -> Optional[OAuthSessionModel]:
         """Update OAuth session tokens"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 current_time = int(time.time())
 
                 db.query(OAuthSession).filter_by(id=session_id).update(
@@ -240,10 +252,12 @@ class OAuthSessionTable:
             log.error(f"Error updating OAuth session tokens: {e}")
             return None
 
-    def delete_session_by_id(self, session_id: str) -> bool:
+    def delete_session_by_id(
+        self, session_id: str, db: Optional[Session] = None
+    ) -> bool:
         """Delete an OAuth session"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 result = db.query(OAuthSession).filter_by(id=session_id).delete()
                 db.commit()
                 return result > 0
@@ -251,10 +265,12 @@ class OAuthSessionTable:
             log.error(f"Error deleting OAuth session: {e}")
             return False
 
-    def delete_sessions_by_user_id(self, user_id: str) -> bool:
+    def delete_sessions_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> bool:
         """Delete all OAuth sessions for a user"""
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 result = db.query(OAuthSession).filter_by(user_id=user_id).delete()
                 db.commit()
                 return True
@@ -262,5 +278,18 @@ class OAuthSessionTable:
             log.error(f"Error deleting OAuth sessions by user ID: {e}")
             return False
 
+    def delete_sessions_by_provider(
+        self, provider: str, db: Optional[Session] = None
+    ) -> bool:
+        """Delete all OAuth sessions for a provider"""
+        try:
+            with get_db_context(db) as db:
+                db.query(OAuthSession).filter_by(provider=provider).delete()
+                db.commit()
+                return True
+        except Exception as e:
+            log.error(f"Error deleting OAuth sessions by provider {provider}: {e}")
+            return False
+
 
 OAuthSessions = OAuthSessionTable()

backend/open_webui/models/prompt_history.py

@@ -0,0 +1,234 @@
+"""Prompt history model for version tracking."""
+
+import time
+import uuid
+from typing import Optional
+import json
+import difflib
+
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db_context
+from open_webui.models.users import Users, UserResponse
+
+from pydantic import BaseModel, ConfigDict
+from sqlalchemy import BigInteger, Column, Text, JSON, Index
+
+####################
+# PromptHistory DB Schema
+####################
+
+
+class PromptHistory(Base):
+    __tablename__ = "prompt_history"
+
+    id = Column(Text, primary_key=True)
+    prompt_id = Column(Text, nullable=False, index=True)
+    parent_id = Column(Text, nullable=True)  # Reference to parent commit
+    snapshot = Column(JSON, nullable=False)
+    user_id = Column(Text, nullable=False)
+    commit_message = Column(Text, nullable=True)
+    created_at = Column(BigInteger, nullable=False)
+
+
+class PromptHistoryModel(BaseModel):
+    id: str
+    prompt_id: str
+    parent_id: Optional[str] = None
+    snapshot: dict
+    user_id: str
+    commit_message: Optional[str] = None
+    created_at: int
+
+    model_config = ConfigDict(from_attributes=True)
+
+
+class PromptHistoryResponse(PromptHistoryModel):
+    """Response model with user info."""
+
+    user: Optional[UserResponse] = None
+
+
+class PromptHistoryTable:
+    def create_history_entry(
+        self,
+        prompt_id: str,
+        snapshot: dict,
+        user_id: str,
+        parent_id: Optional[str] = None,
+        commit_message: Optional[str] = None,
+        db: Optional[Session] = None,
+    ) -> Optional[PromptHistoryModel]:
+        """Create a new history entry (commit) for a prompt."""
+        with get_db_context(db) as db:
+            history = PromptHistory(
+                id=str(uuid.uuid4()),
+                prompt_id=prompt_id,
+                parent_id=parent_id,
+                snapshot=snapshot,
+                user_id=user_id,
+                commit_message=commit_message,
+                created_at=int(time.time()),
+            )
+            db.add(history)
+            db.commit()
+            db.refresh(history)
+            return PromptHistoryModel.model_validate(history)
+
+    def get_history_by_prompt_id(
+        self,
+        prompt_id: str,
+        limit: int = 50,
+        offset: int = 0,
+        db: Optional[Session] = None,
+    ) -> list[PromptHistoryResponse]:
+        """Get all history entries for a prompt, ordered by created_at desc."""
+        with get_db_context(db) as db:
+            entries = (
+                db.query(PromptHistory)
+                .filter(PromptHistory.prompt_id == prompt_id)
+                .order_by(PromptHistory.created_at.desc())
+                .offset(offset)
+                .limit(limit)
+                .all()
+            )
+
+            # Get user info for each entry
+            user_ids = list(set(e.user_id for e in entries))
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
+            users_dict = {user.id: user for user in users}
+
+            return [
+                PromptHistoryResponse(
+                    **PromptHistoryModel.model_validate(entry).model_dump(),
+                    user=(
+                        users_dict.get(entry.user_id).model_dump()
+                        if users_dict.get(entry.user_id)
+                        else None
+                    ),
+                )
+                for entry in entries
+            ]
+
+    def get_history_entry_by_id(
+        self,
+        history_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[PromptHistoryModel]:
+        """Get a specific history entry by ID."""
+        with get_db_context(db) as db:
+            entry = (
+                db.query(PromptHistory).filter(PromptHistory.id == history_id).first()
+            )
+            if entry:
+                return PromptHistoryModel.model_validate(entry)
+            return None
+
+    def get_latest_history_entry(
+        self,
+        prompt_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[PromptHistoryModel]:
+        """Get the most recent history entry for a prompt."""
+        with get_db_context(db) as db:
+            entry = (
+                db.query(PromptHistory)
+                .filter(PromptHistory.prompt_id == prompt_id)
+                .order_by(PromptHistory.created_at.desc())
+                .first()
+            )
+            if entry:
+                return PromptHistoryModel.model_validate(entry)
+            return None
+
+    def get_history_count(
+        self,
+        prompt_id: str,
+        db: Optional[Session] = None,
+    ) -> int:
+        """Get the number of history entries for a prompt."""
+        with get_db_context(db) as db:
+            return (
+                db.query(PromptHistory)
+                .filter(PromptHistory.prompt_id == prompt_id)
+                .count()
+            )
+
+    def compute_diff(
+        self,
+        from_id: str,
+        to_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[dict]:
+        """Compute diff between two history entries."""
+        with get_db_context(db) as db:
+            from_entry = (
+                db.query(PromptHistory).filter(PromptHistory.id == from_id).first()
+            )
+            to_entry = db.query(PromptHistory).filter(PromptHistory.id == to_id).first()
+
+            if not from_entry or not to_entry:
+                return None
+
+            from_snapshot = from_entry.snapshot
+            to_snapshot = to_entry.snapshot
+
+            # Compute diff for content field
+            from_content = from_snapshot.get("content", "")
+            to_content = to_snapshot.get("content", "")
+
+            diff_lines = list(
+                difflib.unified_diff(
+                    from_content.splitlines(keepends=True),
+                    to_content.splitlines(keepends=True),
+                    fromfile=f"v{from_id[:8]}",
+                    tofile=f"v{to_id[:8]}",
+                    lineterm="",
+                )
+            )
+
+            return {
+                "from_id": from_id,
+                "to_id": to_id,
+                "from_snapshot": from_snapshot,
+                "to_snapshot": to_snapshot,
+                "content_diff": diff_lines,
+                "name_changed": from_snapshot.get("name") != to_snapshot.get("name"),
+            }
+
+    def delete_history_by_prompt_id(
+        self,
+        prompt_id: str,
+        db: Optional[Session] = None,
+    ) -> bool:
+        """Delete all history entries for a prompt."""
+        with get_db_context(db) as db:
+            db.query(PromptHistory).filter(
+                PromptHistory.prompt_id == prompt_id
+            ).delete()
+            db.commit()
+            return True
+
+    def delete_history_entry(
+        self,
+        history_id: str,
+        db: Optional[Session] = None,
+    ) -> bool:
+        """Delete a history entry and reparent its children to grandparent."""
+        with get_db_context(db) as db:
+            entry = db.query(PromptHistory).filter_by(id=history_id).first()
+            if not entry:
+                return False
+
+            # Find children that reference this entry as parent
+            children = db.query(PromptHistory).filter_by(parent_id=history_id).all()
+
+            # Reparent children to grandparent
+            for child in children:
+                child.parent_id = entry.parent_id
+
+            db.delete(entry)
+            db.commit()
+            return True
+
+
+PromptHistories = PromptHistoryTable()

backend/open_webui/models/prompts.py

@@ -1,14 +1,17 @@
 import time
+import uuid
 from typing import Optional
 
-from open_webui.internal.db import Base, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 from open_webui.models.groups import Groups
 from open_webui.models.users import Users, UserResponse
+from open_webui.models.prompt_history import PromptHistories
+from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 
-from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Column, String, Text, JSON
 
-from open_webui.utils.access_control import has_access
+from pydantic import BaseModel, ConfigDict, Field
+from sqlalchemy import BigInteger, Boolean, Column, String, Text, JSON, or_, func, cast
 
 ####################
 # Prompts DB Schema
@@ -18,38 +21,35 @@ from open_webui.utils.access_control import has_access
 class Prompt(Base):
     __tablename__ = "prompt"
 
-    command = Column(String, primary_key=True)
+    id = Column(Text, primary_key=True)
+    command = Column(String, unique=True, index=True)
     user_id = Column(String)
-    title = Column(Text)
+    name = Column(Text)
     content = Column(Text)
-    timestamp = Column(BigInteger)
-
-    access_control = Column(JSON, nullable=True)  # Controls data access levels.
-    # Defines access control rules for this entry.
-    # - `None`: Public access, available to all users with the "user" role.
-    # - `{}`: Private access, restricted exclusively to the owner.
-    # - Custom permissions: Specific access control for reading and writing;
-    #   Can specify group or user-level restrictions:
-    #   {
-    #      "read": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      },
-    #      "write": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      }
-    #   }
+    data = Column(JSON, nullable=True)
+    meta = Column(JSON, nullable=True)
+    tags = Column(JSON, nullable=True)
+    is_active = Column(Boolean, default=True)
+    version_id = Column(Text, nullable=True)  # Points to active history entry
+    created_at = Column(BigInteger, nullable=True)
+    updated_at = Column(BigInteger, nullable=True)
 
 
 class PromptModel(BaseModel):
+    id: Optional[str] = None
     command: str
     user_id: str
-    title: str
+    name: str
     content: str
-    timestamp: int  # timestamp in epoch
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+    tags: Optional[list[str]] = None
+    is_active: Optional[bool] = True
+    version_id: Optional[str] = None
+    created_at: Optional[int] = None
+    updated_at: Optional[int] = None
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
 
-    access_control: Optional[dict] = None
     model_config = ConfigDict(from_attributes=True)
 
 
@@ -62,53 +62,152 @@ class PromptUserResponse(PromptModel):
     user: Optional[UserResponse] = None
 
 
+class PromptAccessResponse(PromptUserResponse):
+    write_access: Optional[bool] = False
+
+
+class PromptListResponse(BaseModel):
+    items: list[PromptUserResponse]
+    total: int
+
+
+class PromptAccessListResponse(BaseModel):
+    items: list[PromptAccessResponse]
+    total: int
+
+
 class PromptForm(BaseModel):
+
     command: str
-    title: str
+    name: str  # Changed from title
     content: str
-    access_control: Optional[dict] = None
+    data: Optional[dict] = None
+    meta: Optional[dict] = None
+    tags: Optional[list[str]] = None
+    access_grants: Optional[list[dict]] = None
+    version_id: Optional[str] = None  # Active version
+    commit_message: Optional[str] = None  # For history tracking
+    is_production: Optional[bool] = True  # Whether to set new version as production
 
 
 class PromptsTable:
+    def _get_access_grants(
+        self, prompt_id: str, db: Optional[Session] = None
+    ) -> list[AccessGrantModel]:
+        return AccessGrants.get_grants_by_resource("prompt", prompt_id, db=db)
+
+    def _to_prompt_model(
+        self, prompt: Prompt, db: Optional[Session] = None
+    ) -> PromptModel:
+        prompt_data = PromptModel.model_validate(prompt).model_dump(
+            exclude={"access_grants"}
+        )
+        prompt_data["access_grants"] = self._get_access_grants(prompt_data["id"], db=db)
+        return PromptModel.model_validate(prompt_data)
+
     def insert_new_prompt(
-        self, user_id: str, form_data: PromptForm
+        self, user_id: str, form_data: PromptForm, db: Optional[Session] = None
     ) -> Optional[PromptModel]:
+        now = int(time.time())
+        prompt_id = str(uuid.uuid4())
+
         prompt = PromptModel(
-            **{
-                "user_id": user_id,
-                **form_data.model_dump(),
-                "timestamp": int(time.time()),
-            }
+            id=prompt_id,
+            user_id=user_id,
+            command=form_data.command,
+            name=form_data.name,
+            content=form_data.content,
+            data=form_data.data or {},
+            meta=form_data.meta or {},
+            tags=form_data.tags or [],
+            access_grants=[],
+            is_active=True,
+            created_at=now,
+            updated_at=now,
         )
 
         try:
-            with get_db() as db:
-                result = Prompt(**prompt.model_dump())
+            with get_db_context(db) as db:
+                result = Prompt(**prompt.model_dump(exclude={"access_grants"}))
                 db.add(result)
                 db.commit()
                 db.refresh(result)
+                AccessGrants.set_access_grants(
+                    "prompt", prompt_id, form_data.access_grants, db=db
+                )
+
                 if result:
-                    return PromptModel.model_validate(result)
+                    current_access_grants = self._get_access_grants(prompt_id, db=db)
+                    snapshot = {
+                        "name": form_data.name,
+                        "content": form_data.content,
+                        "command": form_data.command,
+                        "data": form_data.data or {},
+                        "meta": form_data.meta or {},
+                        "tags": form_data.tags or [],
+                        "access_grants": [
+                            grant.model_dump() for grant in current_access_grants
+                        ],
+                    }
+
+                    history_entry = PromptHistories.create_history_entry(
+                        prompt_id=prompt_id,
+                        snapshot=snapshot,
+                        user_id=user_id,
+                        parent_id=None,  # Initial commit has no parent
+                        commit_message=form_data.commit_message or "Initial version",
+                        db=db,
+                    )
+
+                    # Set the initial version as the production version
+                    if history_entry:
+                        result.version_id = history_entry.id
+                        db.commit()
+                        db.refresh(result)
+
+                    return self._to_prompt_model(result, db=db)
                 else:
                     return None
         except Exception:
             return None
 
-    def get_prompt_by_command(self, command: str) -> Optional[PromptModel]:
+    def get_prompt_by_id(
+        self, prompt_id: str, db: Optional[Session] = None
+    ) -> Optional[PromptModel]:
+        """Get prompt by UUID."""
         try:
-            with get_db() as db:
-                prompt = db.query(Prompt).filter_by(command=command).first()
-                return PromptModel.model_validate(prompt)
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(id=prompt_id).first()
+                if prompt:
+                    return self._to_prompt_model(prompt, db=db)
+                return None
         except Exception:
             return None
 
-    def get_prompts(self) -> list[PromptUserResponse]:
-        with get_db() as db:
-            all_prompts = db.query(Prompt).order_by(Prompt.timestamp.desc()).all()
+    def get_prompt_by_command(
+        self, command: str, db: Optional[Session] = None
+    ) -> Optional[PromptModel]:
+        try:
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(command=command).first()
+                if prompt:
+                    return self._to_prompt_model(prompt, db=db)
+                return None
+        except Exception:
+            return None
+
+    def get_prompts(self, db: Optional[Session] = None) -> list[PromptUserResponse]:
+        with get_db_context(db) as db:
+            all_prompts = (
+                db.query(Prompt)
+                .filter(Prompt.is_active == True)
+                .order_by(Prompt.updated_at.desc())
+                .all()
+            )
 
             user_ids = list(set(prompt.user_id for prompt in all_prompts))
 
-            users = Users.get_users_by_user_ids(user_ids) if user_ids else []
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
             users_dict = {user.id: user for user in users}
 
             prompts = []
@@ -117,7 +216,7 @@ class PromptsTable:
                 prompts.append(
                     PromptUserResponse.model_validate(
                         {
-                            **PromptModel.model_validate(prompt).model_dump(),
+                            **self._to_prompt_model(prompt, db=db).model_dump(),
                             "user": user.model_dump() if user else None,
                         }
                     )
@@ -126,42 +225,411 @@ class PromptsTable:
             return prompts
 
     def get_prompts_by_user_id(
-        self, user_id: str, permission: str = "write"
+        self, user_id: str, permission: str = "write", db: Optional[Session] = None
     ) -> list[PromptUserResponse]:
-        prompts = self.get_prompts()
-        user_group_ids = {group.id for group in Groups.get_groups_by_member_id(user_id)}
+        prompts = self.get_prompts(db=db)
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
 
         return [
             prompt
             for prompt in prompts
             if prompt.user_id == user_id
-            or has_access(user_id, permission, prompt.access_control, user_group_ids)
+            or AccessGrants.has_access(
+                user_id=user_id,
+                resource_type="prompt",
+                resource_id=prompt.id,
+                permission=permission,
+                user_group_ids=user_group_ids,
+                db=db,
+            )
         ]
 
+    def search_prompts(
+        self,
+        user_id: str,
+        filter: dict = {},
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> PromptListResponse:
+        with get_db_context(db) as db:
+            from open_webui.models.users import User, UserModel
+
+            # Join with User table for user filtering and sorting
+            query = db.query(Prompt, User).outerjoin(User, User.id == Prompt.user_id)
+            query = query.filter(Prompt.is_active == True)
+
+            if filter:
+                query_key = filter.get("query")
+                if query_key:
+                    query = query.filter(
+                        or_(
+                            Prompt.name.ilike(f"%{query_key}%"),
+                            Prompt.command.ilike(f"%{query_key}%"),
+                            Prompt.content.ilike(f"%{query_key}%"),
+                            User.name.ilike(f"%{query_key}%"),
+                            User.email.ilike(f"%{query_key}%"),
+                        )
+                    )
+
+                view_option = filter.get("view_option")
+                if view_option == "created":
+                    query = query.filter(Prompt.user_id == user_id)
+                elif view_option == "shared":
+                    query = query.filter(Prompt.user_id != user_id)
+
+                # Apply access grant filtering
+                query = AccessGrants.has_permission_filter(
+                    db=db,
+                    query=query,
+                    DocumentModel=Prompt,
+                    filter=filter,
+                    resource_type="prompt",
+                    permission="read",
+                )
+
+                tag = filter.get("tag")
+                if tag:
+                    # Search for tag in JSON array field
+                    like_pattern = f'%"{tag.lower()}"%'
+                    tags_text = func.lower(cast(Prompt.tags, String))
+                    query = query.filter(tags_text.like(like_pattern))
+
+                order_by = filter.get("order_by")
+                direction = filter.get("direction")
+
+                if order_by == "name":
+                    if direction == "asc":
+                        query = query.order_by(Prompt.name.asc())
+                    else:
+                        query = query.order_by(Prompt.name.desc())
+                elif order_by == "created_at":
+                    if direction == "asc":
+                        query = query.order_by(Prompt.created_at.asc())
+                    else:
+                        query = query.order_by(Prompt.created_at.desc())
+                elif order_by == "updated_at":
+                    if direction == "asc":
+                        query = query.order_by(Prompt.updated_at.asc())
+                    else:
+                        query = query.order_by(Prompt.updated_at.desc())
+                else:
+                    query = query.order_by(Prompt.updated_at.desc())
+            else:
+                query = query.order_by(Prompt.updated_at.desc())
+
+            # Count BEFORE pagination
+            total = query.count()
+
+            if skip:
+                query = query.offset(skip)
+            if limit:
+                query = query.limit(limit)
+
+            items = query.all()
+
+            prompts = []
+            for prompt, user in items:
+                prompts.append(
+                    PromptUserResponse(
+                        **self._to_prompt_model(prompt, db=db).model_dump(),
+                        user=(
+                            UserResponse(**UserModel.model_validate(user).model_dump())
+                            if user
+                            else None
+                        ),
+                    )
+                )
+
+            return PromptListResponse(items=prompts, total=total)
+
     def update_prompt_by_command(
-        self, command: str, form_data: PromptForm
+        self,
+        command: str,
+        form_data: PromptForm,
+        user_id: str,
+        db: Optional[Session] = None,
     ) -> Optional[PromptModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 prompt = db.query(Prompt).filter_by(command=command).first()
-                prompt.title = form_data.title
+                if not prompt:
+                    return None
+
+                latest_history = PromptHistories.get_latest_history_entry(
+                    prompt.id, db=db
+                )
+                parent_id = latest_history.id if latest_history else None
+                current_access_grants = self._get_access_grants(prompt.id, db=db)
+
+                # Check if content changed to decide on history creation
+                content_changed = (
+                    prompt.name != form_data.name
+                    or prompt.content != form_data.content
+                    or form_data.access_grants is not None
+                )
+
+                # Update prompt fields
+                prompt.name = form_data.name
                 prompt.content = form_data.content
-                prompt.access_control = form_data.access_control
-                prompt.timestamp = int(time.time())
+                prompt.data = form_data.data or prompt.data
+                prompt.meta = form_data.meta or prompt.meta
+                prompt.updated_at = int(time.time())
+                if form_data.access_grants is not None:
+                    AccessGrants.set_access_grants(
+                        "prompt", prompt.id, form_data.access_grants, db=db
+                    )
+                    current_access_grants = self._get_access_grants(prompt.id, db=db)
+
                 db.commit()
-                return PromptModel.model_validate(prompt)
+
+                # Create history entry only if content changed
+                if content_changed:
+                    snapshot = {
+                        "name": form_data.name,
+                        "content": form_data.content,
+                        "command": command,
+                        "data": form_data.data or {},
+                        "meta": form_data.meta or {},
+                        "access_grants": [
+                            grant.model_dump() for grant in current_access_grants
+                        ],
+                    }
+
+                    history_entry = PromptHistories.create_history_entry(
+                        prompt_id=prompt.id,
+                        snapshot=snapshot,
+                        user_id=user_id,
+                        parent_id=parent_id,
+                        commit_message=form_data.commit_message,
+                        db=db,
+                    )
+
+                    # Set as production if flag is True (default)
+                    if form_data.is_production and history_entry:
+                        prompt.version_id = history_entry.id
+                        db.commit()
+
+                return self._to_prompt_model(prompt, db=db)
         except Exception:
             return None
 
-    def delete_prompt_by_command(self, command: str) -> bool:
+    def update_prompt_by_id(
+        self,
+        prompt_id: str,
+        form_data: PromptForm,
+        user_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[PromptModel]:
         try:
-            with get_db() as db:
-                db.query(Prompt).filter_by(command=command).delete()
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(id=prompt_id).first()
+                if not prompt:
+                    return None
+
+                latest_history = PromptHistories.get_latest_history_entry(
+                    prompt.id, db=db
+                )
+                parent_id = latest_history.id if latest_history else None
+                current_access_grants = self._get_access_grants(prompt.id, db=db)
+
+                # Check if content changed to decide on history creation
+                content_changed = (
+                    prompt.name != form_data.name
+                    or prompt.command != form_data.command
+                    or prompt.content != form_data.content
+                    or form_data.access_grants is not None
+                    or (form_data.tags is not None and prompt.tags != form_data.tags)
+                )
+
+                # Update prompt fields
+                prompt.name = form_data.name
+                prompt.command = form_data.command
+                prompt.content = form_data.content
+                prompt.data = form_data.data or prompt.data
+                prompt.meta = form_data.meta or prompt.meta
+
+                if form_data.tags is not None:
+                    prompt.tags = form_data.tags
+
+                if form_data.access_grants is not None:
+                    AccessGrants.set_access_grants(
+                        "prompt", prompt.id, form_data.access_grants, db=db
+                    )
+                    current_access_grants = self._get_access_grants(prompt.id, db=db)
+
+                prompt.updated_at = int(time.time())
+
                 db.commit()
 
-                return True
+                # Create history entry only if content changed
+                if content_changed:
+                    snapshot = {
+                        "name": form_data.name,
+                        "content": form_data.content,
+                        "command": prompt.command,
+                        "data": form_data.data or {},
+                        "meta": form_data.meta or {},
+                        "tags": prompt.tags or [],
+                        "access_grants": [
+                            grant.model_dump() for grant in current_access_grants
+                        ],
+                    }
+
+                    history_entry = PromptHistories.create_history_entry(
+                        prompt_id=prompt.id,
+                        snapshot=snapshot,
+                        user_id=user_id,
+                        parent_id=parent_id,
+                        commit_message=form_data.commit_message,
+                        db=db,
+                    )
+
+                    # Set as production if flag is True (default)
+                    if form_data.is_production and history_entry:
+                        prompt.version_id = history_entry.id
+                        db.commit()
+
+                return self._to_prompt_model(prompt, db=db)
+        except Exception:
+            return None
+
+    def update_prompt_metadata(
+        self,
+        prompt_id: str,
+        name: str,
+        command: str,
+        tags: Optional[list[str]] = None,
+        db: Optional[Session] = None,
+    ) -> Optional[PromptModel]:
+        """Update only name, command, and tags (no history created)."""
+        try:
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(id=prompt_id).first()
+                if not prompt:
+                    return None
+
+                prompt.name = name
+                prompt.command = command
+
+                if tags is not None:
+                    prompt.tags = tags
+
+                prompt.updated_at = int(time.time())
+                db.commit()
+
+                return self._to_prompt_model(prompt, db=db)
+        except Exception:
+            return None
+
+    def update_prompt_version(
+        self,
+        prompt_id: str,
+        version_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[PromptModel]:
+        """Set the active version of a prompt and restore content from that version's snapshot."""
+        try:
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(id=prompt_id).first()
+                if not prompt:
+                    return None
+
+                history_entry = PromptHistories.get_history_entry_by_id(
+                    version_id, db=db
+                )
+
+                if not history_entry:
+                    return None
+
+                # Restore prompt content from the snapshot
+                snapshot = history_entry.snapshot
+                if snapshot:
+                    prompt.name = snapshot.get("name", prompt.name)
+                    prompt.content = snapshot.get("content", prompt.content)
+                    prompt.data = snapshot.get("data", prompt.data)
+                    prompt.meta = snapshot.get("meta", prompt.meta)
+                    prompt.tags = snapshot.get("tags", prompt.tags)
+                    # Note: command and access_grants are not restored from snapshot
+
+                prompt.version_id = version_id
+                prompt.updated_at = int(time.time())
+                db.commit()
+
+                return self._to_prompt_model(prompt, db=db)
+        except Exception:
+            return None
+
+    def delete_prompt_by_command(
+        self, command: str, db: Optional[Session] = None
+    ) -> bool:
+        """Soft delete a prompt by setting is_active to False."""
+        try:
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(command=command).first()
+                if prompt:
+                    PromptHistories.delete_history_by_prompt_id(prompt.id, db=db)
+                    AccessGrants.revoke_all_access("prompt", prompt.id, db=db)
+
+                    prompt.is_active = False
+                    prompt.updated_at = int(time.time())
+                    db.commit()
+                    return True
+                return False
         except Exception:
             return False
 
+    def delete_prompt_by_id(self, prompt_id: str, db: Optional[Session] = None) -> bool:
+        """Soft delete a prompt by setting is_active to False."""
+        try:
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(id=prompt_id).first()
+                if prompt:
+                    PromptHistories.delete_history_by_prompt_id(prompt.id, db=db)
+                    AccessGrants.revoke_all_access("prompt", prompt.id, db=db)
+
+                    prompt.is_active = False
+                    prompt.updated_at = int(time.time())
+                    db.commit()
+                    return True
+                return False
+        except Exception:
+            return False
+
+    def hard_delete_prompt_by_command(
+        self, command: str, db: Optional[Session] = None
+    ) -> bool:
+        """Permanently delete a prompt and its history."""
+        try:
+            with get_db_context(db) as db:
+                prompt = db.query(Prompt).filter_by(command=command).first()
+                if prompt:
+                    PromptHistories.delete_history_by_prompt_id(prompt.id, db=db)
+                    AccessGrants.revoke_all_access("prompt", prompt.id, db=db)
+
+                    # Delete prompt
+                    db.query(Prompt).filter_by(command=command).delete()
+                    db.commit()
+                    return True
+                return False
+        except Exception:
+            return False
+
+    def get_tags(self, db: Optional[Session] = None) -> list[str]:
+        try:
+            with get_db_context(db) as db:
+                prompts = db.query(Prompt).filter_by(is_active=True).all()
+                tags = set()
+                for prompt in prompts:
+                    if prompt.tags:
+                        for tag in prompt.tags:
+                            if tag:
+                                tags.add(tag)
+                return sorted(list(tags))
+        except Exception:
+            return []
+
 
 Prompts = PromptsTable()

backend/open_webui/models/skills.py

@@ -0,0 +1,339 @@
+import logging
+import time
+from typing import Optional
+
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, get_db, get_db_context
+from open_webui.models.users import Users, UserResponse
+from open_webui.models.groups import Groups
+from open_webui.models.access_grants import AccessGrantModel, AccessGrants
+
+from pydantic import BaseModel, ConfigDict, Field
+from sqlalchemy import JSON, BigInteger, Boolean, Column, String, Text, or_
+
+log = logging.getLogger(__name__)
+
+####################
+# Skills DB Schema
+####################
+
+
+class Skill(Base):
+    __tablename__ = "skill"
+
+    id = Column(String, primary_key=True, unique=True)
+    user_id = Column(String)
+    name = Column(Text, unique=True)
+    description = Column(Text, nullable=True)
+    content = Column(Text)
+    meta = Column(JSON)
+    is_active = Column(Boolean, default=True)
+
+    updated_at = Column(BigInteger)
+    created_at = Column(BigInteger)
+
+
+class SkillMeta(BaseModel):
+    tags: Optional[list[str]] = []
+
+
+class SkillModel(BaseModel):
+    id: str
+    user_id: str
+    name: str
+    description: Optional[str] = None
+    content: str
+    meta: SkillMeta
+    is_active: bool = True
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
+
+    updated_at: int  # timestamp in epoch
+    created_at: int  # timestamp in epoch
+
+    model_config = ConfigDict(from_attributes=True)
+
+
+####################
+# Forms
+####################
+
+
+class SkillUserModel(SkillModel):
+    user: Optional[UserResponse] = None
+
+
+class SkillResponse(BaseModel):
+    id: str
+    user_id: str
+    name: str
+    description: Optional[str] = None
+    meta: SkillMeta
+    is_active: bool = True
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
+    updated_at: int  # timestamp in epoch
+    created_at: int  # timestamp in epoch
+
+
+class SkillUserResponse(SkillResponse):
+    user: Optional[UserResponse] = None
+
+    model_config = ConfigDict(extra="allow")
+
+
+class SkillAccessResponse(SkillUserResponse):
+    write_access: Optional[bool] = False
+
+
+class SkillForm(BaseModel):
+    id: str
+    name: str
+    description: Optional[str] = None
+    content: str
+    meta: SkillMeta = SkillMeta()
+    is_active: bool = True
+    access_grants: Optional[list[dict]] = None
+
+
+class SkillListResponse(BaseModel):
+    items: list[SkillUserResponse] = []
+    total: int = 0
+
+
+class SkillAccessListResponse(BaseModel):
+    items: list[SkillAccessResponse] = []
+    total: int = 0
+
+
+class SkillsTable:
+    def _get_access_grants(
+        self, skill_id: str, db: Optional[Session] = None
+    ) -> list[AccessGrantModel]:
+        return AccessGrants.get_grants_by_resource("skill", skill_id, db=db)
+
+    def _to_skill_model(self, skill: Skill, db: Optional[Session] = None) -> SkillModel:
+        skill_data = SkillModel.model_validate(skill).model_dump(
+            exclude={"access_grants"}
+        )
+        skill_data["access_grants"] = self._get_access_grants(skill_data["id"], db=db)
+        return SkillModel.model_validate(skill_data)
+
+    def insert_new_skill(
+        self,
+        user_id: str,
+        form_data: SkillForm,
+        db: Optional[Session] = None,
+    ) -> Optional[SkillModel]:
+        with get_db_context(db) as db:
+            try:
+                result = Skill(
+                    **{
+                        **form_data.model_dump(exclude={"access_grants"}),
+                        "user_id": user_id,
+                        "updated_at": int(time.time()),
+                        "created_at": int(time.time()),
+                    }
+                )
+                db.add(result)
+                db.commit()
+                db.refresh(result)
+                AccessGrants.set_access_grants(
+                    "skill", result.id, form_data.access_grants, db=db
+                )
+                if result:
+                    return self._to_skill_model(result, db=db)
+                else:
+                    return None
+            except Exception as e:
+                log.exception(f"Error creating a new skill: {e}")
+                return None
+
+    def get_skill_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[SkillModel]:
+        try:
+            with get_db_context(db) as db:
+                skill = db.get(Skill, id)
+                return self._to_skill_model(skill, db=db) if skill else None
+        except Exception:
+            return None
+
+    def get_skill_by_name(
+        self, name: str, db: Optional[Session] = None
+    ) -> Optional[SkillModel]:
+        try:
+            with get_db_context(db) as db:
+                skill = db.query(Skill).filter_by(name=name).first()
+                return self._to_skill_model(skill, db=db) if skill else None
+        except Exception:
+            return None
+
+    def get_skills(self, db: Optional[Session] = None) -> list[SkillUserModel]:
+        with get_db_context(db) as db:
+            all_skills = db.query(Skill).order_by(Skill.updated_at.desc()).all()
+
+            user_ids = list(set(skill.user_id for skill in all_skills))
+
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
+            users_dict = {user.id: user for user in users}
+
+            skills = []
+            for skill in all_skills:
+                user = users_dict.get(skill.user_id)
+                skills.append(
+                    SkillUserModel.model_validate(
+                        {
+                            **self._to_skill_model(skill, db=db).model_dump(),
+                            "user": user.model_dump() if user else None,
+                        }
+                    )
+                )
+            return skills
+
+    def get_skills_by_user_id(
+        self, user_id: str, permission: str = "write", db: Optional[Session] = None
+    ) -> list[SkillUserModel]:
+        skills = self.get_skills(db=db)
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
+
+        return [
+            skill
+            for skill in skills
+            if skill.user_id == user_id
+            or AccessGrants.has_access(
+                user_id=user_id,
+                resource_type="skill",
+                resource_id=skill.id,
+                permission=permission,
+                user_group_ids=user_group_ids,
+                db=db,
+            )
+        ]
+
+    def search_skills(
+        self,
+        user_id: str,
+        filter: dict = {},
+        skip: int = 0,
+        limit: int = 30,
+        db: Optional[Session] = None,
+    ) -> SkillListResponse:
+        try:
+            with get_db_context(db) as db:
+                from open_webui.models.users import User, UserModel
+
+                # Join with User table for user filtering
+                query = db.query(Skill, User).outerjoin(User, User.id == Skill.user_id)
+
+                if filter:
+                    query_key = filter.get("query")
+                    if query_key:
+                        query = query.filter(
+                            or_(
+                                Skill.name.ilike(f"%{query_key}%"),
+                                Skill.description.ilike(f"%{query_key}%"),
+                                Skill.id.ilike(f"%{query_key}%"),
+                                User.name.ilike(f"%{query_key}%"),
+                                User.email.ilike(f"%{query_key}%"),
+                            )
+                        )
+
+                    view_option = filter.get("view_option")
+                    if view_option == "created":
+                        query = query.filter(Skill.user_id == user_id)
+                    elif view_option == "shared":
+                        query = query.filter(Skill.user_id != user_id)
+
+                    # Apply access grant filtering
+                    query = AccessGrants.has_permission_filter(
+                        db=db,
+                        query=query,
+                        DocumentModel=Skill,
+                        filter=filter,
+                        resource_type="skill",
+                        permission="read",
+                    )
+
+                query = query.order_by(Skill.updated_at.desc())
+
+                # Count BEFORE pagination
+                total = query.count()
+
+                if skip:
+                    query = query.offset(skip)
+                if limit:
+                    query = query.limit(limit)
+
+                items = query.all()
+
+                skills = []
+                for skill, user in items:
+                    skills.append(
+                        SkillUserResponse(
+                            **self._to_skill_model(skill, db=db).model_dump(),
+                            user=(
+                                UserResponse(
+                                    **UserModel.model_validate(user).model_dump()
+                                )
+                                if user
+                                else None
+                            ),
+                        )
+                    )
+
+                return SkillListResponse(items=skills, total=total)
+        except Exception as e:
+            log.exception(f"Error searching skills: {e}")
+            return SkillListResponse(items=[], total=0)
+
+    def update_skill_by_id(
+        self, id: str, updated: dict, db: Optional[Session] = None
+    ) -> Optional[SkillModel]:
+        try:
+            with get_db_context(db) as db:
+                access_grants = updated.pop("access_grants", None)
+                db.query(Skill).filter_by(id=id).update(
+                    {**updated, "updated_at": int(time.time())}
+                )
+                db.commit()
+                if access_grants is not None:
+                    AccessGrants.set_access_grants("skill", id, access_grants, db=db)
+
+                skill = db.query(Skill).get(id)
+                db.refresh(skill)
+                return self._to_skill_model(skill, db=db)
+        except Exception:
+            return None
+
+    def toggle_skill_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[SkillModel]:
+        with get_db_context(db) as db:
+            try:
+                skill = db.query(Skill).filter_by(id=id).first()
+                if not skill:
+                    return None
+
+                skill.is_active = not skill.is_active
+                skill.updated_at = int(time.time())
+                db.commit()
+                db.refresh(skill)
+
+                return self._to_skill_model(skill, db=db)
+            except Exception:
+                return None
+
+    def delete_skill_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        try:
+            with get_db_context(db) as db:
+                AccessGrants.revoke_all_access("skill", id, db=db)
+                db.query(Skill).filter_by(id=id).delete()
+                db.commit()
+
+                return True
+        except Exception:
+            return False
+
+
+Skills = SkillsTable()

backend/open_webui/models/tags.py

@@ -3,15 +3,14 @@ import time
 import uuid
 from typing import Optional
 
-from open_webui.internal.db import Base, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 
 
-from open_webui.env import SRC_LOG_LEVELS
 from pydantic import BaseModel, ConfigDict
 from sqlalchemy import BigInteger, Column, String, JSON, PrimaryKeyConstraint, Index
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 
 ####################
@@ -52,8 +51,10 @@ class TagChatIdForm(BaseModel):
 
 
 class TagTable:
-    def insert_new_tag(self, name: str, user_id: str) -> Optional[TagModel]:
-        with get_db() as db:
+    def insert_new_tag(
+        self, name: str, user_id: str, db: Optional[Session] = None
+    ) -> Optional[TagModel]:
+        with get_db_context(db) as db:
             id = name.replace(" ", "_").lower()
             tag = TagModel(**{"id": id, "user_id": user_id, "name": name})
             try:
@@ -70,27 +71,29 @@ class TagTable:
                 return None
 
     def get_tag_by_name_and_user_id(
-        self, name: str, user_id: str
+        self, name: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[TagModel]:
         try:
             id = name.replace(" ", "_").lower()
-            with get_db() as db:
+            with get_db_context(db) as db:
                 tag = db.query(Tag).filter_by(id=id, user_id=user_id).first()
                 return TagModel.model_validate(tag)
         except Exception:
             return None
 
-    def get_tags_by_user_id(self, user_id: str) -> list[TagModel]:
-        with get_db() as db:
+    def get_tags_by_user_id(
+        self, user_id: str, db: Optional[Session] = None
+    ) -> list[TagModel]:
+        with get_db_context(db) as db:
             return [
                 TagModel.model_validate(tag)
                 for tag in (db.query(Tag).filter_by(user_id=user_id).all())
             ]
 
     def get_tags_by_ids_and_user_id(
-        self, ids: list[str], user_id: str
+        self, ids: list[str], user_id: str, db: Optional[Session] = None
     ) -> list[TagModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             return [
                 TagModel.model_validate(tag)
                 for tag in (
@@ -98,9 +101,11 @@ class TagTable:
                 )
             ]
 
-    def delete_tag_by_name_and_user_id(self, name: str, user_id: str) -> bool:
+    def delete_tag_by_name_and_user_id(
+        self, name: str, user_id: str, db: Optional[Session] = None
+    ) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 id = name.replace(" ", "_").lower()
                 res = db.query(Tag).filter_by(id=id, user_id=user_id).delete()
                 log.debug(f"res: {res}")
@@ -110,5 +115,45 @@ class TagTable:
             log.error(f"delete_tag: {e}")
             return False
 
+    def delete_tags_by_ids_and_user_id(
+        self, ids: list[str], user_id: str, db: Optional[Session] = None
+    ) -> bool:
+        """Delete all tags whose id is in *ids* for the given user, in one query."""
+        if not ids:
+            return True
+        try:
+            with get_db_context(db) as db:
+                db.query(Tag).filter(Tag.id.in_(ids), Tag.user_id == user_id).delete(
+                    synchronize_session=False
+                )
+                db.commit()
+                return True
+        except Exception as e:
+            log.error(f"delete_tags_by_ids: {e}")
+            return False
+
+    def ensure_tags_exist(
+        self, names: list[str], user_id: str, db: Optional[Session] = None
+    ) -> None:
+        """Create tag rows for any *names* that don't already exist for *user_id*."""
+        if not names:
+            return
+        ids = [n.replace(" ", "_").lower() for n in names]
+        with get_db_context(db) as db:
+            existing = {
+                t.id
+                for t in db.query(Tag.id)
+                .filter(Tag.id.in_(ids), Tag.user_id == user_id)
+                .all()
+            }
+            new_tags = [
+                Tag(id=tag_id, name=name, user_id=user_id)
+                for tag_id, name in zip(ids, names)
+                if tag_id not in existing
+            ]
+            if new_tags:
+                db.add_all(new_tags)
+                db.commit()
+
 
 Tags = TagTable()

backend/open_webui/models/tools.py

@@ -2,19 +2,16 @@ import logging
 import time
 from typing import Optional
 
-from open_webui.internal.db import Base, JSONField, get_db
+from sqlalchemy.orm import Session
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 from open_webui.models.users import Users, UserResponse
 from open_webui.models.groups import Groups
+from open_webui.models.access_grants import AccessGrantModel, AccessGrants
 
-from open_webui.env import SRC_LOG_LEVELS
-from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Column, String, Text, JSON
-
-from open_webui.utils.access_control import has_access
-
+from pydantic import BaseModel, ConfigDict, Field
+from sqlalchemy import BigInteger, Column, String, Text
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
 
 ####################
 # Tools DB Schema
@@ -24,7 +21,7 @@ log.setLevel(SRC_LOG_LEVELS["MODELS"])
 class Tool(Base):
     __tablename__ = "tool"
 
-    id = Column(String, primary_key=True)
+    id = Column(String, primary_key=True, unique=True)
     user_id = Column(String)
     name = Column(Text)
     content = Column(Text)
@@ -32,23 +29,6 @@ class Tool(Base):
     meta = Column(JSONField)
     valves = Column(JSONField)
 
-    access_control = Column(JSON, nullable=True)  # Controls data access levels.
-    # Defines access control rules for this entry.
-    # - `None`: Public access, available to all users with the "user" role.
-    # - `{}`: Private access, restricted exclusively to the owner.
-    # - Custom permissions: Specific access control for reading and writing;
-    #   Can specify group or user-level restrictions:
-    #   {
-    #      "read": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      },
-    #      "write": {
-    #          "group_ids": ["group_id1", "group_id2"],
-    #          "user_ids":  ["user_id1", "user_id2"]
-    #      }
-    #   }
-
     updated_at = Column(BigInteger)
     created_at = Column(BigInteger)
 
@@ -65,7 +45,7 @@ class ToolModel(BaseModel):
     content: str
     specs: list[dict]
     meta: ToolMeta
-    access_control: Optional[dict] = None
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
 
     updated_at: int  # timestamp in epoch
     created_at: int  # timestamp in epoch
@@ -87,7 +67,7 @@ class ToolResponse(BaseModel):
     user_id: str
     name: str
     meta: ToolMeta
-    access_control: Optional[dict] = None
+    access_grants: list[AccessGrantModel] = Field(default_factory=list)
     updated_at: int  # timestamp in epoch
     created_at: int  # timestamp in epoch
 
@@ -98,12 +78,16 @@ class ToolUserResponse(ToolResponse):
     model_config = ConfigDict(extra="allow")
 
 
+class ToolAccessResponse(ToolUserResponse):
+    write_access: Optional[bool] = False
+
+
 class ToolForm(BaseModel):
     id: str
     name: str
     content: str
     meta: ToolMeta
-    access_control: Optional[dict] = None
+    access_grants: Optional[list[dict]] = None
 
 
 class ToolValves(BaseModel):
@@ -111,48 +95,65 @@ class ToolValves(BaseModel):
 
 
 class ToolsTable:
-    def insert_new_tool(
-        self, user_id: str, form_data: ToolForm, specs: list[dict]
-    ) -> Optional[ToolModel]:
-        with get_db() as db:
-            tool = ToolModel(
-                **{
-                    **form_data.model_dump(),
-                    "specs": specs,
-                    "user_id": user_id,
-                    "updated_at": int(time.time()),
-                    "created_at": int(time.time()),
-                }
-            )
+    def _get_access_grants(
+        self, tool_id: str, db: Optional[Session] = None
+    ) -> list[AccessGrantModel]:
+        return AccessGrants.get_grants_by_resource("tool", tool_id, db=db)
 
+    def _to_tool_model(self, tool: Tool, db: Optional[Session] = None) -> ToolModel:
+        tool_data = ToolModel.model_validate(tool).model_dump(exclude={"access_grants"})
+        tool_data["access_grants"] = self._get_access_grants(tool_data["id"], db=db)
+        return ToolModel.model_validate(tool_data)
+
+    def insert_new_tool(
+        self,
+        user_id: str,
+        form_data: ToolForm,
+        specs: list[dict],
+        db: Optional[Session] = None,
+    ) -> Optional[ToolModel]:
+        with get_db_context(db) as db:
             try:
-                result = Tool(**tool.model_dump())
+                result = Tool(
+                    **{
+                        **form_data.model_dump(exclude={"access_grants"}),
+                        "specs": specs,
+                        "user_id": user_id,
+                        "updated_at": int(time.time()),
+                        "created_at": int(time.time()),
+                    }
+                )
                 db.add(result)
                 db.commit()
                 db.refresh(result)
+                AccessGrants.set_access_grants(
+                    "tool", result.id, form_data.access_grants, db=db
+                )
                 if result:
-                    return ToolModel.model_validate(result)
+                    return self._to_tool_model(result, db=db)
                 else:
                     return None
             except Exception as e:
                 log.exception(f"Error creating a new tool: {e}")
                 return None
 
-    def get_tool_by_id(self, id: str) -> Optional[ToolModel]:
+    def get_tool_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[ToolModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 tool = db.get(Tool, id)
-                return ToolModel.model_validate(tool)
+                return self._to_tool_model(tool, db=db) if tool else None
         except Exception:
             return None
 
-    def get_tools(self) -> list[ToolUserModel]:
-        with get_db() as db:
+    def get_tools(self, db: Optional[Session] = None) -> list[ToolUserModel]:
+        with get_db_context(db) as db:
             all_tools = db.query(Tool).order_by(Tool.updated_at.desc()).all()
 
             user_ids = list(set(tool.user_id for tool in all_tools))
 
-            users = Users.get_users_by_user_ids(user_ids) if user_ids else []
+            users = Users.get_users_by_user_ids(user_ids, db=db) if user_ids else []
             users_dict = {user.id: user for user in users}
 
             tools = []
@@ -161,7 +162,7 @@ class ToolsTable:
                 tools.append(
                     ToolUserModel.model_validate(
                         {
-                            **ToolModel.model_validate(tool).model_dump(),
+                            **self._to_tool_model(tool, db=db).model_dump(),
                             "user": user.model_dump() if user else None,
                         }
                     )
@@ -169,43 +170,56 @@ class ToolsTable:
             return tools
 
     def get_tools_by_user_id(
-        self, user_id: str, permission: str = "write"
+        self, user_id: str, permission: str = "write", db: Optional[Session] = None
     ) -> list[ToolUserModel]:
-        tools = self.get_tools()
-        user_group_ids = {group.id for group in Groups.get_groups_by_member_id(user_id)}
+        tools = self.get_tools(db=db)
+        user_group_ids = {
+            group.id for group in Groups.get_groups_by_member_id(user_id, db=db)
+        }
 
         return [
             tool
             for tool in tools
             if tool.user_id == user_id
-            or has_access(user_id, permission, tool.access_control, user_group_ids)
+            or AccessGrants.has_access(
+                user_id=user_id,
+                resource_type="tool",
+                resource_id=tool.id,
+                permission=permission,
+                user_group_ids=user_group_ids,
+                db=db,
+            )
         ]
 
-    def get_tool_valves_by_id(self, id: str) -> Optional[dict]:
+    def get_tool_valves_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[dict]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 tool = db.get(Tool, id)
                 return tool.valves if tool.valves else {}
         except Exception as e:
             log.exception(f"Error getting tool valves by id {id}")
             return None
 
-    def update_tool_valves_by_id(self, id: str, valves: dict) -> Optional[ToolValves]:
+    def update_tool_valves_by_id(
+        self, id: str, valves: dict, db: Optional[Session] = None
+    ) -> Optional[ToolValves]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 db.query(Tool).filter_by(id=id).update(
                     {"valves": valves, "updated_at": int(time.time())}
                 )
                 db.commit()
-                return self.get_tool_by_id(id)
+                return self.get_tool_by_id(id, db=db)
         except Exception:
             return None
 
     def get_user_valves_by_id_and_user_id(
-        self, id: str, user_id: str
+        self, id: str, user_id: str, db: Optional[Session] = None
     ) -> Optional[dict]:
         try:
-            user = Users.get_user_by_id(user_id)
+            user = Users.get_user_by_id(user_id, db=db)
             user_settings = user.settings.model_dump() if user.settings else {}
 
             # Check if user has "tools" and "valves" settings
@@ -222,10 +236,10 @@ class ToolsTable:
             return None
 
     def update_user_valves_by_id_and_user_id(
-        self, id: str, user_id: str, valves: dict
+        self, id: str, user_id: str, valves: dict, db: Optional[Session] = None
     ) -> Optional[dict]:
         try:
-            user = Users.get_user_by_id(user_id)
+            user = Users.get_user_by_id(user_id, db=db)
             user_settings = user.settings.model_dump() if user.settings else {}
 
             # Check if user has "tools" and "valves" settings
@@ -237,7 +251,7 @@ class ToolsTable:
             user_settings["tools"]["valves"][id] = valves
 
             # Update the user settings in the database
-            Users.update_user_by_id(user_id, {"settings": user_settings})
+            Users.update_user_by_id(user_id, {"settings": user_settings}, db=db)
 
             return user_settings["tools"]["valves"][id]
         except Exception as e:
@@ -246,23 +260,29 @@ class ToolsTable:
             )
             return None
 
-    def update_tool_by_id(self, id: str, updated: dict) -> Optional[ToolModel]:
+    def update_tool_by_id(
+        self, id: str, updated: dict, db: Optional[Session] = None
+    ) -> Optional[ToolModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
+                access_grants = updated.pop("access_grants", None)
                 db.query(Tool).filter_by(id=id).update(
                     {**updated, "updated_at": int(time.time())}
                 )
                 db.commit()
+                if access_grants is not None:
+                    AccessGrants.set_access_grants("tool", id, access_grants, db=db)
 
                 tool = db.query(Tool).get(id)
                 db.refresh(tool)
-                return ToolModel.model_validate(tool)
+                return self._to_tool_model(tool, db=db)
         except Exception:
             return None
 
-    def delete_tool_by_id(self, id: str) -> bool:
+    def delete_tool_by_id(self, id: str, db: Optional[Session] = None) -> bool:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
+                AccessGrants.revoke_all_access("tool", id, db=db)
                 db.query(Tool).filter_by(id=id).delete()
                 db.commit()
 

backend/open_webui/models/users.py

@@ -1,18 +1,35 @@
 import time
 from typing import Optional
 
-from open_webui.internal.db import Base, JSONField, get_db
+from sqlalchemy.orm import Session, defer
+from open_webui.internal.db import Base, JSONField, get_db, get_db_context
 
 
 from open_webui.env import DATABASE_USER_ACTIVE_STATUS_UPDATE_INTERVAL
+
 from open_webui.models.chats import Chats
-from open_webui.models.groups import Groups
+from open_webui.models.groups import Groups, GroupMember
+from open_webui.models.channels import ChannelMember
+
 from open_webui.utils.misc import throttle
+from open_webui.utils.validate import validate_profile_image_url
 
 
-from pydantic import BaseModel, ConfigDict
-from sqlalchemy import BigInteger, Column, String, Text, Date
-from sqlalchemy import or_
+from pydantic import BaseModel, ConfigDict, field_validator, model_validator
+from sqlalchemy import (
+    BigInteger,
+    JSON,
+    Column,
+    String,
+    Boolean,
+    Text,
+    Date,
+    exists,
+    select,
+    cast,
+)
+from sqlalchemy import or_, case, func
+from sqlalchemy.dialects.postgresql import JSONB
 
 import datetime
 
@@ -21,59 +38,73 @@ import datetime
 ####################
 
 
-class User(Base):
-    __tablename__ = "user"
-
-    id = Column(String, primary_key=True)
-    name = Column(String)
-
-    email = Column(String)
-    username = Column(String(50), nullable=True)
-
-    role = Column(String)
-    profile_image_url = Column(Text)
-
-    bio = Column(Text, nullable=True)
-    gender = Column(Text, nullable=True)
-    date_of_birth = Column(Date, nullable=True)
-
-    info = Column(JSONField, nullable=True)
-    settings = Column(JSONField, nullable=True)
-
-    api_key = Column(String, nullable=True, unique=True)
-    oauth_sub = Column(Text, unique=True)
-
-    last_active_at = Column(BigInteger)
-
-    updated_at = Column(BigInteger)
-    created_at = Column(BigInteger)
-
-
 class UserSettings(BaseModel):
     ui: Optional[dict] = {}
     model_config = ConfigDict(extra="allow")
     pass
 
 
+class User(Base):
+    __tablename__ = "user"
+
+    id = Column(String, primary_key=True, unique=True)
+    email = Column(String)
+    username = Column(String(50), nullable=True)
+    role = Column(String)
+
+    name = Column(String)
+
+    profile_image_url = Column(Text)
+    profile_banner_image_url = Column(Text, nullable=True)
+
+    bio = Column(Text, nullable=True)
+    gender = Column(Text, nullable=True)
+    date_of_birth = Column(Date, nullable=True)
+    timezone = Column(String, nullable=True)
+
+    presence_state = Column(String, nullable=True)
+    status_emoji = Column(String, nullable=True)
+    status_message = Column(Text, nullable=True)
+    status_expires_at = Column(BigInteger, nullable=True)
+
+    info = Column(JSON, nullable=True)
+    settings = Column(JSON, nullable=True)
+
+    oauth = Column(JSON, nullable=True)
+    scim = Column(JSON, nullable=True)
+
+    last_active_at = Column(BigInteger)
+    updated_at = Column(BigInteger)
+    created_at = Column(BigInteger)
+
+
 class UserModel(BaseModel):
     id: str
-    name: str
 
     email: str
     username: Optional[str] = None
-
     role: str = "pending"
-    profile_image_url: str
+
+    name: str
+
+    profile_image_url: Optional[str] = None
+    profile_banner_image_url: Optional[str] = None
 
     bio: Optional[str] = None
     gender: Optional[str] = None
     date_of_birth: Optional[datetime.date] = None
+    timezone: Optional[str] = None
+
+    presence_state: Optional[str] = None
+    status_emoji: Optional[str] = None
+    status_message: Optional[str] = None
+    status_expires_at: Optional[int] = None
 
     info: Optional[dict] = None
     settings: Optional[UserSettings] = None
 
-    api_key: Optional[str] = None
-    oauth_sub: Optional[str] = None
+    oauth: Optional[dict] = None
+    scim: Optional[dict] = None
 
     last_active_at: int  # timestamp in epoch
     updated_at: int  # timestamp in epoch
@@ -81,6 +112,44 @@ class UserModel(BaseModel):
 
     model_config = ConfigDict(from_attributes=True)
 
+    @model_validator(mode="after")
+    def set_profile_image_url(self):
+        if not self.profile_image_url:
+            self.profile_image_url = f"/api/v1/users/{self.id}/profile/image"
+        return self
+
+
+class UserStatusModel(UserModel):
+    is_active: bool = False
+
+    model_config = ConfigDict(from_attributes=True)
+
+
+class ApiKey(Base):
+    __tablename__ = "api_key"
+
+    id = Column(Text, primary_key=True, unique=True)
+    user_id = Column(Text, nullable=False)
+    key = Column(Text, unique=True, nullable=False)
+    data = Column(JSON, nullable=True)
+    expires_at = Column(BigInteger, nullable=True)
+    last_used_at = Column(BigInteger, nullable=True)
+    created_at = Column(BigInteger, nullable=False)
+    updated_at = Column(BigInteger, nullable=False)
+
+
+class ApiKeyModel(BaseModel):
+    id: str
+    user_id: str
+    key: str
+    data: Optional[dict] = None
+    expires_at: Optional[int] = None
+    last_used_at: Optional[int] = None
+    created_at: int  # timestamp in epoch
+    updated_at: int  # timestamp in epoch
+
+    model_config = ConfigDict(from_attributes=True)
+
 
 ####################
 # Forms
@@ -94,17 +163,44 @@ class UpdateProfileForm(BaseModel):
     gender: Optional[str] = None
     date_of_birth: Optional[datetime.date] = None
 
+    @field_validator("profile_image_url")
+    @classmethod
+    def check_profile_image_url(cls, v: str) -> str:
+        return validate_profile_image_url(v)
+
+
+class UserGroupIdsModel(UserModel):
+    group_ids: list[str] = []
+
+
+class UserModelResponse(UserModel):
+    model_config = ConfigDict(extra="allow")
+
 
 class UserListResponse(BaseModel):
-    users: list[UserModel]
+    users: list[UserModelResponse]
     total: int
 
 
-class UserInfoResponse(BaseModel):
+class UserGroupIdsListResponse(BaseModel):
+    users: list[UserGroupIdsModel]
+    total: int
+
+
+class UserStatus(BaseModel):
+    status_emoji: Optional[str] = None
+    status_message: Optional[str] = None
+    status_expires_at: Optional[int] = None
+
+
+class UserInfoResponse(UserStatus):
     id: str
     name: str
     email: str
     role: str
+    bio: Optional[str] = None
+    groups: Optional[list] = []
+    is_active: bool = False
 
 
 class UserIdNameResponse(BaseModel):
@@ -112,6 +208,12 @@ class UserIdNameResponse(BaseModel):
     name: str
 
 
+class UserIdNameStatusResponse(UserStatus):
+    id: str
+    name: str
+    is_active: Optional[bool] = None
+
+
 class UserInfoListResponse(BaseModel):
     users: list[UserInfoResponse]
     total: int
@@ -122,18 +224,18 @@ class UserIdNameListResponse(BaseModel):
     total: int
 
 
-class UserResponse(BaseModel):
-    id: str
-    name: str
-    email: str
-    role: str
-    profile_image_url: str
-
-
 class UserNameResponse(BaseModel):
     id: str
     name: str
     role: str
+
+
+class UserResponse(UserNameResponse):
+    email: str
+
+
+class UserProfileImageResponse(UserNameResponse):
+    email: str
     profile_image_url: str
 
 
@@ -149,6 +251,11 @@ class UserUpdateForm(BaseModel):
     profile_image_url: str
     password: Optional[str] = None
 
+    @field_validator("profile_image_url")
+    @classmethod
+    def check_profile_image_url(cls, v: str) -> str:
+        return validate_profile_image_url(v)
+
 
 class UsersTable:
     def insert_new_user(
@@ -158,20 +265,23 @@ class UsersTable:
         email: str,
         profile_image_url: str = "/user.png",
         role: str = "pending",
-        oauth_sub: Optional[str] = None,
+        username: Optional[str] = None,
+        oauth: Optional[dict] = None,
+        db: Optional[Session] = None,
     ) -> Optional[UserModel]:
-        with get_db() as db:
+        with get_db_context(db) as db:
             user = UserModel(
                 **{
                     "id": id,
-                    "name": name,
                     "email": email,
+                    "name": name,
                     "role": role,
                     "profile_image_url": profile_image_url,
                     "last_active_at": int(time.time()),
                     "created_at": int(time.time()),
                     "updated_at": int(time.time()),
-                    "oauth_sub": oauth_sub,
+                    "username": username,
+                    "oauth": oauth,
                 }
             )
             result = User(**user.model_dump())
@@ -183,35 +293,86 @@ class UsersTable:
             else:
                 return None
 
-    def get_user_by_id(self, id: str) -> Optional[UserModel]:
+    def get_user_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 user = db.query(User).filter_by(id=id).first()
                 return UserModel.model_validate(user)
         except Exception:
             return None
 
-    def get_user_by_api_key(self, api_key: str) -> Optional[UserModel]:
+    def get_user_by_api_key(
+        self, api_key: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                user = db.query(User).filter_by(api_key=api_key).first()
-                return UserModel.model_validate(user)
+            with get_db_context(db) as db:
+                user = (
+                    db.query(User)
+                    .join(ApiKey, User.id == ApiKey.user_id)
+                    .filter(ApiKey.key == api_key)
+                    .first()
+                )
+                return UserModel.model_validate(user) if user else None
         except Exception:
             return None
 
-    def get_user_by_email(self, email: str) -> Optional[UserModel]:
+    def get_user_by_email(
+        self, email: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                user = db.query(User).filter_by(email=email).first()
-                return UserModel.model_validate(user)
+            with get_db_context(db) as db:
+                user = (
+                    db.query(User)
+                    .filter(func.lower(User.email) == email.lower())
+                    .first()
+                )
+                return UserModel.model_validate(user) if user else None
         except Exception:
             return None
 
-    def get_user_by_oauth_sub(self, sub: str) -> Optional[UserModel]:
+    def get_user_by_oauth_sub(
+        self, provider: str, sub: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                user = db.query(User).filter_by(oauth_sub=sub).first()
-                return UserModel.model_validate(user)
+            with get_db_context(db) as db:  # type: Session
+                dialect_name = db.bind.dialect.name
+
+                query = db.query(User)
+                if dialect_name == "sqlite":
+                    query = query.filter(User.oauth.contains({provider: {"sub": sub}}))
+                elif dialect_name == "postgresql":
+                    query = query.filter(
+                        User.oauth[provider].cast(JSONB)["sub"].astext == sub
+                    )
+
+                user = query.first()
+                return UserModel.model_validate(user) if user else None
+        except Exception as e:
+            # You may want to log the exception here
+            return None
+
+    def get_user_by_scim_external_id(
+        self, provider: str, external_id: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
+        try:
+            with get_db_context(db) as db:  # type: Session
+                dialect_name = db.bind.dialect.name
+
+                query = db.query(User)
+                if dialect_name == "sqlite":
+                    query = query.filter(
+                        User.scim.contains({provider: {"external_id": external_id}})
+                    )
+                elif dialect_name == "postgresql":
+                    query = query.filter(
+                        User.scim[provider].cast(JSONB)["external_id"].astext
+                        == external_id
+                    )
+
+                user = query.first()
+                return UserModel.model_validate(user) if user else None
         except Exception:
             return None
 
@@ -220,9 +381,11 @@ class UsersTable:
         filter: Optional[dict] = None,
         skip: Optional[int] = None,
         limit: Optional[int] = None,
+        db: Optional[Session] = None,
     ) -> dict:
-        with get_db() as db:
-            query = db.query(User)
+        with get_db_context(db) as db:
+            # Join GroupMember so we can order by group_id when requested
+            query = db.query(User).options(defer(User.profile_image_url))
 
             if filter:
                 query_key = filter.get("query")
@@ -234,14 +397,76 @@ class UsersTable:
                         )
                     )
 
+                channel_id = filter.get("channel_id")
+                if channel_id:
+                    query = query.filter(
+                        exists(
+                            select(ChannelMember.id).where(
+                                ChannelMember.user_id == User.id,
+                                ChannelMember.channel_id == channel_id,
+                            )
+                        )
+                    )
+
+                user_ids = filter.get("user_ids")
+                group_ids = filter.get("group_ids")
+
+                if isinstance(user_ids, list) and isinstance(group_ids, list):
+                    # If both are empty lists, return no users
+                    if not user_ids and not group_ids:
+                        return {"users": [], "total": 0}
+
+                if user_ids:
+                    query = query.filter(User.id.in_(user_ids))
+
+                if group_ids:
+                    query = query.filter(
+                        exists(
+                            select(GroupMember.id).where(
+                                GroupMember.user_id == User.id,
+                                GroupMember.group_id.in_(group_ids),
+                            )
+                        )
+                    )
+
+                roles = filter.get("roles")
+                if roles:
+                    include_roles = [role for role in roles if not role.startswith("!")]
+                    exclude_roles = [role[1:] for role in roles if role.startswith("!")]
+
+                    if include_roles:
+                        query = query.filter(User.role.in_(include_roles))
+                    if exclude_roles:
+                        query = query.filter(~User.role.in_(exclude_roles))
+
                 order_by = filter.get("order_by")
                 direction = filter.get("direction")
 
-                if order_by == "name":
+                if order_by and order_by.startswith("group_id:"):
+                    group_id = order_by.split(":", 1)[1]
+
+                    # Subquery that checks if the user belongs to the group
+                    membership_exists = exists(
+                        select(GroupMember.id).where(
+                            GroupMember.user_id == User.id,
+                            GroupMember.group_id == group_id,
+                        )
+                    )
+
+                    # CASE: user in group → 1, user not in group → 0
+                    group_sort = case((membership_exists, 1), else_=0)
+
+                    if direction == "asc":
+                        query = query.order_by(group_sort.asc(), User.name.asc())
+                    else:
+                        query = query.order_by(group_sort.desc(), User.name.asc())
+
+                elif order_by == "name":
                     if direction == "asc":
                         query = query.order_by(User.name.asc())
                     else:
                         query = query.order_by(User.name.desc())
+
                 elif order_by == "email":
                     if direction == "asc":
                         query = query.order_by(User.email.asc())
@@ -274,41 +499,67 @@ class UsersTable:
             else:
                 query = query.order_by(User.created_at.desc())
 
-            if skip:
+            # Count BEFORE pagination
+            total = query.count()
+
+            # correct pagination logic
+            if skip is not None:
                 query = query.offset(skip)
-            if limit:
+            if limit is not None:
                 query = query.limit(limit)
 
             users = query.all()
             return {
                 "users": [UserModel.model_validate(user) for user in users],
-                "total": db.query(User).count(),
+                "total": total,
             }
 
-    def get_users_by_user_ids(self, user_ids: list[str]) -> list[UserModel]:
-        with get_db() as db:
-            users = db.query(User).filter(User.id.in_(user_ids)).all()
+    def get_users_by_group_id(
+        self, group_id: str, db: Optional[Session] = None
+    ) -> list[UserModel]:
+        with get_db_context(db) as db:
+            users = (
+                db.query(User)
+                .options(defer(User.profile_image_url))
+                .join(GroupMember, User.id == GroupMember.user_id)
+                .filter(GroupMember.group_id == group_id)
+                .all()
+            )
             return [UserModel.model_validate(user) for user in users]
 
-    def get_num_users(self) -> Optional[int]:
-        with get_db() as db:
+    def get_users_by_user_ids(
+        self, user_ids: list[str], db: Optional[Session] = None
+    ) -> list[UserStatusModel]:
+        with get_db_context(db) as db:
+            users = (
+                db.query(User)
+                .options(defer(User.profile_image_url))
+                .filter(User.id.in_(user_ids))
+                .all()
+            )
+            return [UserModel.model_validate(user) for user in users]
+
+    def get_num_users(self, db: Optional[Session] = None) -> Optional[int]:
+        with get_db_context(db) as db:
             return db.query(User).count()
 
-    def has_users(self) -> bool:
-        with get_db() as db:
+    def has_users(self, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
             return db.query(db.query(User).exists()).scalar()
 
-    def get_first_user(self) -> UserModel:
+    def get_first_user(self, db: Optional[Session] = None) -> UserModel:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 user = db.query(User).order_by(User.created_at).first()
                 return UserModel.model_validate(user)
         except Exception:
             return None
 
-    def get_user_webhook_url_by_id(self, id: str) -> Optional[str]:
+    def get_user_webhook_url_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[str]:
         try:
-            with get_db() as db:
+            with get_db_context(db) as db:
                 user = db.query(User).filter_by(id=id).first()
 
                 if user.settings is None:
@@ -322,75 +573,168 @@ class UsersTable:
         except Exception:
             return None
 
-    def update_user_role_by_id(self, id: str, role: str) -> Optional[UserModel]:
+    def get_num_users_active_today(self, db: Optional[Session] = None) -> Optional[int]:
+        with get_db_context(db) as db:
+            current_timestamp = int(datetime.datetime.now().timestamp())
+            today_midnight_timestamp = current_timestamp - (current_timestamp % 86400)
+            query = db.query(User).filter(
+                User.last_active_at > today_midnight_timestamp
+            )
+            return query.count()
+
+    def update_user_role_by_id(
+        self, id: str, role: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                db.query(User).filter_by(id=id).update({"role": role})
-                db.commit()
+            with get_db_context(db) as db:
                 user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+                user.role = role
+                db.commit()
+                db.refresh(user)
+                return UserModel.model_validate(user)
+        except Exception:
+            return None
+
+    def update_user_status_by_id(
+        self, id: str, form_data: UserStatus, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
+        try:
+            with get_db_context(db) as db:
+                user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+                for key, value in form_data.model_dump(exclude_none=True).items():
+                    setattr(user, key, value)
+                db.commit()
+                db.refresh(user)
                 return UserModel.model_validate(user)
         except Exception:
             return None
 
     def update_user_profile_image_url_by_id(
-        self, id: str, profile_image_url: str
+        self, id: str, profile_image_url: str, db: Optional[Session] = None
     ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                db.query(User).filter_by(id=id).update(
-                    {"profile_image_url": profile_image_url}
-                )
-                db.commit()
-
+            with get_db_context(db) as db:
                 user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+                user.profile_image_url = profile_image_url
+                db.commit()
+                db.refresh(user)
                 return UserModel.model_validate(user)
         except Exception:
             return None
 
     @throttle(DATABASE_USER_ACTIVE_STATUS_UPDATE_INTERVAL)
-    def update_user_last_active_by_id(self, id: str) -> Optional[UserModel]:
-        try:
-            with get_db() as db:
-                db.query(User).filter_by(id=id).update(
-                    {"last_active_at": int(time.time())}
-                )
-                db.commit()
-
-                user = db.query(User).filter_by(id=id).first()
-                return UserModel.model_validate(user)
-        except Exception:
-            return None
-
-    def update_user_oauth_sub_by_id(
-        self, id: str, oauth_sub: str
+    def update_last_active_by_id(
+        self, id: str, db: Optional[Session] = None
     ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                db.query(User).filter_by(id=id).update({"oauth_sub": oauth_sub})
-                db.commit()
-
+            with get_db_context(db) as db:
                 user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+                user.last_active_at = int(time.time())
+                db.commit()
+                db.refresh(user)
                 return UserModel.model_validate(user)
         except Exception:
             return None
 
-    def update_user_by_id(self, id: str, updated: dict) -> Optional[UserModel]:
+    def update_user_oauth_by_id(
+        self, id: str, provider: str, sub: str, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
+        """
+        Update or insert an OAuth provider/sub pair into the user's oauth JSON field.
+        Example resulting structure:
+            {
+                "google": { "sub": "123" },
+                "github": { "sub": "abc" }
+            }
+        """
         try:
-            with get_db() as db:
-                db.query(User).filter_by(id=id).update(updated)
+            with get_db_context(db) as db:
+                user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+
+                # Load existing oauth JSON or create empty
+                oauth = user.oauth or {}
+
+                # Update or insert provider entry
+                oauth[provider] = {"sub": sub}
+
+                # Persist updated JSON
+                db.query(User).filter_by(id=id).update({"oauth": oauth})
                 db.commit()
 
-                user = db.query(User).filter_by(id=id).first()
                 return UserModel.model_validate(user)
-                # return UserModel(**user.dict())
+
+        except Exception:
+            return None
+
+    def update_user_scim_by_id(
+        self,
+        id: str,
+        provider: str,
+        external_id: str,
+        db: Optional[Session] = None,
+    ) -> Optional[UserModel]:
+        """
+        Update or insert a SCIM provider/external_id pair into the user's scim JSON field.
+        Example resulting structure:
+            {
+                "microsoft": { "external_id": "abc" },
+                "okta": { "external_id": "def" }
+            }
+        """
+        try:
+            with get_db_context(db) as db:
+                user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+
+                scim = user.scim or {}
+                scim[provider] = {"external_id": external_id}
+
+                db.query(User).filter_by(id=id).update({"scim": scim})
+                db.commit()
+
+                return UserModel.model_validate(user)
+
+        except Exception:
+            return None
+
+    def update_user_by_id(
+        self, id: str, updated: dict, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
+        try:
+            with get_db_context(db) as db:
+                user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+                for key, value in updated.items():
+                    setattr(user, key, value)
+                db.commit()
+                db.refresh(user)
+                return UserModel.model_validate(user)
         except Exception as e:
             print(e)
             return None
 
-    def update_user_settings_by_id(self, id: str, updated: dict) -> Optional[UserModel]:
+    def update_user_settings_by_id(
+        self, id: str, updated: dict, db: Optional[Session] = None
+    ) -> Optional[UserModel]:
         try:
-            with get_db() as db:
-                user_settings = db.query(User).filter_by(id=id).first().settings
+            with get_db_context(db) as db:
+                user = db.query(User).filter_by(id=id).first()
+                if not user:
+                    return None
+
+                user_settings = user.settings
 
                 if user_settings is None:
                     user_settings = {}
@@ -405,15 +749,15 @@ class UsersTable:
         except Exception:
             return None
 
-    def delete_user_by_id(self, id: str) -> bool:
+    def delete_user_by_id(self, id: str, db: Optional[Session] = None) -> bool:
         try:
             # Remove User from Groups
             Groups.remove_user_from_all_groups(id)
 
             # Delete User Chats
-            result = Chats.delete_chats_by_user_id(id)
+            result = Chats.delete_chats_by_user_id(id, db=db)
             if result:
-                with get_db() as db:
+                with get_db_context(db) as db:
                     # Delete User
                     db.query(User).filter_by(id=id).delete()
                     db.commit()
@@ -424,35 +768,89 @@ class UsersTable:
         except Exception:
             return False
 
-    def update_user_api_key_by_id(self, id: str, api_key: str) -> bool:
+    def get_user_api_key_by_id(
+        self, id: str, db: Optional[Session] = None
+    ) -> Optional[str]:
         try:
-            with get_db() as db:
-                result = db.query(User).filter_by(id=id).update({"api_key": api_key})
-                db.commit()
-                return True if result == 1 else False
-        except Exception:
-            return False
-
-    def get_user_api_key_by_id(self, id: str) -> Optional[str]:
-        try:
-            with get_db() as db:
-                user = db.query(User).filter_by(id=id).first()
-                return user.api_key
+            with get_db_context(db) as db:
+                api_key = db.query(ApiKey).filter_by(user_id=id).first()
+                return api_key.key if api_key else None
         except Exception:
             return None
 
-    def get_valid_user_ids(self, user_ids: list[str]) -> list[str]:
-        with get_db() as db:
+    def update_user_api_key_by_id(
+        self, id: str, api_key: str, db: Optional[Session] = None
+    ) -> bool:
+        try:
+            with get_db_context(db) as db:
+                db.query(ApiKey).filter_by(user_id=id).delete()
+                db.commit()
+
+                now = int(time.time())
+                new_api_key = ApiKey(
+                    id=f"key_{id}",
+                    user_id=id,
+                    key=api_key,
+                    created_at=now,
+                    updated_at=now,
+                )
+                db.add(new_api_key)
+                db.commit()
+
+                return True
+
+        except Exception:
+            return False
+
+    def delete_user_api_key_by_id(self, id: str, db: Optional[Session] = None) -> bool:
+        try:
+            with get_db_context(db) as db:
+                db.query(ApiKey).filter_by(user_id=id).delete()
+                db.commit()
+                return True
+        except Exception:
+            return False
+
+    def get_valid_user_ids(
+        self, user_ids: list[str], db: Optional[Session] = None
+    ) -> list[str]:
+        with get_db_context(db) as db:
             users = db.query(User).filter(User.id.in_(user_ids)).all()
             return [user.id for user in users]
 
-    def get_super_admin_user(self) -> Optional[UserModel]:
-        with get_db() as db:
+    def get_super_admin_user(self, db: Optional[Session] = None) -> Optional[UserModel]:
+        with get_db_context(db) as db:
             user = db.query(User).filter_by(role="admin").first()
             if user:
                 return UserModel.model_validate(user)
             else:
                 return None
 
+    def get_active_user_count(self, db: Optional[Session] = None) -> int:
+        with get_db_context(db) as db:
+            # Consider user active if last_active_at within the last 3 minutes
+            three_minutes_ago = int(time.time()) - 180
+            count = (
+                db.query(User).filter(User.last_active_at >= three_minutes_ago).count()
+            )
+            return count
+
+    @staticmethod
+    def is_active(user: UserModel) -> bool:
+        """Compute active status from an already-loaded UserModel (no DB hit)."""
+        if user.last_active_at:
+            three_minutes_ago = int(time.time()) - 180
+            return user.last_active_at >= three_minutes_ago
+        return False
+
+    def is_user_active(self, user_id: str, db: Optional[Session] = None) -> bool:
+        with get_db_context(db) as db:
+            user = db.query(User).filter_by(id=user_id).first()
+            if user and user.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 False
+
 
 Users = UsersTable()

backend/open_webui/retrieval/loaders/external_document.py

@@ -5,10 +5,9 @@ from urllib.parse import quote
 
 from langchain_core.document_loaders import BaseLoader
 from langchain_core.documents import Document
-from open_webui.env import SRC_LOG_LEVELS
+from open_webui.utils.headers import include_user_info_headers
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class ExternalDocumentLoader(BaseLoader):
@@ -18,6 +17,7 @@ class ExternalDocumentLoader(BaseLoader):
         url: str,
         api_key: str,
         mime_type=None,
+        user=None,
         **kwargs,
     ) -> None:
         self.url = url
@@ -26,6 +26,8 @@ class ExternalDocumentLoader(BaseLoader):
         self.file_path = file_path
         self.mime_type = mime_type
 
+        self.user = user
+
     def load(self) -> List[Document]:
         with open(self.file_path, "rb") as f:
             data = f.read()
@@ -42,6 +44,9 @@ class ExternalDocumentLoader(BaseLoader):
         except:
             pass
 
+        if self.user is not None:
+            headers = include_user_info_headers(headers, self.user)
+
         url = self.url
         if url.endswith("/"):
             url = url[:-1]

backend/open_webui/retrieval/loaders/external_web.py

@@ -4,10 +4,8 @@ from typing import Iterator, List, Union
 
 from langchain_core.document_loaders import BaseLoader
 from langchain_core.documents import Document
-from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class ExternalWebLoader(BaseLoader):

backend/open_webui/retrieval/loaders/main.py

@@ -27,13 +27,13 @@ from open_webui.retrieval.loaders.external_document import ExternalDocumentLoade
 
 from open_webui.retrieval.loaders.mistral import MistralLoader
 from open_webui.retrieval.loaders.datalab_marker import DatalabMarkerLoader
+from open_webui.retrieval.loaders.mineru import MinerULoader
 
 
-from open_webui.env import SRC_LOG_LEVELS, GLOBAL_LOG_LEVEL
+from open_webui.env import GLOBAL_LOG_LEVEL, REQUESTS_VERIFY
 
 logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 known_source_ext = [
     "go",
@@ -114,7 +114,7 @@ class TikaLoader:
             endpoint += "/"
         endpoint += "tika/text"
 
-        r = requests.put(endpoint, data=data, headers=headers)
+        r = requests.put(endpoint, data=data, headers=headers, verify=REQUESTS_VERIFY)
 
         if r.ok:
             raw_metadata = r.json()
@@ -131,8 +131,9 @@ class TikaLoader:
 
 
 class DoclingLoader:
-    def __init__(self, url, file_path=None, mime_type=None, params=None):
+    def __init__(self, url, api_key=None, file_path=None, mime_type=None, params=None):
         self.url = url.rstrip("/")
+        self.api_key = api_key
         self.file_path = file_path
         self.mime_type = mime_type
 
@@ -140,68 +141,25 @@ class DoclingLoader:
 
     def load(self) -> list[Document]:
         with open(self.file_path, "rb") as f:
-            files = {
-                "files": (
-                    self.file_path,
-                    f,
-                    self.mime_type or "application/octet-stream",
-                )
-            }
+            headers = {}
+            if self.api_key:
+                headers["X-Api-Key"] = f"{self.api_key}"
 
-            params = {"image_export_mode": "placeholder"}
-
-            if self.params:
-                if self.params.get("do_picture_description"):
-                    params["do_picture_description"] = self.params.get(
-                        "do_picture_description"
+            r = requests.post(
+                f"{self.url}/v1/convert/file",
+                files={
+                    "files": (
+                        self.file_path,
+                        f,
+                        self.mime_type or "application/octet-stream",
                     )
-
-                    picture_description_mode = self.params.get(
-                        "picture_description_mode", ""
-                    ).lower()
-
-                    if picture_description_mode == "local" and self.params.get(
-                        "picture_description_local", {}
-                    ):
-                        params["picture_description_local"] = json.dumps(
-                            self.params.get("picture_description_local", {})
-                        )
-
-                    elif picture_description_mode == "api" and self.params.get(
-                        "picture_description_api", {}
-                    ):
-                        params["picture_description_api"] = json.dumps(
-                            self.params.get("picture_description_api", {})
-                        )
-
-                params["do_ocr"] = self.params.get("do_ocr")
-
-                params["force_ocr"] = self.params.get("force_ocr")
-
-                if (
-                    self.params.get("do_ocr")
-                    and self.params.get("ocr_engine")
-                    and self.params.get("ocr_lang")
-                ):
-                    params["ocr_engine"] = self.params.get("ocr_engine")
-                    params["ocr_lang"] = [
-                        lang.strip()
-                        for lang in self.params.get("ocr_lang").split(",")
-                        if lang.strip()
-                    ]
-
-                if self.params.get("pdf_backend"):
-                    params["pdf_backend"] = self.params.get("pdf_backend")
-
-                if self.params.get("table_mode"):
-                    params["table_mode"] = self.params.get("table_mode")
-
-                if self.params.get("pipeline"):
-                    params["pipeline"] = self.params.get("pipeline")
-
-            endpoint = f"{self.url}/v1/convert/file"
-            r = requests.post(endpoint, files=files, data=params)
-
+                },
+                data={
+                    "image_export_mode": "placeholder",
+                    **self.params,
+                },
+                headers=headers,
+            )
         if r.ok:
             result = r.json()
             document_data = result.get("document", {})
@@ -210,7 +168,6 @@ class DoclingLoader:
             metadata = {"Content-Type": self.mime_type} if self.mime_type else {}
 
             log.debug("Docling extracted text: %s", text)
-
             return [Document(page_content=text, metadata=metadata)]
         else:
             error_msg = f"Error calling Docling API: {r.reason}"
@@ -227,6 +184,7 @@ class DoclingLoader:
 class Loader:
     def __init__(self, engine: str = "", **kwargs):
         self.engine = engine
+        self.user = kwargs.get("user", None)
         self.kwargs = kwargs
 
     def load(
@@ -263,6 +221,7 @@ class Loader:
                 url=self.kwargs.get("EXTERNAL_DOCUMENT_LOADER_URL"),
                 api_key=self.kwargs.get("EXTERNAL_DOCUMENT_LOADER_API_KEY"),
                 mime_type=file_content_type,
+                user=self.user,
             )
         elif self.engine == "tika" and self.kwargs.get("TIKA_SERVER_URL"):
             if self._is_text_file(file_ext, file_content_type):
@@ -271,7 +230,6 @@ class Loader:
                 loader = TikaLoader(
                     url=self.kwargs.get("TIKA_SERVER_URL"),
                     file_path=file_path,
-                    mime_type=file_content_type,
                     extract_images=self.kwargs.get("PDF_EXTRACT_IMAGES"),
                 )
         elif (
@@ -338,6 +296,7 @@ class Loader:
 
                 loader = DoclingLoader(
                     url=self.kwargs.get("DOCLING_SERVER_URL"),
+                    api_key=self.kwargs.get("DOCLING_API_KEY", None),
                     file_path=file_path,
                     mime_type=file_content_type,
                     params=params,
@@ -360,13 +319,34 @@ class Loader:
                     file_path=file_path,
                     api_endpoint=self.kwargs.get("DOCUMENT_INTELLIGENCE_ENDPOINT"),
                     api_key=self.kwargs.get("DOCUMENT_INTELLIGENCE_KEY"),
+                    api_model=self.kwargs.get("DOCUMENT_INTELLIGENCE_MODEL"),
                 )
             else:
                 loader = AzureAIDocumentIntelligenceLoader(
                     file_path=file_path,
                     api_endpoint=self.kwargs.get("DOCUMENT_INTELLIGENCE_ENDPOINT"),
                     azure_credential=DefaultAzureCredential(),
+                    api_model=self.kwargs.get("DOCUMENT_INTELLIGENCE_MODEL"),
                 )
+        elif self.engine == "mineru" and file_ext in [
+            "pdf"
+        ]:  # MinerU currently only supports PDF
+
+            mineru_timeout = self.kwargs.get("MINERU_API_TIMEOUT", 300)
+            if mineru_timeout:
+                try:
+                    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"),
+                api_url=self.kwargs.get("MINERU_API_URL", "http://localhost:8000"),
+                api_key=self.kwargs.get("MINERU_API_KEY", ""),
+                params=self.kwargs.get("MINERU_PARAMS", {}),
+                timeout=mineru_timeout,
+            )
         elif (
             self.engine == "mistral_ocr"
             and self.kwargs.get("MISTRAL_OCR_API_KEY") != ""
@@ -374,21 +354,16 @@ class Loader:
             in ["pdf"]  # Mistral OCR currently only supports PDF and images
         ):
             loader = MistralLoader(
-                api_key=self.kwargs.get("MISTRAL_OCR_API_KEY"), file_path=file_path
-            )
-        elif (
-            self.engine == "external"
-            and self.kwargs.get("MISTRAL_OCR_API_KEY") != ""
-            and file_ext
-            in ["pdf"]  # Mistral OCR currently only supports PDF and images
-        ):
-            loader = MistralLoader(
-                api_key=self.kwargs.get("MISTRAL_OCR_API_KEY"), file_path=file_path
+                base_url=self.kwargs.get("MISTRAL_OCR_API_BASE_URL"),
+                api_key=self.kwargs.get("MISTRAL_OCR_API_KEY"),
+                file_path=file_path,
             )
         else:
             if file_ext == "pdf":
                 loader = PyPDFLoader(
-                    file_path, extract_images=self.kwargs.get("PDF_EXTRACT_IMAGES")
+                    file_path,
+                    extract_images=self.kwargs.get("PDF_EXTRACT_IMAGES"),
+                    mode=self.kwargs.get("PDF_LOADER_MODE", "page"),
                 )
             elif file_ext == "csv":
                 loader = CSVLoader(file_path, autodetect_encoding=True)

backend/open_webui/retrieval/loaders/mineru.py

@@ -0,0 +1,524 @@
+import os
+import time
+import requests
+import logging
+import tempfile
+import zipfile
+from typing import List, Optional
+from langchain_core.documents import Document
+from fastapi import HTTPException, status
+
+log = logging.getLogger(__name__)
+
+
+class MinerULoader:
+    """
+    MinerU document parser loader supporting both Cloud API and Local API modes.
+
+    Cloud API: Uses MinerU managed service with async task-based processing
+    Local API: Uses self-hosted MinerU API with synchronous processing
+    """
+
+    def __init__(
+        self,
+        file_path: str,
+        api_mode: str = "local",
+        api_url: str = "http://localhost:8000",
+        api_key: str = "",
+        params: dict = None,
+        timeout: Optional[int] = 300,
+    ):
+        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
+
+        # 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.page_ranges = self.params.pop("page_ranges", "")
+
+        # Validate API mode
+        if self.api_mode not in ["local", "cloud"]:
+            raise ValueError(
+                f"Invalid API mode: {self.api_mode}. Must be 'local' or 'cloud'"
+            )
+
+        # Validate Cloud API requirements
+        if self.api_mode == "cloud" and not self.api_key:
+            raise ValueError("API key is required for Cloud API mode")
+
+    def load(self) -> List[Document]:
+        """
+        Main entry point for loading and parsing the document.
+        Routes to Cloud or Local API based on api_mode.
+        """
+        try:
+            if self.api_mode == "cloud":
+                return self._load_cloud_api()
+            else:
+                return self._load_local_api()
+        except Exception as e:
+            log.error(f"Error loading document with MinerU: {e}")
+            raise
+
+    def _load_local_api(self) -> List[Document]:
+        """
+        Load document using Local API (synchronous).
+        Posts file to /file_parse endpoint and gets immediate response.
+        """
+        log.info(f"Using MinerU Local API at {self.api_url}")
+
+        filename = os.path.basename(self.file_path)
+
+        # Build form data for Local API
+        form_data = {
+            **self.params,
+            "return_md": "true",
+        }
+
+        # Page ranges (Local API uses start_page_id and end_page_id)
+        if self.page_ranges:
+            # For simplicity, if page_ranges is specified, log a warning
+            # Full page range parsing would require parsing the string
+            log.warning(
+                f"Page ranges '{self.page_ranges}' specified but Local API uses different format. "
+                "Consider using start_page_id/end_page_id parameters if needed."
+            )
+
+        try:
+            with open(self.file_path, "rb") as f:
+                files = {"files": (filename, f, "application/octet-stream")}
+
+                log.info(f"Sending file to MinerU Local API: {filename}")
+                log.debug(f"Local API parameters: {form_data}")
+
+                response = requests.post(
+                    f"{self.api_url}/file_parse",
+                    data=form_data,
+                    files=files,
+                    timeout=self.timeout,
+                )
+                response.raise_for_status()
+
+        except FileNotFoundError:
+            raise HTTPException(
+                status.HTTP_404_NOT_FOUND, detail=f"File not found: {self.file_path}"
+            )
+        except requests.Timeout:
+            raise HTTPException(
+                status.HTTP_504_GATEWAY_TIMEOUT,
+                detail="MinerU Local API request timed out",
+            )
+        except requests.HTTPError as e:
+            error_detail = f"MinerU Local API request failed: {e}"
+            if e.response is not None:
+                try:
+                    error_data = e.response.json()
+                    error_detail += f" - {error_data}"
+                except:
+                    error_detail += f" - {e.response.text}"
+            raise HTTPException(status.HTTP_400_BAD_REQUEST, detail=error_detail)
+        except Exception as e:
+            raise HTTPException(
+                status.HTTP_500_INTERNAL_SERVER_ERROR,
+                detail=f"Error calling MinerU Local API: {str(e)}",
+            )
+
+        # Parse response
+        try:
+            result = response.json()
+        except ValueError as e:
+            raise HTTPException(
+                status.HTTP_502_BAD_GATEWAY,
+                detail=f"Invalid JSON response from MinerU Local API: {e}",
+            )
+
+        # Extract markdown content from response
+        if "results" not in result:
+            raise HTTPException(
+                status.HTTP_502_BAD_GATEWAY,
+                detail="MinerU Local API response missing 'results' field",
+            )
+
+        results = result["results"]
+        if not results:
+            raise HTTPException(
+                status.HTTP_400_BAD_REQUEST,
+                detail="MinerU returned empty results",
+            )
+
+        # Get the first (and typically only) result
+        file_result = list(results.values())[0]
+        markdown_content = file_result.get("md_content", "")
+
+        if not markdown_content:
+            raise HTTPException(
+                status.HTTP_400_BAD_REQUEST,
+                detail="MinerU returned empty markdown content",
+            )
+
+        log.info(f"Successfully parsed document with MinerU Local API: {filename}")
+
+        # Create metadata
+        metadata = {
+            "source": filename,
+            "api_mode": "local",
+            "backend": result.get("backend", "unknown"),
+            "version": result.get("version", "unknown"),
+        }
+
+        return [Document(page_content=markdown_content, metadata=metadata)]
+
+    def _load_cloud_api(self) -> List[Document]:
+        """
+        Load document using Cloud API (asynchronous).
+        Uses batch upload endpoint to avoid need for public file URLs.
+        """
+        log.info(f"Using MinerU Cloud API at {self.api_url}")
+
+        filename = os.path.basename(self.file_path)
+
+        # Step 1: Request presigned upload URL
+        batch_id, upload_url = self._request_upload_url(filename)
+
+        # Step 2: Upload file to presigned URL
+        self._upload_to_presigned_url(upload_url)
+
+        # Step 3: Poll for results
+        result = self._poll_batch_status(batch_id, filename)
+
+        # Step 4: Download and extract markdown from ZIP
+        markdown_content = self._download_and_extract_zip(
+            result["full_zip_url"], filename
+        )
+
+        log.info(f"Successfully parsed document with MinerU Cloud API: {filename}")
+
+        # Create metadata
+        metadata = {
+            "source": filename,
+            "api_mode": "cloud",
+            "batch_id": batch_id,
+        }
+
+        return [Document(page_content=markdown_content, metadata=metadata)]
+
+    def _request_upload_url(self, filename: str) -> tuple:
+        """
+        Request presigned upload URL from Cloud API.
+        Returns (batch_id, upload_url).
+        """
+        headers = {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json",
+        }
+
+        # Build request body
+        request_body = {
+            **self.params,
+            "files": [
+                {
+                    "name": filename,
+                    "is_ocr": self.enable_ocr,
+                }
+            ],
+        }
+
+        # Add page ranges if specified
+        if self.page_ranges:
+            request_body["files"][0]["page_ranges"] = self.page_ranges
+
+        log.info(f"Requesting upload URL for: {filename}")
+        log.debug(f"Cloud API request body: {request_body}")
+
+        try:
+            response = requests.post(
+                f"{self.api_url}/file-urls/batch",
+                headers=headers,
+                json=request_body,
+                timeout=30,
+            )
+            response.raise_for_status()
+        except requests.HTTPError as e:
+            error_detail = f"Failed to request upload URL: {e}"
+            if e.response is not None:
+                try:
+                    error_data = e.response.json()
+                    error_detail += f" - {error_data.get('msg', error_data)}"
+                except:
+                    error_detail += f" - {e.response.text}"
+            raise HTTPException(status.HTTP_400_BAD_REQUEST, detail=error_detail)
+        except Exception as e:
+            raise HTTPException(
+                status.HTTP_500_INTERNAL_SERVER_ERROR,
+                detail=f"Error requesting upload URL: {str(e)}",
+            )
+
+        try:
+            result = response.json()
+        except ValueError as e:
+            raise HTTPException(
+                status.HTTP_502_BAD_GATEWAY,
+                detail=f"Invalid JSON response: {e}",
+            )
+
+        # Check for API error response
+        if result.get("code") != 0:
+            raise HTTPException(
+                status.HTTP_400_BAD_REQUEST,
+                detail=f"MinerU Cloud API error: {result.get('msg', 'Unknown error')}",
+            )
+
+        data = result.get("data", {})
+        batch_id = data.get("batch_id")
+        file_urls = data.get("file_urls", [])
+
+        if not batch_id or not file_urls:
+            raise HTTPException(
+                status.HTTP_502_BAD_GATEWAY,
+                detail="MinerU Cloud API response missing batch_id or file_urls",
+            )
+
+        upload_url = file_urls[0]
+        log.info(f"Received upload URL for batch: {batch_id}")
+
+        return batch_id, upload_url
+
+    def _upload_to_presigned_url(self, upload_url: str) -> None:
+        """
+        Upload file to presigned URL (no authentication needed).
+        """
+        log.info(f"Uploading file to presigned URL")
+
+        try:
+            with open(self.file_path, "rb") as f:
+                response = requests.put(
+                    upload_url,
+                    data=f,
+                    timeout=self.timeout,
+                )
+                response.raise_for_status()
+        except FileNotFoundError:
+            raise HTTPException(
+                status.HTTP_404_NOT_FOUND, detail=f"File not found: {self.file_path}"
+            )
+        except requests.Timeout:
+            raise HTTPException(
+                status.HTTP_504_GATEWAY_TIMEOUT,
+                detail="File upload to presigned URL timed out",
+            )
+        except requests.HTTPError as e:
+            raise HTTPException(
+                status.HTTP_400_BAD_REQUEST,
+                detail=f"Failed to upload file to presigned URL: {e}",
+            )
+        except Exception as e:
+            raise HTTPException(
+                status.HTTP_500_INTERNAL_SERVER_ERROR,
+                detail=f"Error uploading file: {str(e)}",
+            )
+
+        log.info("File uploaded successfully")
+
+    def _poll_batch_status(self, batch_id: str, filename: str) -> dict:
+        """
+        Poll batch status until completion.
+        Returns the result dict for the file.
+        """
+        headers = {
+            "Authorization": f"Bearer {self.api_key}",
+        }
+
+        max_iterations = 300  # 10 minutes max (2 seconds per iteration)
+        poll_interval = 2  # seconds
+
+        log.info(f"Polling batch status: {batch_id}")
+
+        for iteration in range(max_iterations):
+            try:
+                response = requests.get(
+                    f"{self.api_url}/extract-results/batch/{batch_id}",
+                    headers=headers,
+                    timeout=30,
+                )
+                response.raise_for_status()
+            except requests.HTTPError as e:
+                error_detail = f"Failed to poll batch status: {e}"
+                if e.response is not None:
+                    try:
+                        error_data = e.response.json()
+                        error_detail += f" - {error_data.get('msg', error_data)}"
+                    except:
+                        error_detail += f" - {e.response.text}"
+                raise HTTPException(status.HTTP_400_BAD_REQUEST, detail=error_detail)
+            except Exception as e:
+                raise HTTPException(
+                    status.HTTP_500_INTERNAL_SERVER_ERROR,
+                    detail=f"Error polling batch status: {str(e)}",
+                )
+
+            try:
+                result = response.json()
+            except ValueError as e:
+                raise HTTPException(
+                    status.HTTP_502_BAD_GATEWAY,
+                    detail=f"Invalid JSON response while polling: {e}",
+                )
+
+            # Check for API error response
+            if result.get("code") != 0:
+                raise HTTPException(
+                    status.HTTP_400_BAD_REQUEST,
+                    detail=f"MinerU Cloud API error: {result.get('msg', 'Unknown error')}",
+                )
+
+            data = result.get("data", {})
+            extract_result = data.get("extract_result", [])
+
+            # Find our file in the batch results
+            file_result = None
+            for item in extract_result:
+                if item.get("file_name") == filename:
+                    file_result = item
+                    break
+
+            if not file_result:
+                raise HTTPException(
+                    status.HTTP_502_BAD_GATEWAY,
+                    detail=f"File {filename} not found in batch results",
+                )
+
+            state = file_result.get("state")
+
+            if state == "done":
+                log.info(f"Processing complete for {filename}")
+                return file_result
+            elif state == "failed":
+                error_msg = file_result.get("err_msg", "Unknown error")
+                raise HTTPException(
+                    status.HTTP_400_BAD_REQUEST,
+                    detail=f"MinerU processing failed: {error_msg}",
+                )
+            elif state in ["waiting-file", "pending", "running", "converting"]:
+                # Still processing
+                if iteration % 10 == 0:  # Log every 20 seconds
+                    log.info(
+                        f"Processing status: {state} (iteration {iteration + 1}/{max_iterations})"
+                    )
+                time.sleep(poll_interval)
+            else:
+                log.warning(f"Unknown state: {state}")
+                time.sleep(poll_interval)
+
+        # Timeout
+        raise HTTPException(
+            status.HTTP_504_GATEWAY_TIMEOUT,
+            detail="MinerU processing timed out after 10 minutes",
+        )
+
+    def _download_and_extract_zip(self, zip_url: str, filename: str) -> str:
+        """
+        Download ZIP file from CDN and extract markdown content.
+        Returns the markdown content as a string.
+        """
+        log.info(f"Downloading results from: {zip_url}")
+
+        try:
+            response = requests.get(zip_url, timeout=60)
+            response.raise_for_status()
+        except requests.HTTPError as e:
+            raise HTTPException(
+                status.HTTP_400_BAD_REQUEST,
+                detail=f"Failed to download results ZIP: {e}",
+            )
+        except Exception as e:
+            raise HTTPException(
+                status.HTTP_500_INTERNAL_SERVER_ERROR,
+                detail=f"Error downloading results: {str(e)}",
+            )
+
+        # Save ZIP to temporary file and extract
+        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)
+
+                # 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}")
+                    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}"
+                        )
+                    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 Exception as e:
+            raise HTTPException(
+                status.HTTP_500_INTERNAL_SERVER_ERROR,
+                detail=f"Error extracting ZIP: {str(e)}",
+            )
+
+        if not markdown_content:
+            raise HTTPException(
+                status.HTTP_400_BAD_REQUEST,
+                detail="Extracted markdown content is empty",
+            )
+
+        log.info(
+            f"Successfully extracted markdown content ({len(markdown_content)} characters)"
+        )
+        return markdown_content

backend/open_webui/retrieval/loaders/mistral.py

@@ -9,11 +9,10 @@ from typing import List, Dict, Any
 from contextlib import asynccontextmanager
 
 from langchain_core.documents import Document
-from open_webui.env import SRC_LOG_LEVELS, GLOBAL_LOG_LEVEL
+from open_webui.env import GLOBAL_LOG_LEVEL
 
 logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class MistralLoader:
@@ -30,10 +29,9 @@ class MistralLoader:
     - Enhanced error handling with retryable error classification
     """
 
-    BASE_API_URL = "https://api.mistral.ai/v1"
-
     def __init__(
         self,
+        base_url: str,
         api_key: str,
         file_path: str,
         timeout: int = 300,  # 5 minutes default
@@ -55,6 +53,9 @@ class MistralLoader:
         if not os.path.exists(file_path):
             raise FileNotFoundError(f"File not found at {file_path}")
 
+        self.base_url = (
+            base_url.rstrip("/") if base_url else "https://api.mistral.ai/v1"
+        )
         self.api_key = api_key
         self.file_path = file_path
         self.timeout = timeout
@@ -240,7 +241,7 @@ class MistralLoader:
         in a context manager to minimize memory usage duration.
         """
         log.info("Uploading file to Mistral API")
-        url = f"{self.BASE_API_URL}/files"
+        url = f"{self.base_url}/files"
 
         def upload_request():
             # MEMORY OPTIMIZATION: Use context manager to minimize file handle lifetime
@@ -275,7 +276,7 @@ class MistralLoader:
 
     async def _upload_file_async(self, session: aiohttp.ClientSession) -> str:
         """Async file upload with streaming for better memory efficiency."""
-        url = f"{self.BASE_API_URL}/files"
+        url = f"{self.base_url}/files"
 
         async def upload_request():
             # Create multipart writer for streaming upload
@@ -321,7 +322,7 @@ class MistralLoader:
     def _get_signed_url(self, file_id: str) -> str:
         """Retrieves a temporary signed URL for the uploaded file (sync version)."""
         log.info(f"Getting signed URL for file ID: {file_id}")
-        url = f"{self.BASE_API_URL}/files/{file_id}/url"
+        url = f"{self.base_url}/files/{file_id}/url"
         params = {"expiry": 1}
         signed_url_headers = {**self.headers, "Accept": "application/json"}
 
@@ -346,7 +347,7 @@ class MistralLoader:
         self, session: aiohttp.ClientSession, file_id: str
     ) -> str:
         """Async signed URL retrieval."""
-        url = f"{self.BASE_API_URL}/files/{file_id}/url"
+        url = f"{self.base_url}/files/{file_id}/url"
         params = {"expiry": 1}
 
         headers = {**self.headers, "Accept": "application/json"}
@@ -373,7 +374,7 @@ class MistralLoader:
     def _process_ocr(self, signed_url: str) -> Dict[str, Any]:
         """Sends the signed URL to the OCR endpoint for processing (sync version)."""
         log.info("Processing OCR via Mistral API")
-        url = f"{self.BASE_API_URL}/ocr"
+        url = f"{self.base_url}/ocr"
         ocr_headers = {
             **self.headers,
             "Content-Type": "application/json",
@@ -407,7 +408,7 @@ class MistralLoader:
         self, session: aiohttp.ClientSession, signed_url: str
     ) -> Dict[str, Any]:
         """Async OCR processing with timing metrics."""
-        url = f"{self.BASE_API_URL}/ocr"
+        url = f"{self.base_url}/ocr"
 
         headers = {
             **self.headers,
@@ -446,7 +447,7 @@ class MistralLoader:
     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}")
-        url = f"{self.BASE_API_URL}/files/{file_id}"
+        url = f"{self.base_url}/files/{file_id}"
 
         try:
             response = requests.delete(
@@ -467,7 +468,7 @@ class MistralLoader:
             async def delete_request():
                 self._debug_log(f"Deleting file ID: {file_id}")
                 async with session.delete(
-                    url=f"{self.BASE_API_URL}/files/{file_id}",
+                    url=f"{self.base_url}/files/{file_id}",
                     headers=self.headers,
                     timeout=aiohttp.ClientTimeout(
                         total=self.cleanup_timeout

backend/open_webui/retrieval/loaders/tavily.py

@@ -4,10 +4,8 @@ from typing import Iterator, List, Literal, Union
 
 from langchain_core.document_loaders import BaseLoader
 from langchain_core.documents import Document
-from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class TavilyLoader(BaseLoader):

backend/open_webui/retrieval/loaders/youtube.py

@@ -4,10 +4,8 @@ from xml.etree.ElementTree import ParseError
 from typing import Any, Dict, Generator, List, Optional, Sequence, Union
 from urllib.parse import parse_qs, urlparse
 from langchain_core.documents import Document
-from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 ALLOWED_SCHEMES = {"http", "https"}
 ALLOWED_NETLOCS = {
@@ -83,6 +81,7 @@ class YoutubeLoader:
                 TranscriptsDisabled,
                 YouTubeTranscriptApi,
             )
+            from youtube_transcript_api.proxies import GenericProxyConfig
         except ImportError:
             raise ImportError(
                 'Could not import "youtube_transcript_api" Python package. '
@@ -90,10 +89,9 @@ class YoutubeLoader:
             )
 
         if self.proxy_url:
-            youtube_proxies = {
-                "http": self.proxy_url,
-                "https": self.proxy_url,
-            }
+            youtube_proxies = GenericProxyConfig(
+                http_url=self.proxy_url, https_url=self.proxy_url
+            )
             log.debug(f"Using proxy URL: {self.proxy_url[:14]}...")
         else:
             youtube_proxies = None

backend/open_webui/retrieval/models/colbert.py

@@ -5,12 +5,10 @@ import numpy as np
 from colbert.infra import ColBERTConfig
 from colbert.modeling.checkpoint import Checkpoint
 
-from open_webui.env import SRC_LOG_LEVELS
 
 from open_webui.retrieval.models.base_reranker import BaseReranker
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class ColBERT(BaseReranker):

backend/open_webui/retrieval/models/external.py

@@ -4,12 +4,11 @@ from typing import Optional, List, Tuple
 from urllib.parse import quote
 
 
-from open_webui.env import ENABLE_FORWARD_USER_INFO_HEADERS, SRC_LOG_LEVELS
+from open_webui.env import ENABLE_FORWARD_USER_INFO_HEADERS, REQUESTS_VERIFY
 from open_webui.retrieval.models.base_reranker import BaseReranker
-
+from open_webui.utils.headers import include_user_info_headers
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class ExternalReranker(BaseReranker):
@@ -18,10 +17,12 @@ class ExternalReranker(BaseReranker):
         api_key: str,
         url: str = "http://localhost:8080/v1/rerank",
         model: str = "reranker",
+        timeout: Optional[int] = None,
     ):
         self.api_key = api_key
         self.url = url
         self.model = model
+        self.timeout = timeout
 
     def predict(
         self, sentences: List[Tuple[str, str]], user=None
@@ -40,23 +41,20 @@ class ExternalReranker(BaseReranker):
             log.info(f"ExternalReranker:predict:model {self.model}")
             log.info(f"ExternalReranker:predict:query {query}")
 
+            headers = {
+                "Content-Type": "application/json",
+                "Authorization": f"Bearer {self.api_key}",
+            }
+
+            if ENABLE_FORWARD_USER_INFO_HEADERS and user:
+                headers = include_user_info_headers(headers, user)
+
             r = requests.post(
                 f"{self.url}",
-                headers={
-                    "Content-Type": "application/json",
-                    "Authorization": f"Bearer {self.api_key}",
-                    **(
-                        {
-                            "X-OpenWebUI-User-Name": quote(user.name, safe=" "),
-                            "X-OpenWebUI-User-Id": user.id,
-                            "X-OpenWebUI-User-Email": user.email,
-                            "X-OpenWebUI-User-Role": user.role,
-                        }
-                        if ENABLE_FORWARD_USER_INFO_HEADERS and user
-                        else {}
-                    ),
-                },
+                headers=headers,
                 json=payload,
+                timeout=self.timeout,
+                verify=REQUESTS_VERIFY,
             )
 
             r.raise_for_status()

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

@@ -24,10 +24,8 @@ from open_webui.config import (
     CHROMA_CLIENT_AUTH_PROVIDER,
     CHROMA_CLIENT_AUTH_CREDENTIALS,
 )
-from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class ChromaClient(VectorDBBase):
@@ -71,7 +69,11 @@ class ChromaClient(VectorDBBase):
         return self.client.delete_collection(name=collection_name)
 
     def search(
-        self, collection_name: str, vectors: list[list[float | int]], limit: int
+        self,
+        collection_name: str,
+        vectors: list[list[float | int]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
         try:
@@ -80,6 +82,7 @@ class ChromaClient(VectorDBBase):
                 result = collection.query(
                     query_embeddings=vectors,
                     n_results=limit,
+                    where=filter,
                 )
 
                 # chromadb has cosine distance, 2 (worst) -> 0 (best). Re-odering to 0 -> 1

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

@@ -153,7 +153,11 @@ class ElasticsearchClient(VectorDBBase):
 
     # Status: works
     def search(
-        self, collection_name: str, vectors: list[list[float]], limit: int
+        self,
+        collection_name: str,
+        vectors: list[list[float]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         query = {
             "size": limit,

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

@@ -25,10 +25,8 @@ from open_webui.config import (
     MILVUS_DISKANN_MAX_DEGREE,
     MILVUS_DISKANN_SEARCH_LIST_SIZE,
 )
-from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class MilvusClient(VectorDBBase):
@@ -181,7 +179,11 @@ class MilvusClient(VectorDBBase):
         )
 
     def search(
-        self, collection_name: str, vectors: list[list[float | int]], limit: int
+        self,
+        collection_name: str,
+        vectors: list[list[float | int]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
         collection_name = collection_name.replace("-", "_")
@@ -200,23 +202,24 @@ class MilvusClient(VectorDBBase):
     def query(self, collection_name: str, filter: dict, limit: int = -1):
         connections.connect(uri=MILVUS_URI, token=MILVUS_TOKEN, db_name=MILVUS_DB)
 
-        # Construct the filter string for querying
         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}"
             )
             return None
-        filter_string = " && ".join(
-            [
-                f'metadata["{key}"] == {json.dumps(value)}'
-                for key, value in filter.items()
-            ]
-        )
+
+        filter_expressions = []
+        for key, value in filter.items():
+            if isinstance(value, str):
+                filter_expressions.append(f'metadata["{key}"] == "{value}"')
+            else:
+                filter_expressions.append(f'metadata["{key}"] == {value}')
+
+        filter_string = " && ".join(filter_expressions)
 
         collection = Collection(f"{self.collection_prefix}_{collection_name}")
         collection.load()
-        all_results = []
 
         try:
             log.info(
@@ -224,24 +227,25 @@ class MilvusClient(VectorDBBase):
             )
 
             iterator = collection.query_iterator(
-                filter=filter_string,
+                expr=filter_string,
                 output_fields=[
                     "id",
                     "data",
                     "metadata",
                 ],
-                limit=limit,  # Pass the limit directly; -1 means no limit.
+                limit=limit if limit > 0 else -1,
             )
 
+            all_results = []
             while True:
-                result = iterator.next()
-                if not result:
+                batch = iterator.next()
+                if not batch:
                     iterator.close()
                     break
-                all_results += result
+                all_results.extend(batch)
 
-            log.info(f"Total results from query: {len(all_results)}")
-            return self._result_to_get_result([all_results])
+            log.debug(f"Total results from query: {len(all_results)}")
+            return self._result_to_get_result([all_results] if all_results else [[]])
 
         except Exception as e:
             log.exception(

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

@@ -12,7 +12,6 @@ from open_webui.config import (
     MILVUS_HNSW_EFCONSTRUCTION,
     MILVUS_IVF_FLAT_NLIST,
 )
-from open_webui.env import SRC_LOG_LEVELS
 from open_webui.retrieval.vector.main import (
     GetResult,
     SearchResult,
@@ -29,7 +28,6 @@ from pymilvus import (
 )
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 RESOURCE_ID_FIELD = "resource_id"
 
@@ -157,10 +155,13 @@ class MilvusClient(VectorDBBase):
             for item in items
         ]
         collection.insert(entities)
-        collection.flush()
 
     def search(
-        self, collection_name: str, vectors: List[List[float]], limit: int
+        self,
+        collection_name: str,
+        vectors: List[List[float]],
+        filter: Optional[Dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         if not vectors:
             return None
@@ -263,15 +264,23 @@ class MilvusClient(VectorDBBase):
                 else:
                     expr.append(f"metadata['{key}'] == {value}")
 
-        results = collection.query(
+        iterator = collection.query_iterator(
             expr=" and ".join(expr),
             output_fields=["id", "text", "metadata"],
-            limit=limit,
+            limit=limit if limit else -1,
         )
 
-        ids = [res["id"] for res in results]
-        documents = [res["text"] for res in results]
-        metadatas = [res["metadata"] for res in results]
+        all_results = []
+        while True:
+            batch = iterator.next()
+            if not batch:
+                iterator.close()
+                break
+            all_results.extend(batch)
+
+        ids = [res["id"] for res in all_results]
+        documents = [res["text"] for res in all_results]
+        metadatas = [res["metadata"] for res in all_results]
 
         return GetResult(ids=[ids], documents=[documents], metadatas=[metadatas])
 

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

@@ -0,0 +1,427 @@
+from typing import Optional, List, Dict, Any
+import logging
+import re
+import json
+from sqlalchemy import (
+    func,
+    literal,
+    cast,
+    column,
+    create_engine,
+    Column,
+    Integer,
+    MetaData,
+    LargeBinary,
+    select,
+    text,
+    Text,
+    Table,
+    values,
+)
+from sqlalchemy.sql import true
+from sqlalchemy.pool import NullPool, QueuePool
+
+from sqlalchemy.orm import declarative_base, scoped_session, sessionmaker
+from sqlalchemy.dialects.postgresql import JSONB, array
+from pgvector.sqlalchemy import Vector
+from sqlalchemy.ext.mutable import MutableDict
+from sqlalchemy.exc import NoSuchTableError
+
+from sqlalchemy.dialects.postgresql.psycopg2 import PGDialect_psycopg2
+from sqlalchemy.dialects import registry
+
+
+class OpenGaussDialect(PGDialect_psycopg2):
+    name = "opengauss"
+
+    def _get_server_version_info(self, connection):
+        try:
+            version = connection.exec_driver_sql("SELECT version()").scalar()
+            if not version:
+                return (9, 0, 0)
+
+            match = re.search(
+                r"openGauss\s+(\d+)\.(\d+)\.(\d+)(?:-\w+)?", version, re.IGNORECASE
+            )
+            if match:
+                return (int(match.group(1)), int(match.group(2)), int(match.group(3)))
+
+            return super()._get_server_version_info(connection)
+        except Exception:
+            return (9, 0, 0)
+
+
+# Register dialect
+registry.register("opengauss", __name__, "OpenGaussDialect")
+
+from open_webui.retrieval.vector.utils import process_metadata
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
+from open_webui.config import (
+    OPENGAUSS_DB_URL,
+    OPENGAUSS_INITIALIZE_MAX_VECTOR_LENGTH,
+    OPENGAUSS_POOL_SIZE,
+    OPENGAUSS_POOL_MAX_OVERFLOW,
+    OPENGAUSS_POOL_TIMEOUT,
+    OPENGAUSS_POOL_RECYCLE,
+)
+
+from open_webui.env import SRC_LOG_LEVELS
+
+VECTOR_LENGTH = OPENGAUSS_INITIALIZE_MAX_VECTOR_LENGTH
+Base = declarative_base()
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+class DocumentChunk(Base):
+    __tablename__ = "document_chunk"
+
+    id = Column(Text, primary_key=True)
+    vector = Column(Vector(dim=VECTOR_LENGTH), nullable=True)
+    collection_name = Column(Text, nullable=False)
+    text = Column(Text, nullable=True)
+    vmetadata = Column(MutableDict.as_mutable(JSONB), nullable=True)
+
+
+class OpenGaussClient(VectorDBBase):
+    def __init__(self) -> None:
+        if not OPENGAUSS_DB_URL:
+            from open_webui.internal.db import ScopedSession
+
+            self.session = ScopedSession
+        else:
+            engine_kwargs = {"pool_pre_ping": True, "dialect": OpenGaussDialect()}
+
+            if isinstance(OPENGAUSS_POOL_SIZE, int) and OPENGAUSS_POOL_SIZE > 0:
+                engine_kwargs.update(
+                    {
+                        "pool_size": OPENGAUSS_POOL_SIZE,
+                        "max_overflow": OPENGAUSS_POOL_MAX_OVERFLOW,
+                        "pool_timeout": OPENGAUSS_POOL_TIMEOUT,
+                        "pool_recycle": OPENGAUSS_POOL_RECYCLE,
+                        "poolclass": QueuePool,
+                    }
+                )
+            else:
+                engine_kwargs["poolclass"] = NullPool
+
+            engine = create_engine(OPENGAUSS_DB_URL, **engine_kwargs)
+
+            SessionLocal = sessionmaker(
+                autocommit=False, autoflush=False, bind=engine, expire_on_commit=False
+            )
+            self.session = scoped_session(SessionLocal)
+
+        try:
+            connection = self.session.connection()
+            Base.metadata.create_all(bind=connection)
+
+            self.session.execute(
+                text(
+                    "CREATE INDEX IF NOT EXISTS idx_document_chunk_vector "
+                    "ON document_chunk USING ivfflat (vector vector_cosine_ops) WITH (lists = 100);"
+                )
+            )
+            self.session.execute(
+                text(
+                    "CREATE INDEX IF NOT EXISTS idx_document_chunk_collection_name "
+                    "ON document_chunk (collection_name);"
+                )
+            )
+            self.session.commit()
+            log.info("OpenGauss vector database initialization completed.")
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"OpenGauss Initialization failed.: {e}")
+            raise
+
+    def check_vector_length(self) -> None:
+        metadata = MetaData()
+        try:
+            document_chunk_table = Table(
+                "document_chunk", metadata, autoload_with=self.session.bind
+            )
+        except NoSuchTableError:
+            return
+
+        if "vector" in document_chunk_table.columns:
+            vector_column = document_chunk_table.columns["vector"]
+            vector_type = vector_column.type
+            if isinstance(vector_type, Vector):
+                db_vector_length = vector_type.dim
+                if db_vector_length != VECTOR_LENGTH:
+                    raise Exception(
+                        f"Vector dimension mismatch: configured {VECTOR_LENGTH} vs. {db_vector_length} in the database."
+                    )
+            else:
+                raise Exception("The 'vector' column type is not Vector.")
+        else:
+            raise Exception(
+                "The 'vector' column does not exist in the 'document_chunk' table."
+            )
+
+    def adjust_vector_length(self, vector: List[float]) -> List[float]:
+        current_length = len(vector)
+        if current_length < VECTOR_LENGTH:
+            vector += [0.0] * (VECTOR_LENGTH - current_length)
+        elif current_length > VECTOR_LENGTH:
+            vector = vector[:VECTOR_LENGTH]
+        return vector
+
+    def insert(self, collection_name: str, items: List[VectorItem]) -> None:
+        try:
+            new_items = []
+            for item in items:
+                vector = self.adjust_vector_length(item["vector"])
+                new_chunk = DocumentChunk(
+                    id=item["id"],
+                    vector=vector,
+                    collection_name=collection_name,
+                    text=item["text"],
+                    vmetadata=process_metadata(item["metadata"]),
+                )
+                new_items.append(new_chunk)
+            self.session.bulk_save_objects(new_items)
+            self.session.commit()
+            log.info(
+                f"Inserting {len(new_items)} items into collection '{collection_name}'."
+            )
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Failed to insert data: {e}")
+            raise
+
+    def upsert(self, collection_name: str, items: List[VectorItem]) -> None:
+        try:
+            for item in items:
+                vector = self.adjust_vector_length(item["vector"])
+                existing = (
+                    self.session.query(DocumentChunk)
+                    .filter(DocumentChunk.id == item["id"])
+                    .first()
+                )
+                if existing:
+                    existing.vector = vector
+                    existing.text = item["text"]
+                    existing.vmetadata = process_metadata(item["metadata"])
+                    existing.collection_name = collection_name
+                else:
+                    new_chunk = DocumentChunk(
+                        id=item["id"],
+                        vector=vector,
+                        collection_name=collection_name,
+                        text=item["text"],
+                        vmetadata=process_metadata(item["metadata"]),
+                    )
+                    self.session.add(new_chunk)
+            self.session.commit()
+            log.info(
+                f"Inserting/updating {len(items)} items in collection '{collection_name}'."
+            )
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Failed to insert or update data.: {e}")
+            raise
+
+    def search(
+        self,
+        collection_name: str,
+        vectors: List[List[float]],
+        filter: Optional[Dict[str, Any]] = None,
+        limit: int = 10,
+    ) -> Optional[SearchResult]:
+        try:
+            if not vectors:
+                return None
+
+            vectors = [self.adjust_vector_length(vector) for vector in vectors]
+            num_queries = len(vectors)
+
+            def vector_expr(vector):
+                return cast(array(vector), Vector(VECTOR_LENGTH))
+
+            qid_col = column("qid", Integer)
+            q_vector_col = column("q_vector", Vector(VECTOR_LENGTH))
+            query_vectors = (
+                values(qid_col, q_vector_col)
+                .data(
+                    [(idx, vector_expr(vector)) for idx, vector in enumerate(vectors)]
+                )
+                .alias("query_vectors")
+            )
+
+            result_fields = [
+                DocumentChunk.id,
+                DocumentChunk.text,
+                DocumentChunk.vmetadata,
+                (DocumentChunk.vector.cosine_distance(query_vectors.c.q_vector)).label(
+                    "distance"
+                ),
+            ]
+
+            subq = (
+                select(*result_fields)
+                .where(DocumentChunk.collection_name == collection_name)
+                .order_by(
+                    DocumentChunk.vector.cosine_distance(query_vectors.c.q_vector)
+                )
+            )
+            if limit is not None:
+                subq = subq.limit(limit)
+            subq = subq.lateral("result")
+
+            stmt = (
+                select(
+                    query_vectors.c.qid,
+                    subq.c.id,
+                    subq.c.text,
+                    subq.c.vmetadata,
+                    subq.c.distance,
+                )
+                .select_from(query_vectors)
+                .join(subq, true())
+                .order_by(query_vectors.c.qid, subq.c.distance)
+            )
+
+            result_proxy = self.session.execute(stmt)
+            results = result_proxy.all()
+
+            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 row in results:
+                qid = int(row.qid)
+                ids[qid].append(row.id)
+                distances[qid].append((2.0 - row.distance) / 2.0)
+                documents[qid].append(row.text)
+                metadatas[qid].append(row.vmetadata)
+
+            self.session.rollback()
+            return SearchResult(
+                ids=ids, distances=distances, documents=documents, metadatas=metadatas
+            )
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Vector search failed: {e}")
+            return None
+
+    def query(
+        self, collection_name: str, filter: Dict[str, Any], limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        try:
+            query = self.session.query(DocumentChunk).filter(
+                DocumentChunk.collection_name == collection_name
+            )
+
+            for key, value in filter.items():
+                query = query.filter(DocumentChunk.vmetadata[key].astext == str(value))
+
+            if limit is not None:
+                query = query.limit(limit)
+
+            results = query.all()
+
+            if not results:
+                return None
+
+            ids = [[result.id for result in results]]
+            documents = [[result.text for result in results]]
+            metadatas = [[result.vmetadata for result in results]]
+
+            self.session.rollback()
+            return GetResult(ids=ids, documents=documents, metadatas=metadatas)
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Conditional query failed: {e}")
+            return None
+
+    def get(
+        self, collection_name: str, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        try:
+            query = self.session.query(DocumentChunk).filter(
+                DocumentChunk.collection_name == collection_name
+            )
+            if limit is not None:
+                query = query.limit(limit)
+
+            results = query.all()
+
+            if not results:
+                return None
+
+            ids = [[result.id for result in results]]
+            documents = [[result.text for result in results]]
+            metadatas = [[result.vmetadata for result in results]]
+
+            self.session.rollback()
+            return GetResult(ids=ids, documents=documents, metadatas=metadatas)
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Failed to retrieve data: {e}")
+            return None
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[List[str]] = None,
+        filter: Optional[Dict[str, Any]] = None,
+    ) -> None:
+        try:
+            query = self.session.query(DocumentChunk).filter(
+                DocumentChunk.collection_name == collection_name
+            )
+            if ids:
+                query = query.filter(DocumentChunk.id.in_(ids))
+            if filter:
+                for key, value in filter.items():
+                    query = query.filter(
+                        DocumentChunk.vmetadata[key].astext == str(value)
+                    )
+            deleted = query.delete(synchronize_session=False)
+            self.session.commit()
+            log.info(f"Deleted {deleted} items from collection '{collection_name}'")
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Failed to delete data: {e}")
+            raise
+
+    def reset(self) -> None:
+        try:
+            deleted = self.session.query(DocumentChunk).delete()
+            self.session.commit()
+            log.info(f"Reset completed. Deleted {deleted} items")
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Reset failed: {e}")
+            raise
+
+    def close(self) -> None:
+        pass
+
+    def has_collection(self, collection_name: str) -> bool:
+        try:
+            exists = (
+                self.session.query(DocumentChunk)
+                .filter(DocumentChunk.collection_name == collection_name)
+                .first()
+                is not None
+            )
+            self.session.rollback()
+            return exists
+        except Exception as e:
+            self.session.rollback()
+            log.exception(f"Failed to check collection existence: {e}")
+            return False
+
+    def delete_collection(self, collection_name: str) -> None:
+        self.delete(collection_name)
+        log.info(f"Collection '{collection_name}' has been deleted")

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

@@ -113,7 +113,11 @@ class OpenSearchClient(VectorDBBase):
         self.client.indices.delete(index=self._get_index_name(collection_name))
 
     def search(
-        self, collection_name: str, vectors: list[list[float | int]], limit: int
+        self,
+        collection_name: str,
+        vectors: list[list[float | int]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         try:
             if not self.has_collection(collection_name):
@@ -207,7 +211,7 @@ class OpenSearchClient(VectorDBBase):
                 for item in batch
             ]
             bulk(self.client, actions)
-        self.client.indices.refresh(self._get_index_name(collection_name))
+        self.client.indices.refresh(index=self._get_index_name(collection_name))
 
     def upsert(self, collection_name: str, items: list[VectorItem]):
         self._create_index_if_not_exists(
@@ -230,7 +234,7 @@ class OpenSearchClient(VectorDBBase):
                 for item in batch
             ]
             bulk(self.client, actions)
-        self.client.indices.refresh(self._get_index_name(collection_name))
+        self.client.indices.refresh(index=self._get_index_name(collection_name))
 
     def delete(
         self,
@@ -259,7 +263,7 @@ class OpenSearchClient(VectorDBBase):
             self.client.delete_by_query(
                 index=self._get_index_name(collection_name), body=query_body
             )
-        self.client.indices.refresh(self._get_index_name(collection_name))
+        self.client.indices.refresh(index=self._get_index_name(collection_name))
 
     def reset(self):
         indices = self.client.indices.get(index=f"{self.index_prefix}_*")

backend/open_webui/retrieval/vector/dbs/oracle23ai.py

@@ -55,10 +55,8 @@ from open_webui.config import (
     ORACLE_DB_POOL_MAX,
     ORACLE_DB_POOL_INCREMENT,
 )
-from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class Oracle23aiClient(VectorDBBase):
@@ -258,8 +256,7 @@ class Oracle23aiClient(VectorDBBase):
         with connection.cursor() as cursor:
             try:
                 log.info("Creating Table document_chunk")
-                cursor.execute(
-                    """
+                cursor.execute("""
                     BEGIN
                         EXECUTE IMMEDIATE '
                             CREATE TABLE IF NOT EXISTS document_chunk (
@@ -276,12 +273,10 @@ class Oracle23aiClient(VectorDBBase):
                                 RAISE;
                             END IF;
                     END;
-                """
-                )
+                """)
 
                 log.info("Creating Index document_chunk_collection_name_idx")
-                cursor.execute(
-                    """
+                cursor.execute("""
                     BEGIN
                         EXECUTE IMMEDIATE '
                             CREATE INDEX IF NOT EXISTS document_chunk_collection_name_idx
@@ -293,12 +288,10 @@ class Oracle23aiClient(VectorDBBase):
                                 RAISE;
                             END IF;
                     END;
-                """
-                )
+                """)
 
                 log.info("Creating VECTOR INDEX document_chunk_vector_ivf_idx")
-                cursor.execute(
-                    """
+                cursor.execute("""
                     BEGIN
                         EXECUTE IMMEDIATE '
                             CREATE VECTOR INDEX IF NOT EXISTS document_chunk_vector_ivf_idx 
@@ -314,8 +307,7 @@ class Oracle23aiClient(VectorDBBase):
                                 RAISE;
                             END IF;
                     END;
-                """
-                )
+                """)
 
                 connection.commit()
                 log.info("Database initialization completed successfully.")
@@ -523,7 +515,11 @@ class Oracle23aiClient(VectorDBBase):
                 raise
 
     def search(
-        self, collection_name: str, vectors: List[List[Union[float, int]]], limit: int
+        self,
+        collection_name: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         """
         Search for similar vectors in the database.
@@ -717,7 +713,7 @@ class Oracle23aiClient(VectorDBBase):
         )
 
         try:
-            limit = limit or 1000
+            limit = 1000  # Hardcoded limit for get operation
 
             with self.get_connection() as connection:
                 with connection.cursor() as cursor:

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

@@ -1,4 +1,4 @@
-from typing import Optional, List, Dict, Any
+from typing import Optional, List, Dict, Any, Tuple
 import logging
 import json
 from sqlalchemy import (
@@ -22,7 +22,7 @@ from sqlalchemy.pool import NullPool, QueuePool
 
 from sqlalchemy.orm import declarative_base, scoped_session, sessionmaker
 from sqlalchemy.dialects.postgresql import JSONB, array
-from pgvector.sqlalchemy import Vector
+from pgvector.sqlalchemy import Vector, HALFVEC
 from sqlalchemy.ext.mutable import MutableDict
 from sqlalchemy.exc import NoSuchTableError
 
@@ -44,15 +44,21 @@ from open_webui.config import (
     PGVECTOR_POOL_MAX_OVERFLOW,
     PGVECTOR_POOL_TIMEOUT,
     PGVECTOR_POOL_RECYCLE,
+    PGVECTOR_INDEX_METHOD,
+    PGVECTOR_HNSW_M,
+    PGVECTOR_HNSW_EF_CONSTRUCTION,
+    PGVECTOR_IVFFLAT_LISTS,
+    PGVECTOR_USE_HALFVEC,
 )
 
-from open_webui.env import SRC_LOG_LEVELS
-
 VECTOR_LENGTH = PGVECTOR_INITIALIZE_MAX_VECTOR_LENGTH
+USE_HALFVEC = PGVECTOR_USE_HALFVEC
+
+VECTOR_TYPE_FACTORY = HALFVEC if USE_HALFVEC else Vector
+VECTOR_OPCLASS = "halfvec_cosine_ops" if USE_HALFVEC else "vector_cosine_ops"
 Base = declarative_base()
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 def pgcrypto_encrypt(val, key):
@@ -67,7 +73,7 @@ class DocumentChunk(Base):
     __tablename__ = "document_chunk"
 
     id = Column(Text, primary_key=True)
-    vector = Column(Vector(dim=VECTOR_LENGTH), nullable=True)
+    vector = Column(VECTOR_TYPE_FACTORY(dim=VECTOR_LENGTH), nullable=True)
     collection_name = Column(Text, nullable=False)
 
     if PGVECTOR_PGCRYPTO:
@@ -83,9 +89,9 @@ class PgvectorClient(VectorDBBase):
 
         # if no pgvector uri, use the existing database connection
         if not PGVECTOR_DB_URL:
-            from open_webui.internal.db import Session
+            from open_webui.internal.db import ScopedSession
 
-            self.session = Session
+            self.session = ScopedSession
         else:
             if isinstance(PGVECTOR_POOL_SIZE, int):
                 if PGVECTOR_POOL_SIZE > 0:
@@ -114,34 +120,26 @@ class PgvectorClient(VectorDBBase):
             # Ensure the pgvector extension is available
             # Use a conditional check to avoid permission issues on Azure PostgreSQL
             if PGVECTOR_CREATE_EXTENSION:
-                self.session.execute(
-                    text(
-                        """
+                self.session.execute(text("""
                     DO $$
                     BEGIN
                     IF NOT EXISTS (SELECT 1 FROM pg_extension WHERE extname = 'vector') THEN
                         CREATE EXTENSION IF NOT EXISTS vector;
                     END IF;
                     END $$;
-                """
-                    )
-                )
+                """))
 
             if PGVECTOR_PGCRYPTO:
                 # Ensure the pgcrypto extension is available for encryption
                 # Use a conditional check to avoid permission issues on Azure PostgreSQL
-                self.session.execute(
-                    text(
-                        """
+                self.session.execute(text("""
                     DO $$
                     BEGIN
                        IF NOT EXISTS (SELECT 1 FROM pg_extension WHERE extname = 'pgcrypto') THEN
                           CREATE EXTENSION IF NOT EXISTS pgcrypto;
                        END IF;
                     END $$;
-                """
-                    )
-                )
+                """))
 
                 if not PGVECTOR_PGCRYPTO_KEY:
                     raise ValueError(
@@ -157,13 +155,9 @@ class PgvectorClient(VectorDBBase):
             connection = self.session.connection()
             Base.metadata.create_all(bind=connection)
 
-            # Create an index on the vector column if it doesn't exist
-            self.session.execute(
-                text(
-                    "CREATE INDEX IF NOT EXISTS idx_document_chunk_vector "
-                    "ON document_chunk USING ivfflat (vector vector_cosine_ops) WITH (lists = 100);"
-                )
-            )
+            index_method, index_options = self._vector_index_configuration()
+            self._ensure_vector_index(index_method, index_options)
+
             self.session.execute(
                 text(
                     "CREATE INDEX IF NOT EXISTS idx_document_chunk_collection_name "
@@ -177,6 +171,76 @@ class PgvectorClient(VectorDBBase):
             log.exception(f"Error during initialization: {e}")
             raise
 
+    @staticmethod
+    def _extract_index_method(index_def: Optional[str]) -> Optional[str]:
+        if not index_def:
+            return None
+        try:
+            after_using = index_def.lower().split("using ", 1)[1]
+            return after_using.split()[0]
+        except (IndexError, AttributeError):
+            return None
+
+    def _vector_index_configuration(self) -> Tuple[str, str]:
+        if PGVECTOR_INDEX_METHOD:
+            index_method = PGVECTOR_INDEX_METHOD
+            log.info(
+                "Using vector index method '%s' from PGVECTOR_INDEX_METHOD.",
+                index_method,
+            )
+        elif USE_HALFVEC:
+            index_method = "hnsw"
+            log.info(
+                "VECTOR_LENGTH=%s exceeds 2000; using halfvec column type with hnsw index.",
+                VECTOR_LENGTH,
+            )
+        else:
+            index_method = "ivfflat"
+
+        if index_method == "hnsw":
+            index_options = f"WITH (m = {PGVECTOR_HNSW_M}, ef_construction = {PGVECTOR_HNSW_EF_CONSTRUCTION})"
+        else:
+            index_options = f"WITH (lists = {PGVECTOR_IVFFLAT_LISTS})"
+
+        return index_method, index_options
+
+    def _ensure_vector_index(self, index_method: str, index_options: str) -> None:
+        index_name = "idx_document_chunk_vector"
+        existing_index_def = self.session.execute(
+            text("""
+                SELECT indexdef
+                FROM pg_indexes
+                WHERE schemaname = current_schema()
+                  AND tablename = 'document_chunk'
+                  AND indexname = :index_name
+                """),
+            {"index_name": index_name},
+        ).scalar()
+
+        existing_method = self._extract_index_method(existing_index_def)
+        if existing_method and existing_method != index_method:
+            raise RuntimeError(
+                f"Existing pgvector index '{index_name}' uses method '{existing_method}' but configuration now "
+                f"requires '{index_method}'. Automatic rebuild is disabled to prevent long-running maintenance. "
+                "Drop the index manually (optionally after tuning maintenance_work_mem/max_parallel_maintenance_workers) "
+                "and recreate it with the new method before restarting Open WebUI."
+            )
+
+        if not existing_index_def:
+            index_sql = (
+                f"CREATE INDEX IF NOT EXISTS {index_name} "
+                f"ON document_chunk USING {index_method} (vector {VECTOR_OPCLASS})"
+            )
+            if index_options:
+                index_sql = f"{index_sql} {index_options}"
+            self.session.execute(text(index_sql))
+            log.info(
+                "Ensured vector index '%s' using %s%s.",
+                index_name,
+                index_method,
+                f" {index_options}" if index_options else "",
+            )
+
     def check_vector_length(self) -> None:
         """
         Check if the VECTOR_LENGTH matches the existing vector column dimension in the database.
@@ -196,16 +260,19 @@ class PgvectorClient(VectorDBBase):
         if "vector" in document_chunk_table.columns:
             vector_column = document_chunk_table.columns["vector"]
             vector_type = vector_column.type
-            if isinstance(vector_type, Vector):
-                db_vector_length = vector_type.dim
-                if db_vector_length != VECTOR_LENGTH:
-                    raise Exception(
-                        f"VECTOR_LENGTH {VECTOR_LENGTH} does not match existing vector column dimension {db_vector_length}. "
-                        "Cannot change vector size after initialization without migrating the data."
-                    )
-            else:
+            expected_type = HALFVEC if USE_HALFVEC else Vector
+
+            if not isinstance(vector_type, expected_type):
                 raise Exception(
-                    "The 'vector' column exists but is not of type 'Vector'."
+                    "The 'vector' column type does not match the expected type "
+                    f"('{expected_type.__name__}') for VECTOR_LENGTH {VECTOR_LENGTH}."
+                )
+
+            db_vector_length = getattr(vector_type, "dim", None)
+            if db_vector_length is not None and db_vector_length != VECTOR_LENGTH:
+                raise Exception(
+                    f"VECTOR_LENGTH {VECTOR_LENGTH} does not match existing vector column dimension {db_vector_length}. "
+                    "Cannot change vector size after initialization without migrating the data."
                 )
         else:
             raise Exception(
@@ -232,8 +299,7 @@ class PgvectorClient(VectorDBBase):
                     # Ensure metadata is converted to its JSON text representation
                     json_metadata = json.dumps(item["metadata"])
                     self.session.execute(
-                        text(
-                            """
+                        text("""
                             INSERT INTO document_chunk
                             (id, vector, collection_name, text, vmetadata)
                             VALUES (
@@ -242,8 +308,7 @@ class PgvectorClient(VectorDBBase):
                                 pgp_sym_encrypt(:metadata_text, :key)
                             )
                             ON CONFLICT (id) DO NOTHING
-                        """
-                        ),
+                        """),
                         {
                             "id": item["id"],
                             "vector": vector,
@@ -285,8 +350,7 @@ class PgvectorClient(VectorDBBase):
                     vector = self.adjust_vector_length(item["vector"])
                     json_metadata = json.dumps(item["metadata"])
                     self.session.execute(
-                        text(
-                            """
+                        text("""
                             INSERT INTO document_chunk
                             (id, vector, collection_name, text, vmetadata)
                             VALUES (
@@ -299,8 +363,7 @@ class PgvectorClient(VectorDBBase):
                               collection_name = EXCLUDED.collection_name,
                               text = EXCLUDED.text,
                               vmetadata = EXCLUDED.vmetadata
-                        """
-                        ),
+                        """),
                         {
                             "id": item["id"],
                             "vector": vector,
@@ -349,7 +412,8 @@ class PgvectorClient(VectorDBBase):
         self,
         collection_name: str,
         vectors: List[List[float]],
-        limit: Optional[int] = None,
+        filter: Optional[Dict[str, Any]] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         try:
             if not vectors:
@@ -360,11 +424,11 @@ class PgvectorClient(VectorDBBase):
             num_queries = len(vectors)
 
             def vector_expr(vector):
-                return cast(array(vector), Vector(VECTOR_LENGTH))
+                return cast(array(vector), VECTOR_TYPE_FACTORY(VECTOR_LENGTH))
 
             # Create the values for query vectors
             qid_col = column("qid", Integer)
-            q_vector_col = column("q_vector", Vector(VECTOR_LENGTH))
+            q_vector_col = column("q_vector", VECTOR_TYPE_FACTORY(VECTOR_LENGTH))
             query_vectors = (
                 values(qid_col, q_vector_col)
                 .data(
@@ -397,9 +461,47 @@ class PgvectorClient(VectorDBBase):
             )
 
             # Build the lateral subquery for each query vector
+            where_clauses = [DocumentChunk.collection_name == collection_name]
+
+            # Apply metadata filter if provided
+            if filter:
+                for key, value in filter.items():
+                    if isinstance(value, dict) and "$in" in value:
+                        # Handle $in operator: {"field": {"$in": [values]}}
+                        in_values = value["$in"]
+                        if PGVECTOR_PGCRYPTO:
+                            where_clauses.append(
+                                pgcrypto_decrypt(
+                                    DocumentChunk.vmetadata,
+                                    PGVECTOR_PGCRYPTO_KEY,
+                                    JSONB,
+                                )[key].astext.in_([str(v) for v in in_values])
+                            )
+                        else:
+                            where_clauses.append(
+                                DocumentChunk.vmetadata[key].astext.in_(
+                                    [str(v) for v in in_values]
+                                )
+                            )
+                    else:
+                        # Handle simple equality: {"field": "value"}
+                        if PGVECTOR_PGCRYPTO:
+                            where_clauses.append(
+                                pgcrypto_decrypt(
+                                    DocumentChunk.vmetadata,
+                                    PGVECTOR_PGCRYPTO_KEY,
+                                    JSONB,
+                                )[key].astext
+                                == str(value)
+                            )
+                        else:
+                            where_clauses.append(
+                                DocumentChunk.vmetadata[key].astext == str(value)
+                            )
+
             subq = (
                 select(*result_fields)
-                .where(DocumentChunk.collection_name == collection_name)
+                .where(*where_clauses)
                 .order_by(
                     (DocumentChunk.vector.cosine_distance(query_vectors.c.q_vector))
                 )

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

@@ -31,15 +31,12 @@ from open_webui.config import (
     PINECONE_METRIC,
     PINECONE_CLOUD,
 )
-from open_webui.env import SRC_LOG_LEVELS
 from open_webui.retrieval.vector.utils import process_metadata
 
-
 NO_LIMIT = 10000  # Reasonable limit to avoid overwhelming the system
 BATCH_SIZE = 100  # Recommended batch size for Pinecone operations
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class PineconeClient(VectorDBBase):
@@ -393,7 +390,11 @@ class PineconeClient(VectorDBBase):
         )
 
     def search(
-        self, collection_name: str, vectors: List[List[Union[float, int]]], limit: int
+        self,
+        collection_name: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         """Search for similar vectors in a collection."""
         if not vectors or not vectors[0]:

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

@@ -22,12 +22,10 @@ from open_webui.config import (
     QDRANT_TIMEOUT,
     QDRANT_HNSW_M,
 )
-from open_webui.env import SRC_LOG_LEVELS
 
 NO_LIMIT = 999999999
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class QdrantClient(VectorDBBase):
@@ -147,7 +145,11 @@ class QdrantClient(VectorDBBase):
         )
 
     def search(
-        self, collection_name: str, vectors: list[list[float | int]], limit: int
+        self,
+        collection_name: str,
+        vectors: list[list[float | int]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
         if limit is None:

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

@@ -13,7 +13,6 @@ from open_webui.config import (
     QDRANT_TIMEOUT,
     QDRANT_HNSW_M,
 )
-from open_webui.env import SRC_LOG_LEVELS
 from open_webui.retrieval.vector.main import (
     GetResult,
     SearchResult,
@@ -30,7 +29,6 @@ TENANT_ID_FIELD = "tenant_id"
 DEFAULT_DIMENSION = 384
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 def _tenant_filter(tenant_id: str) -> models.FieldCondition:
@@ -256,7 +254,11 @@ class QdrantClient(VectorDBBase):
         )
 
     def search(
-        self, collection_name: str, vectors: List[List[float | int]], limit: int
+        self,
+        collection_name: str,
+        vectors: List[List[float | int]],
+        filter: Optional[Dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         """
         Search for the nearest neighbor items based on the vectors with tenant isolation.

backend/open_webui/retrieval/vector/dbs/s3vector.py

@@ -6,13 +6,11 @@ from open_webui.retrieval.vector.main import (
     SearchResult,
 )
 from open_webui.config import S3_VECTOR_BUCKET_NAME, S3_VECTOR_REGION
-from open_webui.env import SRC_LOG_LEVELS
 from typing import List, Optional, Dict, Any, Union
 import logging
 import boto3
 
 log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 class S3VectorClient(VectorDBBase):
@@ -117,15 +115,16 @@ class S3VectorClient(VectorDBBase):
 
     def has_collection(self, collection_name: str) -> bool:
         """
-        Check if a vector index (collection) exists in the S3 vector bucket.
+        Check if a vector index exists using direct lookup.
+        This avoids pagination issues with list_indexes() and is significantly faster.
         """
-
         try:
-            response = self.client.list_indexes(vectorBucketName=self.bucket_name)
-            indexes = response.get("indexes", [])
-            return any(idx.get("indexName") == collection_name for idx in indexes)
+            self.client.get_index(
+                vectorBucketName=self.bucket_name, indexName=collection_name
+            )
+            return True
         except Exception as e:
-            log.error(f"Error listing indexes: {e}")
+            log.error(f"Error checking if index '{collection_name}' exists: {e}")
             return False
 
     def delete_collection(self, collection_name: str) -> None:
@@ -296,7 +295,11 @@ class S3VectorClient(VectorDBBase):
             raise
 
     def search(
-        self, collection_name: str, vectors: List[List[Union[float, int]]], limit: int
+        self,
+        collection_name: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         """
         Search for similar vectors in a collection using multiple query vectors.

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

@@ -0,0 +1,352 @@
+import weaviate
+import re
+import uuid
+from typing import Any, Dict, List, Optional, Union
+
+from open_webui.retrieval.vector.main import (
+    VectorDBBase,
+    VectorItem,
+    SearchResult,
+    GetResult,
+)
+from open_webui.retrieval.vector.utils import process_metadata
+from open_webui.config import (
+    WEAVIATE_HTTP_HOST,
+    WEAVIATE_GRPC_HOST,
+    WEAVIATE_HTTP_PORT,
+    WEAVIATE_GRPC_PORT,
+    WEAVIATE_API_KEY,
+    WEAVIATE_HTTP_SECURE,
+    WEAVIATE_GRPC_SECURE,
+    WEAVIATE_SKIP_INIT_CHECKS,
+)
+
+
+def _convert_uuids_to_strings(obj: Any) -> Any:
+    """
+    Recursively convert UUID objects to strings in nested data structures.
+
+    This function handles:
+    - UUID objects -> string
+    - Dictionaries with UUID values
+    - Lists/Tuples with UUID values
+    - Nested combinations of the above
+
+    Args:
+        obj: Any object that might contain UUIDs
+
+    Returns:
+        The same object structure with UUIDs converted to strings
+    """
+    if isinstance(obj, uuid.UUID):
+        return str(obj)
+    elif isinstance(obj, dict):
+        return {key: _convert_uuids_to_strings(value) for key, value in obj.items()}
+    elif isinstance(obj, (list, tuple)):
+        return type(obj)(_convert_uuids_to_strings(item) for item in obj)
+    elif isinstance(obj, (str, int, float, bool, type(None))):
+        return obj
+    else:
+        return obj
+
+
+class WeaviateClient(VectorDBBase):
+    def __init__(self):
+        self.url = WEAVIATE_HTTP_HOST
+        try:
+            # Build connection parameters
+            connection_params = {
+                "http_host": WEAVIATE_HTTP_HOST,
+                "http_port": WEAVIATE_HTTP_PORT,
+                "http_secure": WEAVIATE_HTTP_SECURE,
+                "grpc_host": WEAVIATE_GRPC_HOST,
+                "grpc_port": WEAVIATE_GRPC_PORT,
+                "grpc_secure": WEAVIATE_GRPC_SECURE,
+                "skip_init_checks": WEAVIATE_SKIP_INIT_CHECKS,
+            }
+
+            # Only add auth_credentials if WEAVIATE_API_KEY exists and is not empty
+            if WEAVIATE_API_KEY:
+                connection_params["auth_credentials"] = (
+                    weaviate.classes.init.Auth.api_key(WEAVIATE_API_KEY)
+                )
+
+            self.client = weaviate.connect_to_custom(**connection_params)
+            self.client.connect()
+        except Exception as e:
+            raise ConnectionError(f"Failed to connect to Weaviate: {e}") from e
+
+    def _sanitize_collection_name(self, collection_name: str) -> str:
+        """Sanitize collection name to be a valid Weaviate class name."""
+        if not isinstance(collection_name, str) or not collection_name.strip():
+            raise ValueError("Collection name must be a non-empty string")
+
+        # Requirements for a valid Weaviate class name:
+        # The collection name must begin with a capital letter.
+        # The name can only contain letters, numbers, and the underscore (_) character. Spaces are not allowed.
+
+        # Replace hyphens with underscores and keep only alphanumeric characters
+        name = re.sub(r"[^a-zA-Z0-9_]", "", collection_name.replace("-", "_"))
+        name = name.strip("_")
+
+        if not name:
+            raise ValueError(
+                "Could not sanitize collection name to be a valid Weaviate class name"
+            )
+
+        # Ensure it starts with a letter and is capitalized
+        if not name[0].isalpha():
+            name = "C" + name
+
+        return name[0].upper() + name[1:]
+
+    def has_collection(self, collection_name: str) -> bool:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        return self.client.collections.exists(sane_collection_name)
+
+    def delete_collection(self, collection_name: str) -> None:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if self.client.collections.exists(sane_collection_name):
+            self.client.collections.delete(sane_collection_name)
+
+    def _create_collection(self, collection_name: str) -> None:
+        self.client.collections.create(
+            name=collection_name,
+            vector_config=weaviate.classes.config.Configure.Vectors.self_provided(),
+            properties=[
+                weaviate.classes.config.Property(
+                    name="text", data_type=weaviate.classes.config.DataType.TEXT
+                ),
+            ],
+        )
+
+    def insert(self, collection_name: str, items: List[VectorItem]) -> None:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if not self.client.collections.exists(sane_collection_name):
+            self._create_collection(sane_collection_name)
+
+        collection = self.client.collections.get(sane_collection_name)
+
+        with collection.batch.fixed_size(batch_size=100) as batch:
+            for item in items:
+                item_uuid = str(uuid.uuid4()) if not item["id"] else str(item["id"])
+
+                properties = {"text": item["text"]}
+                if item["metadata"]:
+                    clean_metadata = _convert_uuids_to_strings(
+                        process_metadata(item["metadata"])
+                    )
+                    clean_metadata.pop("text", None)
+                    properties.update(clean_metadata)
+
+                batch.add_object(
+                    properties=properties, uuid=item_uuid, vector=item["vector"]
+                )
+
+    def upsert(self, collection_name: str, items: List[VectorItem]) -> None:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if not self.client.collections.exists(sane_collection_name):
+            self._create_collection(sane_collection_name)
+
+        collection = self.client.collections.get(sane_collection_name)
+
+        with collection.batch.fixed_size(batch_size=100) as batch:
+            for item in items:
+                item_uuid = str(item["id"]) if item["id"] else None
+
+                properties = {"text": item["text"]}
+                if item["metadata"]:
+                    clean_metadata = _convert_uuids_to_strings(
+                        process_metadata(item["metadata"])
+                    )
+                    clean_metadata.pop("text", None)
+                    properties.update(clean_metadata)
+
+                batch.add_object(
+                    properties=properties, uuid=item_uuid, vector=item["vector"]
+                )
+
+    def search(
+        self,
+        collection_name: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[dict] = None,
+        limit: int = 10,
+    ) -> Optional[SearchResult]:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if not self.client.collections.exists(sane_collection_name):
+            return None
+
+        collection = self.client.collections.get(sane_collection_name)
+
+        result_ids, result_documents, result_metadatas, result_distances = (
+            [],
+            [],
+            [],
+            [],
+        )
+
+        for vector_embedding in vectors:
+            try:
+                response = collection.query.near_vector(
+                    near_vector=vector_embedding,
+                    limit=limit,
+                    return_metadata=weaviate.classes.query.MetadataQuery(distance=True),
+                )
+
+                ids = [str(obj.uuid) for obj in response.objects]
+                documents = []
+                metadatas = []
+                distances = []
+
+                for obj in response.objects:
+                    properties = dict(obj.properties) if obj.properties else {}
+                    documents.append(properties.pop("text", ""))
+                    metadatas.append(_convert_uuids_to_strings(properties))
+
+                # Weaviate has cosine distance, 2 (worst) -> 0 (best). Re-ordering to 0 -> 1
+                raw_distances = [
+                    (
+                        obj.metadata.distance
+                        if obj.metadata and obj.metadata.distance
+                        else 2.0
+                    )
+                    for obj in response.objects
+                ]
+                distances = [(2 - dist) / 2 for dist in raw_distances]
+
+                result_ids.append(ids)
+                result_documents.append(documents)
+                result_metadatas.append(metadatas)
+                result_distances.append(distances)
+            except Exception:
+                result_ids.append([])
+                result_documents.append([])
+                result_metadatas.append([])
+                result_distances.append([])
+
+        return SearchResult(
+            **{
+                "ids": result_ids,
+                "documents": result_documents,
+                "metadatas": result_metadatas,
+                "distances": result_distances,
+            }
+        )
+
+    def query(
+        self, collection_name: str, filter: Dict, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if not self.client.collections.exists(sane_collection_name):
+            return None
+
+        collection = self.client.collections.get(sane_collection_name)
+
+        weaviate_filter = None
+        if filter:
+            for key, value in filter.items():
+                prop_filter = weaviate.classes.query.Filter.by_property(name=key).equal(
+                    value
+                )
+                weaviate_filter = (
+                    prop_filter
+                    if weaviate_filter is None
+                    else weaviate.classes.query.Filter.all_of(
+                        [weaviate_filter, prop_filter]
+                    )
+                )
+
+        try:
+            response = collection.query.fetch_objects(
+                filters=weaviate_filter, limit=limit
+            )
+
+            ids = [str(obj.uuid) for obj in response.objects]
+            documents = []
+            metadatas = []
+
+            for obj in response.objects:
+                properties = dict(obj.properties) if obj.properties else {}
+                documents.append(properties.pop("text", ""))
+                metadatas.append(_convert_uuids_to_strings(properties))
+
+            return GetResult(
+                **{
+                    "ids": [ids],
+                    "documents": [documents],
+                    "metadatas": [metadatas],
+                }
+            )
+        except Exception:
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if not self.client.collections.exists(sane_collection_name):
+            return None
+
+        collection = self.client.collections.get(sane_collection_name)
+        ids, documents, metadatas = [], [], []
+
+        try:
+            for item in collection.iterator():
+                ids.append(str(item.uuid))
+                properties = dict(item.properties) if item.properties else {}
+                documents.append(properties.pop("text", ""))
+                metadatas.append(_convert_uuids_to_strings(properties))
+
+            if not ids:
+                return None
+
+            return GetResult(
+                **{
+                    "ids": [ids],
+                    "documents": [documents],
+                    "metadatas": [metadatas],
+                }
+            )
+        except Exception:
+            return None
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[List[str]] = None,
+        filter: Optional[Dict] = None,
+    ) -> None:
+        sane_collection_name = self._sanitize_collection_name(collection_name)
+        if not self.client.collections.exists(sane_collection_name):
+            return
+
+        collection = self.client.collections.get(sane_collection_name)
+
+        try:
+            if ids:
+                for item_id in ids:
+                    collection.data.delete_by_id(uuid=item_id)
+            elif filter:
+                weaviate_filter = None
+                for key, value in filter.items():
+                    prop_filter = weaviate.classes.query.Filter.by_property(
+                        name=key
+                    ).equal(value)
+                    weaviate_filter = (
+                        prop_filter
+                        if weaviate_filter is None
+                        else weaviate.classes.query.Filter.all_of(
+                            [weaviate_filter, prop_filter]
+                        )
+                    )
+
+                if weaviate_filter:
+                    collection.data.delete_many(where=weaviate_filter)
+        except Exception:
+            pass
+
+    def reset(self) -> None:
+        try:
+            for collection_name in self.client.collections.list_all().keys():
+                self.client.collections.delete(collection_name)
+        except Exception:
+            pass

backend/open_webui/retrieval/vector/factory.py

@@ -53,6 +53,10 @@ class Vector:
                 from open_webui.retrieval.vector.dbs.pgvector import PgvectorClient
 
                 return PgvectorClient()
+            case VectorType.OPENGAUSS:
+                from open_webui.retrieval.vector.dbs.opengauss import OpenGaussClient
+
+                return OpenGaussClient()
             case VectorType.ELASTICSEARCH:
                 from open_webui.retrieval.vector.dbs.elasticsearch import (
                     ElasticsearchClient,
@@ -67,6 +71,10 @@ class Vector:
                 from open_webui.retrieval.vector.dbs.oracle23ai import Oracle23aiClient
 
                 return Oracle23aiClient()
+            case VectorType.WEAVIATE:
+                from open_webui.retrieval.vector.dbs.weaviate import WeaviateClient
+
+                return WeaviateClient()
             case _:
                 raise ValueError(f"Unsupported vector type: {vector_type}")
 

backend/open_webui/retrieval/vector/main.py

@@ -53,7 +53,11 @@ class VectorDBBase(ABC):
 
     @abstractmethod
     def search(
-        self, collection_name: str, vectors: List[List[Union[float, int]]], limit: int
+        self,
+        collection_name: str,
+        vectors: List[List[Union[float, int]]],
+        filter: Optional[Dict] = None,
+        limit: int = 10,
     ) -> Optional[SearchResult]:
         """Search for similar vectors in a collection."""
         pass

backend/open_webui/retrieval/vector/type.py

@@ -11,3 +11,5 @@ class VectorType(StrEnum):
     PGVECTOR = "pgvector"
     ORACLE23AI = "oracle23ai"
     S3VECTOR = "s3vector"
+    WEAVIATE = "weaviate"
+    OPENGAUSS = "opengauss"

backend/open_webui/retrieval/vector/utils.py

@@ -4,6 +4,7 @@ KEYS_TO_EXCLUDE = ["content", "pages", "tables", "paragraphs", "sections", "figu
 
 
 def filter_metadata(metadata: dict[str, any]) -> dict[str, any]:
+    # Removes large/redundant fields from metadata dict.
     metadata = {
         key: value for key, value in metadata.items() if key not in KEYS_TO_EXCLUDE
     }
@@ -13,16 +14,15 @@ def filter_metadata(metadata: dict[str, any]) -> dict[str, any]:
 def process_metadata(
     metadata: dict[str, any],
 ) -> dict[str, any]:
+    # Removes large fields and converts non-serializable types (datetime, list, dict) to strings.
+    result = {}
     for key, value in metadata.items():
-        # Remove large fields
+        # Skip large fields
         if key in KEYS_TO_EXCLUDE:
-            del metadata[key]
-
+            continue
         # Convert non-serializable fields to strings
-        if (
-            isinstance(value, datetime)
-            or isinstance(value, list)
-            or isinstance(value, dict)
-        ):
-            metadata[key] = str(value)
-    return metadata
+        if isinstance(value, (datetime, list, dict)):
+            result[key] = str(value)
+        else:
+            result[key] = value
+    return result