* fix: a rejected ask_user call ending the turn with no reply
The documented behaviour of the built-in ask_user tool is that a call breaking its rules comes back to the model as an error. Instead the reply stopped there: the error was recorded as the tool result, the model was never asked again, and the user was left with a dead chat and no answer.
The rejection is now handed back like any other failed tool result, so the model sees it and can correct itself within the normal tool-call iteration limit. Any ordinary tool the model emitted in the same turn still runs.
A call rejected for arriving alongside other ask_user calls also left those siblings without a result, which the UI shows as a tool call stuck on "Executing..." forever. Every invalid call now gets its own result. Two ask_user calls on their own also reported the wrong reason, saying the call must be made by itself rather than that only one is allowed per turn.
Fixes#29077
* Keep the original ask_user validation order
Restores the pre-existing check order and the unchanged output id fallback, so this change only alters the return shape needed for staging, and trims a comment that narrated the lines below it.
* Correct the ask_user sibling-call error message
* Shorten the ask_user sibling-call error message
* Drop the untrue sibling-call claim from the ask_user error
The ask_user error text told the user and the model "The others ran.", but that sentence is written into the turn output before any sibling tool call has executed, so it can be plainly false. Under a saved chat with tool approval set to ask, the turn pauses right afterwards and the siblings sit at pending/queued, so the user reads "The others ran" directly above the approval prompt for tools that have not run, and reads it again beside the rejection result if they decline. When the model sends two ask_user calls and nothing else, nothing runs at all and the sentence is emitted twice.
The staging helper cannot see what happens to the sibling calls, so it no longer narrates it. The remaining two sentences hold in every flow: ask_user really is dropped from the executed calls whenever this error is set, and calling it on its own is always the right retry.
On Windows hosts the built-in code interpreter fails immediately with "Failed to fetch dynamically imported module: .../pyodide/pyodide.asm.mjs", and the browser console shows the server answered with a MIME type of "text/plain". Code execution is unusable for those users.
Python's mimetypes module reads the Windows registry after loading its own table, so a stray registry entry silently replaces the correct type for an extension and Starlette then labels the file with it. Browsers enforce strict MIME checking for module scripts and streaming WASM compilation, so the pyodide loader gets refused. The same workaround already existed for .js; this extends it to the two other extensions pyodide ships, and moves it out of the frontend-build branch so the unconditionally mounted /static assets are covered as well.
Fixes#29133
* fix: stop labelling a counted schedule as a one-off
The schedule label treated any rule whose text contained COUNT=1 as a single run, so counts such as 10, 12 and 14 were shown as "Once" together with the date of the first run, on the automations list and on the automation page alike. The label now matches a count of exactly one.
This covers the two places that render the label. The schedule editor reads the count the same way and changes separately. Rules that carry a start date still fall through to the raw rule text, exactly as they already did without a count; that parsing gap changes separately too.
Verified in a browser against the same build without these lines: ten ordinary schedules render identically in both places, and a genuine single-run schedule is still labelled as one.
* fix: stop counted schedules being rewritten as one-offs on save
The schedule editor decides that a rule is a one-off by looking for the text COUNT=1 anywhere in it. A rule that runs ten times carries COUNT=10, which contains that text, so opening such an automation shows it as a single run and saving writes a genuine one-off rule back. One open and save is enough to silently turn a ten run schedule into a one run schedule, with whatever date happened to sit in the rule. The check now requires that no further digit follows, the same test the two schedule labels already use.
The same screens also failed to read counted rules at all. Both label helpers and the editor parser split the stored rule on semicolons after stripping the RRULE prefix, so when the rule carries a DTSTART line the first piece is that whole line and the frequency is never found. The automations list then printed the raw rule text where a human label belongs, and the editor fell back to a plain daily schedule, quietly discarding the weekly or monthly settings on the next save. All three sites now drop the DTSTART part before splitting. They split on whitespace, so the newline form and the space separated form are both handled, matching how the backend already strips it.
The Redis-backed model registry skips its write when the content signature matches what is already stored. That skip has never worked across processes. Two of the values it hashes come out of Python sets, and set iteration order varies with each process's hash seed, so every worker computed a different signature for identical content and every worker rewrote the whole registry on every refresh.
Sorting both makes the signature depend on content alone. Measured on a 120 model registry, 522 KiB serialized: a refresh whose content already matches drops from GET, HKEYS, HSET and SET at 5.1 ms to a single GET at 2.2 ms per worker, and the 522 KiB write leaves the wire entirely.
Verified across 12 child processes with 12 distinct hash seeds: 12 different signatures before, 1 after. Filter execution order is unaffected, because the filter pipeline re-sorts by priority and id before running.
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
A chat could hard-fail in the browser with "TypeError: can't access property content" and stop rendering the assistant message until a reload.
Streamed response items and content parts were placed at the index the provider reports. That index is not bounded by the length of the array the client has built up, so an entry could land past the end and leave a gap behind. Spreading the array on the next event turned that gap into a real empty entry, and the renderer then dereferenced it while looking for message text.
Output items that would land past the end are now appended, since later events locate them by id anyway. Content and summary parts are padded up to the index instead, because a part carries no id and the text streamed for it is addressed by that same index. The renderer and the structured editor now skip an empty entry as well, so chats already saved in the broken state still display and edit.
Fixes#29244
On Redis Cluster deployments the stop button never stopped a running response when the request landed on a different instance than the one streaming it. The pub/sub listener that carries the stop signal between instances never managed to subscribe, so the command was published to a channel nobody was listening on.
The listener subscribes through a cluster client that connects lazily, and redis-py resolves the pub/sub node from a slot cache that is still empty at that point, which fails with a bare KeyError. Awaiting initialize() first fills that cache. It is a no-op on standalone and Sentinel clients, so nothing has to branch on the deployment type, and it stays inside the reconnect loop so a failover refreshes the cache instead of resubscribing against a stale one.
Before 0.11.1 the listener died on that first exception and cross-instance stop never worked at all. The reconnect loop added in 0.11.1 turned it into a startup window plus KeyError retry spam in the logs. Reported upstream as redis/redis-py#4296.
Fixes#19840
Fills the two remaining untranslated strings in the German catalog and corrects a number of existing entries.
The catalog addresses the user formally with "Sie" in over two hundred strings but had drifted to the informal "du" in around thirty, including "Wählen Sie ein Modell" sitting directly alongside "Wähle eine Option". Those now use "Sie", or the infinitive where the surrounding labels already use it. The "Du" chat bubble label and the two model-facing system prompts are deliberately left informal.
"Explored" carried a trailing ": " that the English source and every other translated locale lack. The summary text beside it only renders when there are tool calls or code interpreter runs, so a details group containing only reasoning items rendered a dangling "Untersucht:" in German.
"Persistent" was left in English beside its already translated sibling option "Ephemeral" ("Flüchtig"), leaving a half translated storage dropdown. The calendar tool description promised listing, searching, creating, updating and deleting calendars, when it acts on calendar events.
The remaining changes fix compounds written as two words ("Skill Beschreibung", "Datei upload", "Audio tag"), replace the non-word "managen" with "verwalten", correct a grammatical gender and a plural, and normalize the only two ellipsis characters in the file to the three periods used by every other entry.
Only de-DE is touched. No catalog regeneration, no other locale files.
On a phone the file preview zoom bar shows plus and minus buttons that duplicate pinch-to-zoom and sit on top of an already small preview. They are hidden on coarse pointers now, so touch users get the preview area back and still zoom the way they expect.
The check is the pointer type rather than the viewport width, because what matters is whether the person can pinch, not how narrow their window is. A narrow desktop window keeps the buttons, a tablet does not.
The zoom percentage doubles as the reset control and has no gesture equivalent, so it stays visible, along with all page and slide navigation. Keyboard zoom is unaffected. Word document previews are left alone: they have no pinch support at all, so hiding their buttons would remove zooming entirely.
Fixes#29152
process_pipeline_inlet_filter() and its outlet counterpart construct and tear
down an aiohttp ClientSession, with its own connector and cookie jar, on
every chat completion and every task generation request just to iterate an
empty filter list. On deployments without pipelines, which is the default,
that is wasted setup on every message.
Both functions now return the payload untouched before the session is
created when there is nothing to call. The per-call saving is small, a few
microseconds of object construction per request on the pinned aiohttp; the
point is that requests stop paying setup for a feature that is not
configured.
Clicking a pinned model in the sidebar started a new chat with that model but sent the previous model's tool_ids and skill_ids, and they stayed until the page was reloaded. The model dropdown was unaffected, and so was temporary chat.
A pinned entry links to /?model=<id>, which runs the new-chat path. That path applied the model's defaults and then restored the composer draft over the top, and the draft still held the selection from whichever model was active when it was written. The draft save is debounced, so the stale value was reliably the one read back.
The draft is now restored before the defaults are applied, so the model always decides which tools and skills are active while the unsent prompt, files and approval mode are still kept. Starting a new chat with several models selected now clears the selection instead of carrying the draft's over, since there are no per-model defaults to apply in that case.
Co-authored-by: Claude <noreply@anthropic.com>
173 keys in the de-DE catalog still had empty values, so German users saw those strings rendered in English: the whole terminal file browser, the tool-call approval prompts, chat variables, the model manager, calendar navigation, the accessibility labels for zoom, camera and call controls, and several admin settings panels.
Each string was translated against its actual call site rather than in isolation, so the grammatical form fits the widget it renders in (imperatives on buttons, nouns on labels and select options, participles on toasts). Terminology and the formal "Sie" register follow what the catalog already uses elsewhere, and technical literals were left alone on purpose: the CSV header hint mirrors the file that "Download CSV Template" actually produces, and the MIME pattern, the snake_case variable placeholder and the product names stay verbatim.
Only value strings changed; key order and formatting are untouched.
* fix: stop streaming responses breaking on a duplicate output key
With reasoning-capable models the chat froze mid-stream: the first chunk of the answer appeared, nothing followed, and the whole message only showed up once generation finished. The browser console showed a Svelte each_key_duplicate error.
When a stream event addresses an output slot past the end of the array, the missing slots were filled with the event's own item, id included, so a gap of two left two entries claiming the same id. The next chunk for that item was matched by id, landed in the first of the two, and the rendered list ended up with two items sharing a key, which Svelte refuses to update.
Only the addressed slot now takes the event's item, and the slots before it are anonymous placeholders. Replayed the reported event sequence against the real code: keys are unique again and the chunks stay in order instead of being split across the copies.
* fix: stream reasoning deltas when the provider also sends reasoning_details
Providers such as OpenRouter emit reasoning_details alongside the reasoning
text on the same delta. Merging those details cleared the pending event
unconditionally, discarding the response.reasoning_text.delta that had just
been built, so the client received no reasoning until the response completed
and the thinking block only appeared after generation finished.
The event is now only dropped when the details were all there was to report.
Details persistence is unchanged.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014uuEg4AXPs9zE3vVUfN1Fj
---------
Co-authored-by: Claude <noreply@anthropic.com>
After a tool call, the model's thinking was streamed into the chat as if it were the main response, and only jumped into the collapsed Thoughts section once the turn finished. Every further tool call repeated it.
Each tool round appended an empty placeholder message item to the output and sent it to the browser, then dropped it again from the copy used to offset the next round's item indices. The browser therefore held one item more than the backend counted, so the first thinking chunk of the next round was written into that leftover message item and rendered as normal text until the finished output replaced it.
The placeholder is removed. It was never needed: a message item is already created when actual content arrives, and dropping it also stops an empty assistant message being sent back to the model on the follow-up request.
Co-authored-by: Claude <noreply@anthropic.com>
Any `<$...>` run in a chat message was treated as an inline skill mention and removed before the request reached the model, so text like `<$(=MonthStart($(vMaxMonthEndINC)))"}, [Registration day] >` silently vanished mid-message and the model only saw the part before it.
The mention regexes accepted any character except `|` and `>` as the skill id, so they matched far more than real mentions. Skill ids are already validated as `[a-z0-9_-]+` when a skill is created, so both regexes now require that charset. Ordinary text passes through untouched while `<$id>`, `<$id|Label>` and `</id|Label>` still resolve and strip as before.
Verified against the reported message (now preserved verbatim) and the three mention forms.
Co-authored-by: Claude <noreply@anthropic.com>
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
Every chat request hands each builtin tool a fresh copy of its cached spec, because callers mutate what they get. That copy was a full deepcopy of a nested dict, repeated per tool per message.
The builder now caches the spec already serialized, so a request only parses it back. Parsing is what produces the independent tree callers mutate, and the cached value becomes an immutable string, so a request can no longer reach the cached object at all.
Measured on CPython 3.12 with a 1.1 KB spec and 20 builtin tools per request:
| | before | after |
|---|---|---|
| stdlib json, the default | 276.2 us | 66.4 us |
| orjson | 279.5 us | 37.5 us |
Builtin specs are plain JSON by construction: pydantic normalizes every default before it reaches the schema, so a tuple, set, enum or datetime cannot appear in one, and an unserializable default is dropped rather than embedded.
"RAG_METADATA_MAX_VALUE_CHARS" ships unset, and unset means no bound at all, so the limit only protects the deployments that already knew to configure it. A small Office document is a zip archive, and one crafted to expand enormously during extraction can turn a few hundred kilobytes into gigabytes of metadata held in memory; uploading it a handful of times is enough to exhaust a server and take Open WebUI down with it.
When no explicit limit is configured, the bound now follows "RAG_FILE_MAX_SIZE" instead of being absent, on the reasoning that a document cannot legitimately carry more metadata than the file itself is allowed to be. That keeps the number from being an arbitrary guess: it is whatever the administrator already decided an upload may weigh. Setting "RAG_METADATA_MAX_VALUE_CHARS" explicitly still wins, and a deployment that leaves both unset is unchanged, which is the same posture the upload limit itself takes.
"RAG_FILE_MAX_SIZE" is in MB and is treated as unset when it is zero, matching how the document loader already reads it.
Every streamed delta saves a snapshot of the in-progress response so a reconnecting client can resume it, and each save rebuilt the assistant text from scratch. On the Chat Completions path that re-joined every accumulated chunk, including on saves carrying no new text, so a long answer followed by a large tool call re-joined the whole answer once per argument chunk. The Responses API path never collects those chunks and reads the text back out of the output items instead, where the blank check copied it in full every time.
The joined string is now kept and reused until another chunk arrives, since content_parts is only ever appended to; the nonlocal declaration that suggested otherwise was already dead and is dropped, and inlining the single-use helper removes an unreachable branch with it. The blank check in get_output_text now tests the text rather than allocating a stripped copy of it, which is equivalent for all twelve of its callers. Text streaming on the Chat Completions path is unchanged, since a text delta always appends before it saves.
| stream | before | after |
| --- | --- | --- |
| 20k-char answer, 2000 tool-argument chunks | 21.4 ms | 0.06 ms |
| Responses API, 40k deltas, 200k chars | 80.7 ms | 50.5 ms |
Without Redis nothing extra is retained, since the snapshot store already held that string; with Redis one copy of the response text stays alive while the stream runs.
Tag detection for the code interpreter ran regardless of the tool-calling mode, so a model in Native (Agentic) Mode that emitted <code_interpreter> blocks in ordinary reply text had that code sent to the executor. Native mode never teaches the tag format and exposes execute_code as a builtin tool, so the parser had nothing legitimate to pick up there.
Gates detection on the legacy mode, matching the condition that already decides whether the tag prompt is injected at all. The five authorization checks are unchanged, and native mode keeps executing through the tool.
Deployments on native mode whose models emit the tags unprompted will now see them rendered as text.
Some providers send one very large piece of a streamed answer in a single go: a long reasoning trace, a code execution result, a turn with many tool calls, or a response echo carrying a big tool list. Anything past 128 KB in one line killed the chat mid-answer with a misleading `400, message: Got more than 131072 bytes when reading`. Nothing was rejected upstream, that is our own reader giving up on an oversized line.
Open WebUI already had code that assembles lines itself with no such limit, but it only ran when CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE was set. Unset is the default, and in that case the raw capped reader was used instead, so a default install always broke. That path now always assembles lines, and the setting goes back to being what its name says: an optional cap, off by default. It applies to the Ollama stream as well, since both now share the same reader.
The assembly loop only splits once a line actually completes, because the old one re-concatenated and re-split the whole buffer on every network chunk. Without that, allowing long lines would have traded an error for multi-second event loop stalls.
| | 20 MB in one line | 200k small lines |
| --- | --- | --- |
| before | 4249 ms | 27.3 ms |
| after | 37 ms | 25.2 ms |
* fix: apply the SSRF checks to redirect targets on every web fetch path
Two guards protect server-side fetches: a private-IP check and the operator's `WEB_FETCH_FILTER_LIST`. Neither reached a redirect hop on the aiohttp paths, and the filter list never reached one on the requests paths either.
aiohttp answers IP-literal hosts itself without consulting a resolver, so `_SSRFSafeResolver` was never invoked for a hop such as `http://169.254.169.254/` and the private-IP check simply did not run. With redirect following enabled, a submitted public URL that redirects to an IP literal reached loopback, RFC1918 and cloud-metadata addresses, and the response body was returned to the caller. The filter list was consulted only in `validate_url`, on the originally submitted URL, so a redirect to a filter-listed host was fetched without it ever being applied.
`_SSRFSafeResolver` is replaced by `_SSRFSafeConnector`, which hooks `_resolve_host` so the IP check also covers the IP-literal shortcut and both DNS cache paths. The filter list moves to a per-request hook on each transport, `connect()` for aiohttp and `send()` for the requests adapter, because those see the request destination: at the connection layer a proxied request presents the proxy's host, and a pooled connection skips resolution entirely. This covers every hop, including redirects, on all five aiohttp call sites and both requests sessions. The Playwright loader already validated each hop and is unchanged.
Both gaps required `AIOHTTP_CLIENT_ALLOW_REDIRECTS=true`, which is not the default.
Two behaviour changes for operators. The filter list now applies to redirect targets rather than only to submitted URLs. Under a forward proxy it is evaluated against the request destination instead of the proxy, which also fixes allowlist entries rejecting every fetch in proxied deployments.
* refac: match the web fetch filter list against resolved addresses
The filter list is now evaluated against the hostname together with the addresses it resolves to, at URL validation and on each connection, on both transports. An IPv6 address is also matched by the IPv4 address it carries.
* refac: screen outbound fetch addresses against reserved ranges ipaddress misses
`ipaddress.is_global` was the only test behind the web-fetch address check, and it answers a narrower question than "may we fetch this". Several special-purpose ranges are globally routable by registry while nothing on them is a legitimate destination, so they passed. Classification now screens those ranges on top of `is_global`, and applies the same screen to the IPv4 address embedded in an IPv6 transition encoding rather than only to the literal. All three checkpoints share the predicate, so they all inherit it.
The range list is the exact complement of what CPython's `ipaddress` already models, checked entry by entry against both IANA special-purpose registries. Prefixes IANA marks globally reachable are deliberately left out, so no real destination changes behaviour. Verified against 31 addresses covering every entry, their transition-encoded forms, and public controls in both families: 31/31 expected after, 18/31 before.
* refac: match web fetch filter entries that name an address or a range
A filter entry that parses as an address or a CIDR range is matched by containment rather than by DNS label suffix, so a range covers the addresses inside it and an address matches however it is spelled. A range entry previously matched nothing at all, silently.
The built-in list gains the special-purpose networks that ipaddress.is_global reports as reachable while nothing on them is a legitimate destination, so taking an address out of reach is a WEB_FETCH_FILTER_LIST change rather than a release. Those entries hold whether or not local web fetch is enabled; the private-address rule still follows the toggle.
Deleting an external knowledge base now clears its connection only when an admin removes the last knowledge base referencing it, matching the connection delete route.
The binary branch of the web fetch read the entire response body into memory
before writing it out. It now streams in blocks, applies the configured file
size limit the same way the sibling URL endpoint already does, and removes the
temporary file when a download fails partway instead of leaving it behind.
stop_item_tasks() ran unconditionally while create_task() only ran when the
update carried data, so an update without a content snapshot cancelled the
pending save without scheduling a replacement and the edits were never
written.
With delta streaming most websocket frames are tiny per-token deltas, and
per-message-deflate pays zlib work on every outgoing frame per subscriber for
near-zero gain there; under heavy streaming that shows up as measurable server
CPU. The frames that still benefit are the rare large ones (final message,
sources), and even a 100k token message is only a few hundred KB uncompressed,
which any network delivers without noticeable delay.
UVICORN_WS_PER_MESSAGE_DEFLATE=false (default true, current behavior) disables
the extension in every entry point: open-webui serve and dev, start.sh both
invocations, start_windows.bat and dev.sh. Verified against a running
instance: with the flag off the server declines the client-offered
permessage-deflate extension, with defaults it still negotiates it.
Presence tracking writes each user's last_active_at on every authenticated request, every API key request and every websocket heartbeat. The throttle for it already exists but ships unset, and unset means no throttle at all, so a stock deployment pays one UPDATE plus COMMIT per user per request. The 30 second frontend heartbeat alone is 2 write transactions per minute per open tab, before any actual UI traffic.
Defaulting the throttle to 60 seconds collapses that to at most one write per user per worker per minute. Presence is only ever read at minute granularity, so nothing visible changes.
60 rather than the 300 to 500 the docs currently suggest, because a user counts as active for 3 minutes after their last write and that window is hardcoded in the backend and again in the frontend. Any interval at or above 180 seconds makes people who are actively using the instance drop out of the active user count. Letting the window follow the interval instead would need the value shipped to the client, so that is a separate change.
0 still disables the throttle, and now costs nothing at all: the decorator returns the undecorated function instead of a wrapper that re-checks a constant on every call.
Closes#28165
nltk.download picks the first entry of nltk.data.path that already exists and is
writable. None do here, so punkt_tab lands in /root/nltk_data, and /root is mode
0700. The corpus is then unreachable whenever the container does not run as
root:
nltk.data.find('tokenizers/punkt_tab')
LookupError: Resource punkt_tab not found.
/usr/local/share/nltk_data is already on nltk.data.path, so nothing changes at
the read side.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Moving a folder under one of its own subfolders was accepted. A folder in a parent loop is never a root, so it and everything under it silently disappeared from the sidebar, and there was no way to get it back from the UI.
The move is now rejected with a 400, folders whose parent chain loops are put back at the root on the next folder list, and the folder tree traversals skip ids they have already visited so existing data in that state stays workable.
* fix: stop unsaved group settings from persisting into the group list state
* fix: key the groups list so an open modal cannot rebind to another group
With `DATABASE_ENABLE_IAM_TOKEN_AUTH=true` and `VECTOR_DB=pgvector`, startup failed at vector store initialisation with `fe_sendauth: no password supplied`, so the two features could not be used together.
`PgvectorClient` builds its own engine and never got the `do_connect` listener that refreshes the RDS IAM token, and the `ScopedSession` branch that would have reused the instrumented main engine is unreachable because `PGVECTOR_DB_URL` defaults to `DATABASE_URL` and is therefore never falsy.
The pgvector engine now goes through `enable_iam_token_auth()` like the main and Alembic engines. Since a token authenticates exactly one host/port/user, that function now attaches the listener only to engines pointing at the same target, so a `PGVECTOR_DB_URL` aimed at a separate database keeps the password from its own URL instead of having it overwritten; the skip is logged with both identities.
Fixes#27752
* fix: use the pooled client timeout for the Anthropic Messages passthrough
The native `/api/v1/messages` passthrough still referenced `openai.AIOHTTP_CLIENT_TIMEOUT`, which stopped existing when `routers/openai.py` moved onto `session_pool.get_client_timeout()`. Every passthrough request therefore raised `AttributeError: module 'open_webui.routers.openai' has no attribute 'AIOHTTP_CLIENT_TIMEOUT'` before it was sent, and the surrounding handler turned that into a 502 "Open WebUI: Server Connection Error", so Anthropic-format clients such as Cline could not reach any model at all.
Use `get_client_timeout(stream=...)` like the OpenAI and Ollama proxies do, so the configured `AIOHTTP_CLIENT_TIMEOUT` applies and streaming requests additionally get the idle-read timeout.
Fixes#27595
* fix: authenticate native Anthropic requests with x-api-key
The Anthropic Messages passthrough and the token-count forwarding both build their upstream request through `get_anthropic_request_target`, which sends the connection key as `Authorization: Bearer <key>`. Anthropic's OpenAI-compatible `/chat/completions` endpoint accepts that, which is why the model works in the chat UI, but the native `/v1/messages` and `/v1/messages/count_tokens` endpoints do not: they require the key in `x-api-key` and reject a bearer token with 401 `Invalid bearer token` (and `jwt auth is not yet supported on count_tokens`). They also require an `anthropic-version` header, which was never sent.
For `api.anthropic.com` connections, send `anthropic-version` and move the key into `x-api-key`, dropping the bearer header. Connections using session, OAuth or Entra ID auth keep their token untouched, LiteLLM passthrough connections are unaffected, and admin-configured custom headers still win over both defaults.
Fixes#27695
Streamed responses are scanned for reasoning and code interpreter tags. To work out where the last complete tag ended, the scanner searched backwards from the start of the accumulated text on every chunk, once per tag set. Ordinary prose contains no angle bracket, so that search never stopped early and read the entire response back every time. The cost grows with the square of the response length, and this scanning is on unless a model turns it off.
The two positions are now carried forward as the text grows, so each chunk only scans the characters it added.
Measured on CPython 3.12, a 270 KB response streamed in 27000 chunks:
| response text | before | after |
|---|---|---|
| no newlines | 7690 ms | 40.6 ms |
| with newlines | 5695 ms | 41.7 ms |
The carried positions match a full rescan at every step of 36282 randomized replays, covering text with no markers, newlines only, dense markers, real tags and truncation part way through.
The timer scheduler polls once a second and cancels on every message send and chat open, the sidebar lists chats ordered by `updated_at`, and the folder badges count unread chats per folder. None of those could be served by an index, so each call read most of the `chat` table, and because `meta` sits after the chat payload column SQLite had to walk every row's overflow pages to get there. On a large history that stalls the sidebar, every chat switch and every send, and the idle poll alone burns about a quarter of a CPU core.
Timers now keep their due time in a dedicated `chat.timer_at` column behind a partial index, and the chat list, unread and unfinished-reply queries each get an index matching their filter and ordering. Existing pending timers are backfilled from their meta by the migration. Dropping the `internal` and `type` checks also makes a forked timer chat inert, where a fork used to copy `meta` verbatim and become a second claim target that could fire a duplicate timer.
Measured on SQLite, same rows returned:
| query | before | after |
|---|---|---|
| idle timer poll (2000 chats, 0.43 GB) | 170 ms | 0.04 ms |
| cancel on send and chat open (4000 chats, 377 MB) | 200 ms | 0.04 ms |
| sidebar chat list (15000 chats, 1.26 GB) | 157 ms | 1.8 ms |
| folder unread badges (15000 chats, 1.4 GB) | 54 ms | 0.2 ms |
PostgreSQL 17 serves all of them as index-only scans with no sort node. Exercised through fresh install, upgrade with seeded data, downgrade and re-upgrade on SQLite and PostgreSQL 17.
Fixes#27622
75 entries in the French locale had an empty value. i18next falls back to the
key when a value is empty, so French users were shown raw English strings
across permission settings, empty states, form placeholders and error toasts.
- fill those 75 entries; no key is added, removed or reordered
- follow the conventions already present in the file: "Entrez ..." for input
hints, "Échec de ..." for failures, Chat -> Conversation, Token left as is
- interpolation placeholders preserved; no existing translation modified
`x in d` and `x in d.keys()` are identical for a plain dict, so the `.keys()` call builds a throwaway view and reads as if it were doing something. Both sites operate on a plain dict: `combined` in `merge_and_sort_query_results` is a local `dict()`, and `ui_settings` comes from `UserSettings.model_dump()` where `ui` is annotated `dict | None` and is already guarded against None on the preceding line.
No behaviour change, and no measurable speedup either, so this is a readability cleanup rather than a performance one.
Sites where `.keys()` is load-bearing are left alone: the `list(d.keys())` snapshots taken before mutating during iteration, and the places where `.keys()` is the iteration or comprehension source rather than a membership test.
A stream filter function, or a provider that puts something other than a string in a delta, makes the streaming handler concatenate a string with a non-string. That raises TypeError, and the broad handler wrapped around the whole per-chunk block swallows it at debug level and moves on. The chunk's text never reaches the message the user sees, and nothing above debug level says why.
The content and reasoning fields are now coerced to text once, where they are read off the delta, ahead of every consumer. The coercion is guarded on truthiness, so falsy values such as an empty list still skip the block exactly as before, and the accumulated content receives byte for byte what it received previously.
Checked against 14 delta shapes covering strings, empty values, numbers, booleans, None, lists, dicts and a content array: the truthiness gate and the accumulated content are identical before and after.
The ComfyUI workflow.json Edit drawer (CodeEditorModal -> Drawer) used
z-999, while the admin Settings dialog (Modal) uses z-9999. Since both
are appended to <body>, the drawer rendered behind the settings dialog
and appeared to open 'in the background' (#27647).
Add an optional zIndexClass prop to Drawer (default z-999, preserving
existing behaviour for all other callers) and pass z-99999 from
CodeEditorModal so the editor surfaces above any enclosing modal.
Task lists in Notes serialized to markdown as `- [ ] [ ]` with the item text pushed onto a separate line after a blank line, so previewing or downloading a note produced a broken checklist, and checking an item left the second `[ ]` behind as plain text.
TipTap renders each task item as a checkbox inside a label plus a block-wrapped body. The GFM turndown plugin matches that checkbox and emits its own `[ ]`, which landed next to the marker the task item rule already writes, and the block wrapper left blank lines around the text that the old leading-whitespace strip could not remove.
Register a rule that drops the checkbox so the task item rule is the only source of the marker, and trim the block wrapper while indenting continuation lines so nested lists and code fences stay inside the item.
Fixes#26067
The web search error message was a lambda with a passthrough branch that returned whatever it was handed. Since #28942 both call sites pass no arguments, so that branch is unreachable, and it is the trap that let a caller drop a raw exception object into an HTTP response body and turn an intended 400 into an unserialisable 500.
A plain string constant removes the trap and lines the message up with every other fixed message in that file. Behaviour is unchanged: the response detail comes out byte for byte identical, because the enum already overrides __str__ to render members as their value. Verified on Python 3.11 and 3.12, both producing the same string and the same JSON body.
Any failure during a web search comes back to the client as a bare HTTP 500 with nothing in it. The handler tries to build a 400 whose detail is the caught exception object itself, FastAPI cannot serialise that into a response body, so rendering the error response fails and the request falls through to the generic 500 handler. In chat this surfaces as a web search that fails with no explanation at all, and the most common trigger is simply selecting a search engine without configuring its API key.
This routes the failure through the standard error formatter, which is what the sibling handler for content loading failures in the same function already does. Web search failures now return 400 with a readable message, and the exception itself keeps going to the server log exactly as before.
Passing str(e) into the response was the other option and was rejected: the rest of the backend deliberately keeps provider exception text out of client responses and in the log, and provider exceptions here can carry request details that should not be echoed back.
The DuckDuckGo search path catches RatelimitException from the ddgs library. That exception is defined by the library but never raised anywhere in it, checked against the pinned 9.14.4 and against 9.11.3, so the handler could never run. The two fallbacks around it were dead for the same reason: ddgs.text() returns a non-empty list or raises, so None and an empty list are not outcomes it can produce.
Removing all three leaves one call and changes nothing observable. A refused or rate limited search already came out as a failed search, with the error shown to the user and the traceback in the log, and it still does.
The backend argument is now passed as backend or 'auto' rather than conditionally omitted, because 'auto' is the library's own default for that parameter, so every configured value including unset and empty resolves exactly as before. Verified by running the old and the new function side by side against a stubbed library covering normal results, the domain filter, all four backend settings and a failing search, with identical results in every case.
Two components under the admin settings area are not imported by any route or component. The admin settings shell went dead when the admin settings route became a redirect into the settings modal, which imports every admin tab directly and carries its own tab list and search. The model selector beside it lost its last importer in a separate models refactor. Every child component the shell used is still imported by the settings modal, so nothing goes with them.
This removes around 590 lines that still turn up in every search across the admin area.
The Mistral OCR loader has a full async pipeline beside its synchronous one: an async load, its own upload, signed URL, OCR, delete and retry helpers, a pooled session and a batch loader on top. The only way in was the batch loader, which nothing calls, so the entire async half was unreachable. Everything that loads documents goes through the synchronous path, and the shared loader entry point runs it in a worker thread. The Datalab loader carries a public request status poller with no caller either, since its own load inlines the polling it needs.
With the async half gone, the retry classifier's two aiohttp branches can no longer be reached, since the only retried calls are synchronous, so those go with it along with the aiohttp import that existed solely to feed them, and a timeout attribute that nothing reads any more. The class docstring loses the three bullets that only described the removed pipeline, and four docstrings stop calling themselves the sync version of something that no longer has an async counterpart.
This removes around 350 lines and leaves one code path per loader instead of one live path and one that cannot be entered.
With "show emoji in call" enabled, voice mode stayed completely silent and no request ever reached the configured TTS server. Reasoning models served with a reasoning parser return `message.content` as null and put the text in `reasoning_content`, and the emoji helper called `.replace()` on that null value and threw.
The call overlay ran the emoji request first, inside the same `try` block as speech synthesis, so that error skipped the entire TTS section. The audio cache was never filled, and the playback loop kept re-queueing the same content every 200 ms without ever playing it. Read aloud was unaffected because it synthesizes speech directly, which is why the failure looked specific to voice mode.
Fixed on both sides: the optional chain in `generateEmoji` now covers `content`, and the emoji request in the call overlay gets its own catch, matching the speech synthesis call directly below it. An emoji failure now costs the emoji instead of the whole reply.
With WEBSOCKET_MANAGER=redis on a multi-node deployment, the usage pool cleanup task could stop permanently for the whole cluster. Nodes that lost the startup lock race gave up for good after three attempts, and the winner died on a single failed renew or on any Redis connection error, releasing the lock with nobody left to take it over. From then on expired entries accumulated in the usage pool until a node restarted, so /api/usage over-reported models in use and every disconnect handler walked an ever-growing pool.
The task now retries lock acquisition forever like the session pool cleanup does, and any error is logged and answered by releasing the lock and returning to acquisition, so a transient failure costs one cleanup cycle and every node stays a takeover candidate. The delete of an emptied model entry is KeyError-guarded because a disconnect handler on another node can remove the same key between the sweep's snapshot and its delete; unguarded, that race was a permanent task killer that needed nothing rarer than a chat finishing while its tab closed.
Saving a streaming response serialized the payload with orjson, decoded it to
str, scanned it for the three Unicode line separators and let redis-py encode
it straight back to UTF-8: on an 8 MB non-ASCII chat that is 6.9 ms and ~22 MB
of transient buffers per write, synchronously on the event loop.
json_codec now exposes dumps_bytes, which returns the serialized payload as
UTF-8 bytes without the line-separator escaping, and the two Redis writes in
tasks.py use it. That escaping only protects line-framed protocols such as
SSE; every reader of these Redis values re-parses them before anything is
served, and the escaped and raw forms parse identically, so mixed versions
during a rolling deploy interoperate both ways. The same write drops to
0.9 ms and one 8 MB buffer (7.5x), with 31-66% saved on KB-sized writes.
With ENABLE_ORJSON off, dumps_bytes wraps stdlib json, behaviour unchanged.
The str path keeps the escaping but applies it with chained str.replace
instead of a translate table, cutting a separator-containing 8 MB payload
from 312 ms to 5.7 ms with byte-identical output.
Exporting workspace models loaded every model row, built a full response object with its owner for each, and only then dropped the ones the caller may not see. On a large model table that made the export endpoint slow in proportion to models the user cannot even access.
The owner-or-grant check now happens in the query itself, reusing the permission filter this file already applies to the paginated list endpoint, so only visible rows are ever hydrated. The by-user wrapper had one caller left and is gone with it.
Measured with 500 workspace models of which 3 are visible to the caller: 5 queries and ~12.7 ms before, 4 queries and ~2.8 ms after. The resulting set is unchanged for owner, public, direct-user, group and multi-grant entries, and base model entries stay excluded as before.
Every streamed event that persists to a chat (status updates, citations,
file attachments, message content) serialized the entire conversation JSON
three times: a null-byte check of the stored row, a second sanitize of the
whole blob after merging in the event payload and the flush of the UPDATE
itself. The middle pass rescans megabytes of already-clean history for null
bytes that can only come from the small incoming payload, so long chats pay
for their full history on every single event.
The write paths now sanitize just the incoming message, message id and
status dict and keep the row-level sanitize, so legacy rows with null bytes
still self-heal as before. Median per-event write time (sqlite, orjson):
1 MB chat 15.0 ms to 11.1 ms, 4 MB 65.2 ms to 52.4 ms, 10 MB 159.9 ms to
124.5 ms, roughly 20 percent less per event. As a side effect the
chat_message dual write now receives the sanitized message; previously null
bytes in non-content fields were cleaned in the blob but written raw to
chat_message, which failed that insert on PostgreSQL. Verified byte-identical
rows against the previous implementation across nine scenarios covering null
bytes in every input, legacy dirty rows, a missing title and a NULL chat
column.
Listing a user's folders re-checks which entries they may still see, and it resolved their group membership again for every folder, then again inside the collection and note branches for every entry. A comment in that helper claims one membership fetch for the whole listing, but the caller invokes it once per folder, so the claim never held.
The listing now resolves membership once, and only when some folder actually carries entries, then threads it through the file, collection and note checks. Callers that do not supply it are unchanged and still resolve for themselves.
Measured with twenty folders holding six files, two knowledge bases and two notes each: 245 queries and ~145 ms before, 186 and ~117 ms after. The folders returned, and the entries the integrity pass writes back, are unchanged. That was checked against entries the caller owns, entries shared through a group, entries shared with nobody, another user's files, and an unrecognised entry type.
The tool callable now takes its connection's cookie jar as a parameter, matching how its headers are already passed and how the terminal tool factory in the same module builds its callables.
A calendar event's `meta` is a free-form dict, so `meta.alert_minutes` can hold any JSON type, while the upcoming-events lookup assumed it was a number and compared it directly. It now ignores a value that is not numeric and falls back to the default alert window for that event.
Handled on the read side rather than on the write path so events already stored with a non-numeric value are covered too. Numeric values are untouched, including the negative "no alert" sentinel.
A chat attached via the "+" menu or dropped from the sidebar references an
existing chat by id and carries no url. add_file_context() filtered on
`file.get('url')`, so the reference was dropped from <attached_files>
entirely and the model was never told it existed.
When the RAG file-context path is enabled the chat content still reaches
the model as <source> context, which masked this. With file_context
disabled that path is skipped, and get_attached_knowledge() only promotes
collection/note items into <attached_knowledge> - so an attached chat was
visible in the UI but invisible to the model, which then reported having
no chat attachments despite having a view_chat tool available.
Keep chat references and emit their id so the model can resolve them with
view_chat. The url attribute is now conditional, since a chat has none;
the id guard it replaces was dead once the filter guarantees a url or a
chat id.
Co-authored-by: Claude <noreply@anthropic.com>
The tool-call continuation re-submits with bypass_system_prompt=True, but only
routers/openai.py and routers/ollama.py checked it, so pipe and manifold models
had the system prompt applied again on every continuation. Since
add_or_update_system_message() prepends rather than replaces, N tool-call rounds
left N+1 copies of the system prompt in the payload.
Checking whether a user may reach a file loaded and validated every workspace model that user can access, then scanned each model's knowledge list in Python for one file id. Folder listings run that check once per file, so opening a folder of twenty files rebuilt the whole accessible-model set twenty times, and the same check sits on every retrieval and download path.
The lookup now runs the other way round: the database returns the models that attach the file, and only those are access-checked. The text match on the metadata column is a prefilter and the knowledge entries still decide, so a file id that merely appears in a description grants nothing; file ids are server-generated uuids, so the match can only be too wide, never too narrow.
Measured with 500 accessible workspace models: a single check drops from 9 queries and ~20 ms to 6 and ~2.6 ms, and a twenty-file folder listing from 180 queries and ~680 ms to 120 and ~56 ms. A 72-case matrix over owner, public, direct-user and group grants, for both read and write, returns exactly what it returned before, and write still requires the model owner to own the file. The check also no longer writes to the database while answering a read-only question.
Listing skills ran one database query per skill in the instance. A non-admin opening the list on a workspace with 500 skills issued over 500 queries, the paginated list re-resolved the caller's group membership once per row, and every chat message carrying a skill loaded every skill the user can read, full body and owner included, to use the two or three it actually referenced.
Skills now arrive already filtered: the owner-or-grant check runs in the query as an EXISTS subquery, the same way prompts and the search endpoints already do it, the per-item write flag uses the existing batch grant lookup, and the chat path asks only for the skill ids the request names.
Measured with 500 skills of which 3 are visible to the caller: 504 queries and ~300 ms before, 4 queries and ~2.6 ms after. The resulting set is unchanged for owner, public, direct-user, group and multi-grant entries, for both read and write.
Three methods on KnowledgeTable have no callers anywhere in the repository. get_knowledge_bases_by_user_id loaded every knowledge base and filtered them in Python, which search_knowledge_bases already does in SQL with pagination. get_knowledge_by_id_and_user_id duplicates check_access_by_user_id with the permission hardcoded to write. update_knowledge_data_by_id writes a data column that a migration dropped, so it could only ever raise and return None through its own except block.
What remains is one per-entry access helper and one SQL-filtered list path, so nobody reaches for the slower or the broken variant by accident.
No behaviour change.
Assembling the builtin tools for a chat message fetched the chat row a second time to answer one question: whether this is a note chat. The caller had loaded that same row a few lines earlier, from the same id in the same metadata dict, and had already evaluated the same predicate for its own note handling. So every message with builtin tools enabled read the whole conversation blob twice.
The caller now works the flag out once and passes it down. Tool assembly no longer touches a chat model at all, so the two files cannot drift apart when the shape of that metadata changes.
Measured with a stub request across five chat shapes, a note chat, a plain chat, an internal chat that is not a note, a chat id with no row behind it, and an unsaved chat id: the returned tool set is identical in every case and the query count drops from six to five. The note tools are still enabled for a note chat with the notes feature switched off, which is the only thing that predicate decides.
Opening the shared folder list fetched every shared folder in its own query, fetched a chunk of them a second time to walk their children, and looked up each distinct owner separately. With forty folders shared with a user that is over a hundred queries before any subtree work starts.
The folders and their owners now come back in one query each, and the inheritance pass reuses the rows already in hand. Both folder listings also gained an explicit order: the sidebar merges shared subfolders in response order without sorting them, and neither query had an ORDER BY, so on Postgres a folder rename could reshuffle its siblings.
Measured with forty shared folders and no subtrees: 181 queries and ~105 ms before, 92 and ~66 ms after. With subtrees attached, 203 folders in total, it is 341 queries before against 252 after; the remainder is the recursive child walk, which this change deliberately leaves alone. The returned set, permissions and owner names are unchanged, including for a grant pointing at a deleted folder row, a folder the caller owns that is also shared with them, a folder whose owner record is gone, and a child folder that is itself directly shared.
Saving a chat rewrote its message rows one at a time. Each message took its own session out of the pool and committed on its own, and the save endpoint hands over the entire merged history rather than only what changed, so a two hundred message chat cost two hundred sessions and two hundred commits on every save.
The messages now go through a single select and a single commit. The field mapping for the insert and the update branch moved into two small helpers, so the batch and the single-message path cannot drift apart.
Measured on a two hundred message chat with one message edited: 201 queries and 200 transactions before, 2 queries and 1 transaction after, ~149 ms against ~6 ms. Re-saving an unchanged history now costs one select and no writes at all.
One behaviour change worth stating: a message the database cannot store used to be skipped on its own, and now costs the rest of that same save. This table is a rebuildable fast path, so the reader falls back to the history on the chat row and re-triggers the backfill, and the next save reconciles everything still present. A per-message retry was tried and dropped, because a commit that lands but still raises would re-apply the usage merge and double the recorded token counts.
Changing a password left every other logged-in device working until the JWT expired on its own, up to four weeks with the default settings. The hardening docs already promise the opposite: with Redis configured a password change is supposed to put the user's tokens on the revocation list, but only sign-out and OIDC back-channel logout ever wrote to it.
Both password-change paths, self-service and an admin resetting someone's password, now stamp the per-user revocation marker that token validation already checks, so every session issued before the change stops working. The acting device is signed out as well and asked to sign in again, which is the safer default when the password is being changed precisely because the old one may be compromised. Without Redis nothing can be revoked, as before, and the backend now logs a warning saying so.
The marker is written through one shared helper, so its lifetime follows the configured JWT lifetime instead of a fixed 30 days and never expires at all when JWT_EXPIRES_IN disables expiry. Back-channel logout picks that up too, where a long or disabled JWT lifetime previously let the marker expire while the tokens it revoked were still valid. API keys keep working, they are separate credentials with their own lifecycle.
Discussed in #28647.
The Python test suite was deleted in 4527c747b but its dependencies stayed behind, so pytest, pytest-docker and the docker SDK still install into every image variant, and moto joins them for anyone running pip install open-webui[all]. No Python test file remains in the repository, nothing imports these packages, and no CI job runs pytest. They are removed from backend/requirements.txt and from the all extra, which are the only two channels they ship through.
netcat-openbsd goes for the same reason. It was added in January 2024 without a consumer and nc has never been invoked anywhere in the repository, in any script, workflow or compose file. Both the readiness wait and the healthcheck use curl, and the Ollama install script does not ask for it either.
uv.lock is regenerated output, not hand-edited. It drops three of the four packages plus three transitives that nothing else needs, with no version changes and no additions. pytest stays locked because pytest-asyncio in the dev group still requires it. The dependency markers it adds on the CUDA and numpy entries are inert: each one is a superset of the condition its parent already installs under, and the resolved default install set is identical before and after.
This saves roughly 2 MB uncompressed, which is nothing next to the image as a whole. The point is that a production image stops shipping a test framework and a Docker socket client it never uses.
Everything else stays and is load-bearing. The container installs pip packages at runtime for user-authored tools and functions, so it needs git and a working compiler for anything that is not a prebuilt wheel, and libmariadb-dev for the manual MariaDB install. zstd is required for updating Ollama inside the bundled image. black looks dev-only but backs the code formatting endpoint.
Ref: https://github.com/open-webui/open-webui/discussions/28716
The legacy function-calling path acted on the client-supplied `features` dict after checking only the per-user permission, so a user who still held `features.web_search` or `features.image_generation` could keep triggering web searches and image generation after an administrator had switched those off instance-wide. The native function-calling path already gates the equivalent builtin tools on `web.search.enable` and `image_generation.enable` in `get_builtin_tools`, so the two paths disagreed and the admin-level switch did not actually stop the outbound provider calls it was turned off to stop.
Gate the legacy web search handler on `web.search.enable` at its call site, and gate `chat_image_generation_handler` on the two image switches internally. The image handler needs the check inside it because `image_generation.enable` and `images.edit.enable` are independent: editing stays available when generation is disabled, matching the `/images/generations` and `/images/edit` routes and the native `generate_image`/`edit_image` tools. The handler calls `image_generations`/`image_edits` directly and so bypasses the route guards, which is why the check has to live at the caller.
The "Creating image" status event moves below the new guard so a disabled configuration returns without leaving an unresolved progress indicator in the chat.
Seven route handlers declare a request-scoped database session as a FastAPI dependency and then never touch it. Three of them are `GET /api/v1/users/user/settings`, `/user/status` and `/user/info`, which the frontend hits on every page load, and all three carry a comment saying the user object is already available, so the parameter is leftover from the refactor that removed the refetch. The other four are admin-only external-knowledge connection endpoints that read their data from the config store.
Measured on a route with and without the dependency, 20k requests, best of 5:
| | µs per request |
| --- | --- |
| no dependency | 16.18 |
| unused session dependency | 62.85 |
The dependency costs about three times as much as everything else the request does put together. It is worth being precise about why, because the obvious guess is wrong: this is not database I/O and not connection pool pressure. SQLAlchemy connects lazily, so a session that is never used checks out zero connections, verified by watching the pool's counter stay at zero across the request. The cost is FastAPI resolving an extra async-generator dependency onto the request's exit stack, plus constructing and closing the session object.
Deleting the seven parameters is the whole change. An AST scan over the backend finds exactly these seven handlers before and none after.
With database session sharing enabled, which the docs recommend for PostgreSQL and for multi-replica deployments, the knowledge pending-files and file process-status endpoints each pinned one pooled connection for as long as their SSE stream stayed open, up to one and two hours respectively. A file wedged in processing keeps a stream open for the full duration, so a handful of users sitting on that page can consume every connection in the pool, and the held transactions sit idle and block autovacuum on those tables.
Both handlers took a request-scoped session for their access checks, and FastAPI only releases a yield dependency once the response body has finished streaming, so the session outlived the handler by the whole life of the stream. Neither generator ever used it. They no longer take that dependency, and the queries they run already open their own short-lived sessions when none is passed. This is the approach the chat completion endpoints already use for the same long-response problem.
Measured against a pool with capacity 11: before, at most 11 concurrent streams could ever be open and every further attempt failed, deterministically across repeat runs. After, 25 of 25 opened. Non-stream latency is unchanged, within run-to-run noise, and behaviour is identical whether session sharing is on or off.
Three searches LIKE against cast(json_col AS text), which means they have to match
bytes a JSON encoder wrote. Encoders disagree on non-ASCII: stdlib escapes it to
\uXXXX, orjson writes it raw. Which one produced a row depends on the codec in force
when it was written, so any single pattern finds only half the table.
models.py hard-codes the stdlib spelling, with a comment asserting SQLite stores
JSON via json.dumps(ensure_ascii=True). Model.meta is a JSONField, which has
serialised through JSONCodec since ENABLE_ORJSON was introduced, so on that setting
it stores raw UTF-8 and the escaped pattern matches nothing: non-ASCII workspace
model tag search is broken today. prompts.py and automations.py hard-code the
opposite spelling and miss rows written the other way.
json_text_variants returns both spellings a string can take inside serialised JSON,
collapsing to one for ASCII, and the three call sites OR over them. Rows written
under either setting are now found under either setting, which also covers a
database holding a mix of the two.
Case handling is unchanged. models.py keeps matching non-ASCII tags case-sensitively
on SQLite, whose LOWER() is ASCII-only and would not fold the stored text the way
str.lower() folds the tag. ASCII tags collapse to a single variant and take exactly
the query they took before.
Verified on SQLite across every combination of codec-that-wrote-the-row and
codec-the-app-is-running, for an ASCII and a CJK tag, over all three call sites: 24
of 24 match, against 12 of 24 before. Quoting still bounds whole-tag matches, so
searching "weather" does not match a row tagged "weathervane".
Co-authored-by: Claude <noreply@anthropic.com>
periodic_usage_pool_cleanup, periodic_session_pool_cleanup and
scheduler_worker_loop were started with asyncio.create_task and their
handles discarded. The event loop keeps only a weak reference to a task,
so a task with no other referent can be garbage collected while it is
suspended at an await. All three are while True loops meant to run for
the process lifetime, and if one is collected the failure is silent:
pool entries stop being cleaned up, or automations and calendar alerts
stop firing, with nothing logged.
Six lines above, redis_task_command_listener is already stored on
app.state and cancelled on shutdown. This applies the same treatment to
the other three.
Closes#28052
POST /api/v1/models/sync only worked when every model in the payload was new. As soon as one id already existed, the whole call blew up and the endpoint still answered HTTP 200 with an empty list, so nothing was updated and well-behaved clients saw a success. Only a first-ever sync into an empty catalogue went through.
The update branch splatted the model dump (which already carries user_id and updated_at) and then passed both again as explicit keyword arguments, which is a duplicate-keyword TypeError before SQLAlchemy ever sees it. The insert branch right below merged the same values into a dict first, so it never collided.
Fixed by building that dict once and using it for both branches, matching how sync_functions already does it. Left the broad exception handler alone: it is the reason the failure was silent, but changing the error contract of sync_models is a separate call.
Fixes#28033
The Playwright loader's route interceptor now performs each intercepted request with the same requests/aiohttp clients the other web loader paths already use and fulfills the page with that response, rather than having the browser issue it. Redirect handling, header forwarding and cookie delivery to the browser are unchanged.
Two consequences worth knowing. Page requests now leave from the backend instead of the browser, so with PLAYWRIGHT_WS_URL set they originate from a different host, and TLS is verified against certifi plus AIOHTTP_CLIENT_SSL_CERT_FILE rather than the browser's own trust store. And because the synchronous interceptor blocks, sub-resources on that path fetch one at a time: 30 assets at 40ms went from 2.01s to 3.01s, and 8 assets at 500ms from 1.05s to 4.50s. The asynchronous path is unaffected, at 0.65s and 1.05s respectively.
Both the OAuth and SCIM user lookups now compare the nested JSON value with SQLAlchemy's subscript operator, which emits the correct SQL for each supported database on its own. This replaces the hand-written sqlite and postgresql branches and the column-level contains() call they used.
* fix: preserve user info in agentic RAG tools
* fix: preserve user info in file access checks
---------
Co-authored-by: Damien SPINELLI <damien.spinelli@external.list.lu>
The /openai/responses endpoint forwarded the prefixed model id (e.g.
"myprovider.gpt-4o") to the upstream provider instead of the stripped
native name, causing "model not found" errors when a connection has a
Prefix ID configured.
generate_chat_completion() already strips the prefix before forwarding;
apply the same strip_provider_model_prefix() call in responses() after
the urlIdx routing (which needs the prefixed id) and re-serialize the
body afterwards.
Also fixes the Azure non-v1 deployment path, which built the deployment
URL from the prefixed model name.
Co-authored-by: Claude <noreply@anthropic.com>
A skill ID goes straight into the path of every mutating skill endpoint (/api/v1/skills/id/{id}/...), but create only replaced spaces with hyphens. An ID containing a "/" was stored verbatim as the primary key, so the route never matched, the request fell through to the SPA static mount and the client got 405 Method Not Allowed. The skill could not be opened, edited, toggled or deleted, by admins either, and since skill.name is UNIQUE it could not be recreated under a corrected ID. Percent-encoding does not help: uvicorn decodes the path before Starlette routes it, so the only remaining fix was a direct database write.
Create now rejects any ID outside [a-z0-9_-] with 400 instead of silently storing an unreachable one. Two frontend paths that fed unsanitized IDs into it are fixed as well: the manual "Skill ID" field, which was bound with no sanitization at all and is the path that reproduces on every version, and the markdown import, which put the raw frontmatter name into the ID before opening the editor in clone mode, where the reactive slugify is disabled.
Existing rows with an unreachable ID are not repaired here; rewriting a primary key would also have to re-point the access grants keyed on it.
Fixes#27655
With ENABLE_OAUTH_GROUP_MANAGEMENT enabled, every SSO login reconciles the user's group membership against the IdP claims, adding and removing them from groups and, with ENABLE_OAUTH_GROUP_CREATION, creating groups that do not exist yet. None of it emitted an event, so the same membership change was observable when an admin made it through the UI or when it arrived over SCIM, but invisible when the IdP drove it. That is the path that changes membership most often.
Emits group.member_added and group.member_removed per membership transition and group.created for each auto-created group, using the same payload keys as the groups router. The member events are published only when the write returned a group, so a failed or no-op write emits nothing, and both loops already run only on an actual transition. update_user_groups takes the request so the events can be published; it has a single caller.
The tool callable now takes its connection's cookie jar as a parameter, matching how its headers are already passed and how the terminal tool factory in the same module builds its callables.
The gate now applies the same authorship condition the channel message update route already uses, so both paths agree on which messages a caller may modify.
Changing "Default Pinned Models" in admin settings had no effect for anyone who had already opened Open WebUI once. The sidebar copied the admin default into that user's own settings the first time it rendered and saved it to the server, which marked them as having customized their pins, so every later change to the default was ignored for them. Merely loading the page was enough, the user never had to touch a pin.
The default is now resolved for display only, through a shared store that falls back to the admin list while the user has no pins of their own, the same way default models already work. Nothing is written to the user's settings until they actually pin, unpin or reorder something, at which point their choice takes over for good. Unpinning everything still persists an empty list rather than snapping back to the default.
Users whose settings were already overwritten by the old behaviour keep that copy, since a stored pin list cannot be told apart from a deliberate one.
Fixes a drag-reorder path that mixed sidebar positions with stored ones, and stops the sidebar section reopening itself after any unrelated settings change.
Every toggle row in the chat integrations menu (filters, Web Search, Image, Code Interpreter, Tools, Skills) is a button whose on/off state was carried only by the decorative Switch inside it. Screen readers announced the row name and nothing else, so there was no way to tell whether a tool or feature was active without looking at it.
Each row button now carries aria-pressed, and the Switch wrapper is marked inert so the nested role=switch stops competing with the row for the announcement and stops adding a nameless tab stop. Hit testing skips inert content, so clicking the switch still toggles the row.
Tool rows that are not yet authenticated omit aria-pressed: activating those starts an OAuth redirect rather than toggling, so announcing them as an unpressed toggle would be wrong.
The Web Search, Image and Code Interpreter rows also had a state-flipping aria-label ("Disable Web Search") on top of aria-pressed, which announces as "Disable Web Search, pressed" and reads as the opposite of the truth. Removed: the visible row text already names each control.
Fixes#17150
* fix: enforce the image generation flag on the legacy chat feature path
* fix: refresh the config store when image generation is disabled in admin settings
* fix: hide active feature pills when the feature is no longer available
Revoking a user's `features.memories` permission removed their access to the memories API and to the native function-calling memory tools, but their stored memories were still injected into the system context on the legacy function-calling path.
The branch in `process_chat_payload` only checked the client-supplied `features['memory']` flag plus the global `memories.system_context.enable` switch, with no user-permission check. `add_memory_context` did not compensate: it only checks `model_allows_memory`, which is a model capability rather than a permission, and the one call inside it that does check the permission (`query_memory`) has its 403 swallowed by a `try/except`, so `Memories.get_memories_by_user_id` and the neighbourhood scan still fed the system prompt.
Gate the branch with the same permission check the native path already performs in `get_builtin_tools`, matching the neighbouring `web_search` and `image_generation` branches.
Only the caller's own memories were injected into the caller's own context, so there was no cross-user exposure. The practical effect was that the permission toggle did not do what its name implies: an admin who revoked it still got memory content injected for that user.
Clicking "Merged Response" in a multi-model chat often made the merging model answer with "It appears that the responses provided from the other models were empty".
The merge handler collected each model's answer via `history.messages[id].content`. Assistant messages are persisted by the backend with `output` only (`upsert_message_to_chat_by_id_and_message_id` writes `done`/`role`/`output`, never `content`), so `content` is only populated in the browser session that generated the responses, where the streaming handler mirrors it. Once the chat is reloaded from the database, every assistant message has `content: ''` and the merge request is sent with a list of empty strings, which is exactly what the merging model then reports. That is why the failure looks random: merging works right after generating, and fails after a refresh or when reopening the chat.
Read the responses through `getOutputText(message.output) || message.content`, the same fallback already used by every other read site (`ResponseMessage`, `Overview/Node`, `SearchModal`, `ChatItem`, `ChatMenu`, `Navbar/Menu`).
Fixes#26962
Only the frontend added `stream_options: {include_usage: true}` to the completion payload, gated on the model's `usage` capability. Every backend-initiated run builds its own payload (automations, timers, subagents, channels) and omitted it, so those responses came back without token counts and never rendered the usage block, even with the capability enabled on the model.
Set it in `chat_completion` instead, the single handler all of those callers go through, and drop the two duplicate copies (the Anthropic-compat handler and the frontend). Capabilities are read from the resolved model before the custom-model fallback can rebind it, and the flag is applied after the model's `stream_response` override so a non-streaming model is unaffected.
Fixes#27653
Every action button on a response message gates its visibility on
'isLastMessage || ($settings?.highContrastMode ?? false)' so that buttons
stay visible (not hover-only) when High Contrast mode is on. The two
Regenerate buttons — the one inside RegenerateMenu and the fallback in the
{:else} branch — were missed and still gated on isLastMessage alone, so on
any non-last message the Regenerate icon was invisible until mouse-over even
with High Contrast enabled (#27638).
Add the same highContrastMode clause to both buttons, matching the other
action buttons in this file.
The "expand input" button was positioned with `fixed top-0 right-0`. That only kept it near the composer by accident: `#message-input-container` sets `backdrop-blur-sm`, and a backdrop-filter makes an element the containing block for fixed descendants, so the button resolved to the top-right corner of the entire composer instead of the text area it belongs to.
That corner is already taken. The `@`-tagged model chip renders as the first row of the same container with its dismiss button at the right end, so with a multi-line prompt and a tagged model the two controls are drawn on top of each other. The attached-files row has the same problem: the button paints over the first thumbnail and its remove button.
Anchor the button to the wrapper that holds the text area instead, using `relative`/`absolute`, so it always sits at the top-right of the input row and below whatever rows precede it. With no chip and no files the position is unchanged. As a side effect the button is no longer a child of the `overflow-auto` scroller, so it can no longer be clipped or scrolled out of view on long prompts.
Fixes#26736
`get_message_list` moves through `messages_map` by key but tracked each message's own `id` field, which the message body does not have to carry. Track the key instead.
The SharingPermissions model defaulted public_tools and public_notes to
True while the config defaults (USER_PERMISSIONS_WORKSPACE_TOOLS_ALLOW_PUBLIC_SHARING
and USER_PERMISSIONS_NOTES_ALLOW_PUBLIC_SHARING) are both False.
On an instance whose stored user.permissions config predates these keys,
GET /api/v1/users/default/permissions fills the gap from the model and
reports both as enabled, while has_permission fills it from
DEFAULT_USER_PERMISSIONS and denies. Saving any unrelated permission then
persists the model's True, granting public tool and note sharing the admin
never enabled.
The chat delete endpoints gate on chat.delete for non-admins, but none of
the UI affordances that reach them were conditioned on it, so a user
without the permission is offered controls that always fail with
"Access prohibited".
Gates every entry point on the same condition the backend enforces,
matching the existing chat.share / chat.export / chat.import gates in the
same components:
- Sidebar chat menu (ChatMenu) Delete item — also covers the Search
modal's menu, which reuses ChatMenu
- Sidebar Shift+hover inline Delete button (ChatItem)
- Sidebar hidden #delete-chat-button keyboard-shortcut target (ChatItem)
- Chat navbar menu Delete item (Navbar/Menu)
- Search modal Shift-held inline Delete button (SearchModal)
- Settings -> Data Controls -> Delete All Chats
- Settings -> Archived Chats per-row Delete
The key generation loop redirected input from a non-existent file
(`SET /p WEBUI_SECRET_KEY=<!random!>>%KEY_FILE%`), printing "The system
cannot find the file specified." once per iteration and leaving the key
file empty, so startup failed with "WEBUI_SECRET_KEY is not set".
Build a fixed-length alphanumeric key by indexing into a charset with
%RANDOM% and write it once with `<nul set /p`. Also quote the key file
path and use delayed expansion so paths with spaces work.
Claude-Session: https://claude.ai/code/session_01CmgBivWjad68mX4yBVWMi2
Co-authored-by: Claude <noreply@anthropic.com>
With ENABLE_OTEL and ENABLE_OTEL_LOGS set, InterceptHandler builds the message once for loguru and then hands the same LogRecord to the OpenTelemetry handler, whose _translate calls record.getMessage() a second time. That used to be free, because the message was already a finished f-string with nothing to substitute. Now that log calls pass lazy %-args, the second call re-runs the whole interpolation, so every exported record is formatted twice.
The two getMessage() calls on a 78 kB retrieval record:
before 373.0 us
after 0.1 us
Stamping the built message back onto the record makes the second call a plain string return. msg and args are both in OpenTelemetry's _RESERVED_ATTRS, so neither ever reaches the exported attributes. The isinstance guard matters: _translate exports a non-str msg such as the dicts routers/audio.py logs as a typed body rather than a string, so those records are left untouched, and they have no %-args to format twice anyway. Body, attributes and severity were compared against LoggingHandler._translate for str, dict, list, int, None, exception and exc_info records.
A message referencing a file that no longer exists rendered a broken
image, because nothing anywhere noticed the failed load. What the user saw
depended on the caller's alt text: in a response the alt is the message
content, so the entire reply was rendered inside the image frame, and the
preview still opened full-screen on a dead image.
Image.svelte now tracks a failed load and renders an 'Image unavailable'
placeholder instead, suppressing the preview for a source that cannot be
shown. The failed state resets when the source changes, so a replaced or
corrected URL is retried. An onError callback lets a caller react without
changing behaviour for the eight existing call sites, which are untouched.
Responses now pass the file name (or 'Generated Image') as alt rather than
the message content, which was never a description of the image.
SRC_LOG_LEVELS became an empty dict when per-module log levels were dropped, and env.py keeps it only as a legacy name. opengauss.py is the last thing in the tree that still indexes it, at module scope, so importing the module raises KeyError: 'RAG' and any deployment on VECTOR_DB=opengauss dies the first time it touches the vector store. The factory imports it lazily, which is why nothing else trips over it. Deleting the line is the whole fix: every other vector backend takes getLogger(__name__) and inherits the root level.
colbert.py passes an argument to a message with no placeholder to consume it:
log.info('ColBERT: Loading model', name)
At INFO, which is the default, logging evaluates 'ColBERT: Loading model' % ('colbert-ir/colbertv2.0',) and raises TypeError: not all arguments converted during string formatting. The record is swallowed by handleError, so loading a ColBERT reranker prints '--- Logging error ---' plus a traceback to stderr instead of the model name. Adding %s prints the name and drops the traceback.
When ENABLE_OAUTH_PERSISTENT_CONFIG is off (the default), oauth.* config is
never persisted and is read from environment variables, but the admin panel
still let admins edit the OAuth/OIDC fields and silently dropped every save on
restart, which kept confusing users who missed the docs warning
(open-webui/open-webui#28247).
The OAuth/OIDC section is now read-only in that case: the admin oauth config
endpoint reports the flag and the UI wraps the section in a disabled fieldset,
slightly dimmed with every control inert but all values still visible, plus a
note naming the env var. Saving skips the OAuth POST since nothing can change.
With the flag enabled the section behaves exactly as before.
Known limits: the guard is UI-side only (the POST endpoint keeps accepting
writes, unchanged), and disabled fields mean values cannot be selected and the
masked client secret cannot be revealed while read-only. Switch.svelte gains a
disabled:cursor-not-allowed style that applies to any disabled switch app-wide.
Co-authored-by: Tim Baek <tim@openwebui.com>
The labeler only listened for issues.opened, so it judged a title exactly
once. Reporters who omit the prefix are asked by the triage bot to add one,
and the title they then fix is never looked at again, leaving a genuine bug
report unlabelled until somebody notices it by hand.
It now also runs on issues.edited. Body only edits return immediately, so
the extra runs are limited to titles actually changing, and a bug label a
maintainer has already removed is not restored: on an edit the issue events
are checked for a previous removal first. The opened path is unchanged and
makes no additional API call.
The title pattern also required a delimiter after the bracket, so the
bracketed form the comment advertises, "[Bug] something is broken", never
matched unless it happened to be written "[Bug]: something is broken".
Both forms match now, while "[bug/perf]" keeps matching and titles that
merely mention the word, such as "[UI Bug]" or "fixed a typo", still do not.
A message sent while an attachment is still uploading is held in the chat
queue until the file settles. For files that release works, because the
message input calls onUpdate once the upload completes, which refreshes the
queued entry and asks the queue to run.
Attaching a URL takes a different path. uploadWeb sets the item to uploaded
but calls neither onUpdate nor processNextInQueue, and the queue is
otherwise only revisited when a generation finishes or when the chat is
mounted while idle. In a new chat with nothing generating, none of those
happen, so a message queued behind a URL waits with no event able to
release it until the chat is reloaded.
uploadWeb now asks the queue to run when it is done, the same way the file
path already does.
Logging in through CyberArk Identity dies at the callback with "Unsupported {'app_id'} in header" and the user sees "The email or password provided is incorrect". Any provider that puts a vendor-specific parameter in the ID token header hits this; CAS was already patched by name, CyberArk is the next one.
Authlib 1.7 verifies ID tokens with joserfc, which rejects header parameters it does not recognise. The old fix registered `client_id` so CAS would work, which only ever fixes one provider at a time. This turns off the unknown-header rejection instead, so any private header parameter is ignored rather than fatal. Signature verification, the algorithm allowlist, `crit` handling and value validation of registered headers all still run, so nothing that actually protects the token is relaxed.
Fixes#28062
Since v0.11.0 shipped aiodns, aiohttp silently switched every outbound request from the OS resolver to c-ares. On some Windows hosts the bundled c-ares 1.34.6 (pycares 5) discovers only 127.0.0.1:53 as nameserver, so every external provider lookup fails (#28013). In Docker the long-lived c-ares channel intermittently stops resolving container names while Docker's embedded DNS keeps answering, which wipes the Ollama model list and fails all in-flight chats with a misleading "Model not found" (#28215).
This restores the pre-0.11 ThreadedResolver (OS resolver) by default and gates the c-ares path behind a new env var, AIOHTTP_CLIENT_ASYNC_DNS_RESOLVER, off by default. The event-loop DNS perf improvement is now opt-in for deployments whose resolver setup is known to work with c-ares, instead of a process-wide side effect of the package being installed.
aiodns is also downgraded and pinned to 3.6.1 (pycares<5), the last release before the broken c-ares 1.34.6 build, so opting in does not hit the Windows regression. The hardcoded AsyncResolver in the Mistral OCR loader now follows the same switch. Simply removing aiodns instead was not an option because opting in would then be impossible, and #28215 showed the Docker failure is c-ares itself, not aiodns 4.x.
Newer OpenAI chat models put metadata on the opening line of a colon fence block, like :::writing{variant="email" id="48173" subject="Short question" recipient="mail@example.com"}. The tokenizer matched that line and discarded it, so every block rendered under the same generic "Writing" heading no matter what it contained.
The opening line is now parsed into an attributes map on the token and the header uses it: the subject becomes the title, the recipient follows it and the full string is reachable on hover when the row is too narrow for it. Blocks without metadata render exactly as before, and the other fence types get the parsed attributes for free.
Attributes are read only from inside the {...} braces, not from the whole opening line. Scanning the whole line turned ordinary prose containing key="value" into metadata, and it backtracked quadratically: a 40k character opening line took 586ms to parse, and that runs again on every re-lex while the message streams. Anchored to the braces it is 0.0ms.
Nothing here turns the recipient into a link or a send action. That metadata is model output and can be steered by whatever is in the context, so a prefilled mail action is a separate decision rather than a side effect of parsing.
Adding a model ID that was already on the list in the connection settings modal simply appended it again, so the same model could sit in the whitelist any number of times. The arena model modal had the same flaw, its dropdown kept offering models that were already selected.
The connection modal now rejects a duplicate with a toast and trims the input first; surrounding whitespace renders invisibly in the list, so an untrimmed ID would slip past the duplicate check and still show up as a visually identical row. The arena modal instead filters already-added models out of the dropdown, matching the existing model selector in the admin settings, so a duplicate can no longer be picked at all. Both modals also drop duplicates when loading a stored list, so configs that already contain them are cleaned on their next save.
Until such a config is re-saved, one residual effect of old data remains: a duplicated ID in an arena model's stored list keeps double weight in the random model draw. New duplicates can no longer be created through the UI.
Fixes#28249
Fetching a URL and saving it were reported as one thing. Everything from
reading the URL to writing the vector database sat inside a single try,
whose handler blamed the knowledge base, so a page that could not be
fetched, parsed or resolved was reported as a knowledge base error even
though nothing had reached the knowledge base yet. Reading the URL now has
its own handler that names the URL, and the knowledge base message is left
to the step that actually touches it.
When YouTube refused a transcript the reason was discarded earlier still:
the loader caught the error, logged it, and returned an empty document
list, so the empty result failed downstream and even the salvageable
explanation was gone before a message was produced. The loader now raises
YoutubeTranscriptError carrying a readable reason, mapped from the
transcript library's own exception types. Blocked requests mention that a
proxy can be configured, and disabled, age restricted, unavailable and
missing language cases each say what actually happened.
URLs that attach successfully are unaffected.
Chat creation and chat moves each carried their own copy of the same folder_id validation, resolving the folder and checking ownership and shared write access in slightly different ways. Both now call a single has_folder_write_access helper, which the chat-completions creation path uses as well, so ownership, inherited write grants and nonexistent or malformed ids behave identically everywhere a chat folder_id is set. The owner case also costs one query fewer than before.
JSONField serializes with JSONCodec, but columns declared as SQLAlchemy's own JSON
type go through the engine's serializer instead, and no engine set one. That left
Chat.chat - the largest blob the app stores - on stdlib json.dumps/loads no matter
what ENABLE_ORJSON was set to, while the rest of the app used the codec. SQLAlchemy
invokes it once per write and once per read, so every chat read and write paid a
full stdlib pass over the whole conversation on top of whatever the caller did.
Both engine constructors are now wrapped so the codec is wired in by default and
cannot be missed by a call site that forgets it; an explicit json_serializer still
wins. The 10 create_engine/create_async_engine calls in this module go through the
wrappers. Vector-store engines (pgvector, mariadb, opengauss) are separate databases
and are left alone.
Serializing and deserializing chat-shaped blobs, median of 11 runs:
| chat blob | write | read |
| --- | --- | --- |
| 600 msgs (2.8 MB) | 10.1 -> 1.7 ms | 8.4 -> 3.7 ms |
| 3000 msgs (14.2 MB) | 51.9 -> 8.0 ms | 48.5 -> 27.8 ms |
| 6000 msgs (28.5 MB) | 105.9 -> 29.5 ms | 112.7 -> 80.5 ms |
With ENABLE_ORJSON off JSONCodec is stdlib json, so this is a no-op until the flag
is set - the change cannot regress a default deployment.
With it on, a round-trip probe through a native JSON column returns objects equal to
the stdlib ones on all 12 shapes tried: ASCII, CJK, emoji, astral-plane, unicode
keys, null bytes, lone surrogates, floats, ints above 2**63 and 2**64, line
separators, empty and deeply nested. Stored text changes for non-ASCII, which is
written as raw UTF-8 rather than backslash-uXXXX escapes and is correspondingly
smaller. Nothing queries that text by escape except two Postgres safety filters in
chats.py, and both still hold: a null byte is escaped identically by both codecs,
and the title filter reads a text column rather than JSON. The ->> and json_extract
searches decode the string before matching, so escaping cannot reach them.
Two differences are inherent to JSONCodec and already apply to every JSONField
column: ints beyond 2**64-1 come back as float, and NaN/Infinity serialize to null
rather than the bare literals stdlib emits - the latter being invalid JSON that a
Postgres json column rejects today. Neither shape occurs in chat blobs. Alembic
builds its own engine and stays on stdlib, which is fine in both directions since
each codec reads the other's output.
Claude-Session: https://claude.ai/code/session_014BXoM6QiFJKisxcxKAXii8
Co-authored-by: Claude <noreply@anthropic.com>
The Responses API handler had a branch for response.output_item.done whose own comment said it was handled specifically below, but it never ran. The generic branch matching any response.*.done event came first in the chain and matched this event too, so it fell through and returned the accumulated output unchanged, leaving the dedicated branch below unreachable since the feature was added.
Moving the dedicated branch above the generic one makes the event apply. On a compliant stream this changes nothing, since response.completed replaces the whole output with the same data straight afterwards. It matters when a provider is less tidy: one that never sends response.content_part.added leaves the assistant's own reply unextractable from the next turn's context, and one that omits response.content_part.done drops the annotations that only arrive with the finished item. Both are repaired by honouring the event.
Worth knowing: the item replaces whatever the deltas accumulated, with no guard against a provider sending back less than it streamed. A reasoning item arriving without its content would therefore lose the reasoning body, which is the same shape of provider brokenness that #27800 already needed a guard for.
An interactive prompt raised by __event_call__ was meant to come back as an error dictionary when it timed out. It never did: sio.call raises socketio.exceptions.TimeoutError, which does not inherit from the builtin TimeoutError the handler was catching, so the exception escaped into plugin code instead. Because that exception carries no message, the call sites that wrap plugin calls in except Exception as e turned it into an empty string, so a timed-out prompt looked like an empty answer rather than a failure, and the error branches written for it were dead.
The handler now catches socketio's class alongside the builtin, so a timeout returns the intended error dictionary and a plugin can tell the two apart.
The session eviction that sat inside that handler is removed rather than switched on. It had never executed, and it is wrong in both directions: it compares the pool entry by value, which the heartbeat rewrites every thirty seconds, so it would usually not fire, and when it did fire on a short timeout it would evict a live tab whose user had simply not answered yet, with nothing to restore the entry short of a reload. Genuinely dead sessions are already reaped on missed heartbeats by periodic_session_pool_cleanup.
WEBSOCKET_EVENT_CALLER_TIMEOUT is unset by default, which means no timeout at all, so this only affects deployments that set it.
The generic response.*.delta branch could leave the streaming handler in two states that crash the caller. It bound its result only inside the guard that checks the target item exists, but returned that result outside the guard, so a delta arriving before its output item, or carrying an index past the end, raised UnboundLocalError. Separately, an event name with only two dot-separated parts failed the length check and fell off the end of the branch, so the function returned None and both call sites raised TypeError unpacking it.
Where the response is streamed to a browser both crashes were swallowed at debug level and cost a chunk. On the direct API path there is no handler between here and the server, so the caller kept its 200 while the body was cut short with no [DONE], and the outlet filters never ran.
The return now sits inside the guard with a branch-level fallback that hands back the accumulated output untouched, which is what the sibling done branch and every other skip path in this function already do. Deltas whose item exists behave exactly as before.
Dropping an orphan delta is deliberate rather than synthesizing the missing item: response.output_item.added appends without regard to output_index, so a placeholder would be duplicated when the real item arrives, and a fabricated function_call would have no name or call id.
Raising GLOBAL_LOG_LEVEL to WARNING buys quieter output but not less work: 241 INFO call sites interpolate their payload into an f-string before the logging call gets to drop it. The heaviest is get_doc, which logs every chunk id and metadata dict in a collection, so on the full-context retrieval path that is the entire knowledge base, once per chat request.
That one line at WARNING, CPython 3.12:
| knowledge base | payload | before | after |
| -------------- | ------- | -------- | ------- |
| top-k of 3 | 1.2 kB | 3.8 us | 0.07 us |
| 500 chunks | 201 kB | 583.6 us | 0.08 us |
| 5000 chunks | 2.0 MB | 5.8 ms | 0.15 us |
The lazy form log.info('query_doc:result %s %s', result.ids, result.metadatas) hands the payload to record.getMessage(), which the InterceptHandler only reaches once a record has passed the level check. Output at INFO is byte-identical. Two sites that already built their message eagerly, one str concat and one % operator, move to the same lazy form.
`backend/open_webui/models/chats.py` imports `json`, but the module contains no `json.` references. The only remaining matches in the file are SQL function names such as `json_each` and `json_typeof` inside `text()` strings, which are unrelated to the import.
Noticed while profiling the chat read/write path: the import made it look as though the module serialized locally, when all of that happens in `utils/misc.py:sanitize_data_for_db`.
One-line deletion, no behaviour change.
The `response.completed` handler replaced the accumulated output with the terminal event's `output` whenever that key was present, guarded only by `is not None`. An empty array satisfies that guard, so a provider that finishes the stream with `"output": []` wiped everything collected from `response.output_item.added`, `response.output_text.delta` and `response.output_item.done`.
The assistant message was then persisted with `output: []` and empty content, which shows up as a reply that renders correctly while streaming and disappears the moment the stream ends.
Fall back to the accumulated output when the terminal array is empty. A spec-compliant `response.completed` still wins, since a populated array is truthy, and when nothing was streamed the accumulated output is empty too, so the fallback cannot invent content.
Fixes#27789
Permission checks are the most repeated database work in a request, and every one of them asks the same question: which groups is this user in. Today that question cannot use an index.
`group_member` has only its primary key and a `(group_id, user_id)` unique constraint. That constraint leads on `group_id`, so a lookup by `user_id` has to walk the entire membership table, every time. `Groups.get_groups_by_member_id` sits under `has_permission`, `has_access`, `check_model_access` and the `AccessGrants` fallbacks, so an ordinary chat completion pays that walk several times before the model is even called, and the admin user list pays it once per row.
The cost scales with total memberships across all users rather than with the size of any one user's, so it stays invisible on a small instance and then arrives all at once on a large one.
Measured on SQLite, timing the real join from `get_groups_by_member_id`:
| memberships | before | after |
|---|---|---|
| 5,000 | 0.04 ms | 0.03 ms |
| 50,000 | 0.10 ms | 0.04 ms |
| 200,000 | 1.33 ms | 0.04 ms |
| 500,000 | 2.94 ms | 0.04 ms |
The after column is flat because the lookup becomes a seek instead of a scan. Concretely: on a deployment with 500k memberships, say 10,000 users in 50 groups each, one chat completion currently spends roughly 15 ms of database time answering the same question over and over. Afterwards it is under 0.2 ms. On a small install you will not be able to measure the difference, and that is fine, the point is that the curve stops bending.
The index is `(user_id, group_id)`. The trailing column makes those lookups index-only, since `group_id` is the column they select. Queries that lead on `group_id`, such as `get_group_user_ids_by_id` and the `chat_messages` subqueries, are already served by the existing unique constraint and are unaffected.
What to expect when the migration runs: on PostgreSQL this is a plain `CREATE INDEX`, which takes a SHARE lock, so reads continue while writes to `group_member` block until it completes. The table holds one row per membership, so expect sub-second even on the numbers above. `CONCURRENTLY` cannot be used here because the migration runner wraps the upgrade in a transaction, and it is not warranted at this table size.
GLOBAL_LOG_LEVEL defaults to INFO, so every log.debug(...) in the backend is discarded, but the message is built first: 187 call sites interpolate their payload into an f-string before the logging call runs, so the work happens on every request and the result is thrown away. The worst one sits in process_chat_payload and stringifies the whole request body, full conversation history included, once per chat completion.
That one line with DEBUG disabled, CPython 3.12:
| conversation | payload | before | after |
| ------------ | ------- | -------- | ------- |
| 4 messages | 1.2 kB | 3.4 us | 0.07 us |
| 20 messages | 17 kB | 24.8 us | 0.07 us |
| 60 messages | 123 kB | 216.6 us | 0.07 us |
The lazy form log.debug('form_data: %s', form_data) hands the payload to record.getMessage(), which the InterceptHandler only reaches once a record has passed the level check. With DEBUG enabled the emitted lines are byte-identical, f'{x=}' sites included: those map to %r. MistralLoader._debug_log callers get the same treatment, since that wrapper already forwards *args.
`get_permissions` deep-copies the default permission tree with a `json.loads(json.dumps(...))` round trip before merging group permissions into it. It runs on signin, signup, the permissions endpoint, OAuth, and the chat-completion middleware.
It now goes through `JSONCodec`, which selects orjson when `ENABLE_ORJSON` is set. The intermediate string never leaves the expression, so neither the escaping nor the separator differences between the two backends are observable; only the resulting object is used.
`default_permissions` always originates from `Config.get('user.permissions')`, a SQLAlchemy `JSON` column, so the tree is JSON-native by construction and the round trip is exact.
Note for anyone tempted to simplify this to `copy.deepcopy`: measured on the real `DEFAULT_USER_PERMISSIONS` shape over 200k iterations, `deepcopy` takes 3.51s against 1.45s for the stdlib round trip and 0.40s for orjson. The round trip is the fast option, not a workaround.
With `ENABLE_ORJSON` unset, which is the default, `JSONCodec` is stdlib `json` and this call site behaves exactly as before.
The code interpreter parses every message from the Jupyter kernel websocket with stdlib `json`, in a loop that runs for the duration of an execution. Messages carrying large stdout or a base64 image payload are the expensive ones.
It now goes through `JSONCodec`, which selects orjson when `ENABLE_ORJSON` is set. Every consumer of the parsed message reads strings only: `content.text`, `content.data['text/plain']` and `['image/png']`, `content.traceback`, and `content.execution_state`. Jupyter renders large integers into `text/plain` as strings rather than JSON numbers, so no numeric round trip is involved.
The one-shot `execute_request` message this module sends keeps stdlib `json`; it is a small fixed-shape dict sent once per execution.
With `ENABLE_ORJSON` unset, which is the default, `JSONCodec` is stdlib `json` and this call site behaves exactly as before.
`_json_to_metadata` parses the metadata of every result row returned by search and get. It now goes through `JSONCodec`, which selects orjson when `ENABLE_ORJSON` is set.
The text it parses is produced by Oracle's own `JSON_SERIALIZE`, and the column is a native `JSON` type, so the database normalises whatever was written and the reader never depends on the writer's escaping.
The matching `_metadata_to_json` write deliberately keeps stdlib `json`: it passes `default=self._decimal_handler`, orjson accepts none of stdlib's keyword arguments, and dropping the handler would turn a currently successful insert of a `Decimal` into a hard failure. The read side has no such constraint.
With `ENABLE_ORJSON` unset, which is the default, `JSONCodec` is stdlib `json` and this call site behaves exactly as before.
`routers/ollama.py` serializes the outbound body with stdlib `json` on six inference paths: `/api/chat`, the OpenAI-compatible completions and chat completions proxies, embeddings, the Anthropic messages proxy, and responses. All six carry a full conversation or an embedding batch.
They now go through `JSONCodec`, which selects orjson when `ENABLE_ORJSON` is set. Every one is passed to `send_request`, which hands it to aiohttp as `data=`; aiohttp encodes `str` as UTF-8 and derives `Content-Length` from the encoded bytes. None is hashed, cached, length-measured or persisted.
Admin model management keeps stdlib: `/api/unload`, `/api/pull`, `/api/delete` and `/api/show` serialize fixed one- or two-key dicts, as do the blob download and upload progress events and the error frames. Codec dispatch on those costs about what it saves.
With `ENABLE_ORJSON` unset, which is the default, `JSONCodec` is stdlib `json` and this call site behaves exactly as before.
`passthrough_anthropic_messages` in `main.py` serializes the full request payload with stdlib `json` before sending it upstream. It is the largest single serialization on that path, since the body carries the whole conversation.
It now goes through `JSONCodec`, which selects orjson when `ENABLE_ORJSON` is set. The result is passed to aiohttp as `data=`, which encodes `str` as UTF-8 and sets `Content-Length` from the encoded bytes. The payload originates from a parsed request dict, so it holds only JSON-native types, and the serialized string is never hashed, compared or persisted.
The remaining stdlib `json` calls in this module are left alone: two are a debug log line and a fixed Ollama unload payload, and one parses an upstream error body.
With `ENABLE_ORJSON` unset, which is the default, `JSONCodec` is stdlib `json` and this call site behaves exactly as before.
`request.app.state.MODELS` is a `RedisDict` when Redis is configured. Unpacking it with `{**pool}` makes Python call `keys()` and then `__getitem__` once per key, which is one HKEYS plus one HGET per model, issued sequentially through a synchronous client. At 200 models that is 201 blocking Redis round trips per call.
`RedisDict.items()` is a single HGETALL, so `dict(pool.items())` fetches the same data in one round trip. `utils/chat.py:184` already does exactly this and carries a comment explaining why; these ten call sites were missed.
They are on the direct-connection branch of the task endpoints (title, tags, follow-up, autocomplete, query generation and the rest), of `chat_completed`, and of context compaction, so they run for background tasks fired on ordinary chat turns.
Behaviour is unchanged. The merged mapping is identical, the explicitly added direct model still overrides any pool entry with the same id, and when Redis is not configured the pool is a plain dict where `dict(d.items())` and `{**d}` are equivalent.
It also closes a race. `RedisDict.set` writes with HSET and then HDELs the stale keys, so a key returned by HKEYS could be deleted before its HGET arrived, raising `KeyError` out of the dict literal and failing the request mid model refresh. The old path could likewise observe a mix of pre- and post-refresh entries. HGETALL is atomic, so the caller now always sees one coherent snapshot.
Fourteen modules import `json` without using it. Ruff flags every one with F401, and a word-boundary search for `json` in each file matches only the import line itself, including inside strings, comments and annotations.
Two exclusions, both deliberate. Migration files are left alone: the import is equally dead there, but those files are frozen history and not worth the churn. `models/chats.py` has the same dead import and is handled in its own change, so it is skipped here to avoid two changes touching the same line.
No behaviour change.
orjson emits U+2028, U+2029 and U+0085 raw, where stdlib `json.dumps` escapes them under its default `ensure_ascii=True`. Python treats all three as line boundaries, so with `ENABLE_ORJSON` set, one of them inside model output splits a `data: {...}` SSE frame in half. Both halves then fail to parse and the delta is dropped with no error.
`utils/middleware.py` reassembles frames with `splitlines()`, so an affected response silently loses content on the direct API path. External clients are exposed as well: httpx's `LineDecoder` reimplements the same line-boundary semantics, so any SDK reading the OpenAI-compatible stream through `aiter_lines` breaks on a raw separator.
The three characters are escaped on the way out of `ORJSONCodec.dumps`. That restores parity with stdlib and fixes every reader at once, rather than patching one consumer and leaving external clients broken. They are the complete set: of the ten code points `splitlines()` treats as boundaries, the other seven are below U+0020, where JSON already forces an escape.
The membership guard is load bearing. Calling `translate` unconditionally costs roughly 1.5 us on a typical SSE chunk against 0.115 us for the serialization it wraps, so it would spend more than orjson saves. The three scans cost about 0.04 us.
Payloads containing none of the three are returned unchanged, byte for byte. With `ENABLE_ORJSON` unset, which is the default, none of this code runs.
U+2028 and U+2029 are common in text extracted from PDFs and word processor documents, so the realistic trigger is a model quoting an uploaded file back to the user.
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
The Valkey backend serializes chunk metadata on every insert and parses it back on every result row in `get` and `query`. Both directions now go through `JSONCodec`, which selects orjson when `ENABLE_ORJSON` is set.
The stored `metadata_json` field is never matched against as text. `_build_filter_expression` only emits TAG predicates, and the TAG fields are `id`, `hash`, `file_id`, `source` and `knowledge_base_id`; `metadata_json` appears only as a return field that is immediately re-parsed. So rows written with escaped non-ASCII and rows written raw are indistinguishable to every reader, and no migration is needed.
`process_metadata` already stringifies datetimes and strips null bytes and lone surrogates before the write, so the two backends cannot disagree about what is serializable here.
Both read `except` clauses widen from `(json.JSONDecodeError, TypeError)` to `(ValueError, TypeError)`. The codec falls back to engineio's codec, which installs `parse_int=_safe_int` and raises a bare `ValueError` for integer literals longer than 100 characters; the narrower clause would have let that escape and abort a search instead of yielding empty metadata. `json.JSONDecodeError` is a `ValueError` subclass, so this is a strict superset. That removes the module's last use of stdlib `json`, so the import goes with it.
With `ENABLE_ORJSON` unset, which is the default, `JSONCodec` is stdlib `json` and this call site behaves exactly as before.
RAG vector search runs in a thread pool, but then calls `future.result()` on the event loop thread, so the whole worker freezes until every collection answers. Every other user's token stream stops for that long. It's the default retrieval path.
Now `asyncio.gather` over `asyncio.to_thread`, matching what `routers/retrieval.py:2779` already does for the same call.
Measured with 3 queries across 4 collections, 60 ms search, and a second request wanting a turn every 5 ms:
| | before | after |
|---|---|---|
| RAG call | 62.0 ms | 61.2 ms |
| other request's turns | 0 | 7 |
| its worst stall | 62.5 ms | 16.0 ms |
Same results, same order, same `(result, error)` contract. Cancellation now lands mid-search instead of after every thread finishes. Threads move from an unbounded per-call pool to the loop's bounded shared one.
Swap the JSON encoder/decoder used across the backend from stdlib json to
orjson when ENABLE_ORJSON is set — HTTP request bodies, JSONResponse
bodies, upstream provider responses, SSE chunks, and socket.io/Redis
payloads.
The flag defaults to off, in which case the app uses stdlib json and
engineio's codec verbatim, so default behaviour is unchanged.
- json_codec exports JSONCodec (stdlib json or the orjson codec) and
SOCKETIO_JSON (engineio's codec or the orjson codec); call sites import
JSONCodec and stay implementation-agnostic
- apply_orjson_http_json() is a no-op when the flag is off, leaving
starlette's Request.json / JSONResponse.render untouched
- the orjson codec falls back to the stdlib for inputs orjson rejects
(non-str dict keys, ints beyond 64 bits, NaN literals)
- orjson is imported only when the flag is on
- FastAPI(default_response_class=...) is deliberately not used: an
explicit default disables the Pydantic direct-to-bytes fast path for
response_model routes
build_matcher compiled a caller-supplied pattern with Python's backtracking re and ran it over every line of every reachable file, with no timeout, no thread offload and no length caps. is_regex_pattern promotes any pattern containing a metacharacter, and a bare pipe counts, so no explicit regex flag is needed to reach the compiler. The search loop is synchronous inside an async handler, and UVICORN_WORKERS defaults to 1, so the cost lands on every other user of the instance. MAX_GREP_RESULTS bounds how many matches are reported, not how much work is done.
Backtracking cost is exponential in the length of the text being matched, so capping the pattern or the line does not bound it: the subject in the measurements below is 30 characters. `(x|x)*y` against a line of 30 x took 80 seconds, `(a+)+$` against 32 a took 169 seconds, and the same subject with a literal pattern took 0.6 microseconds.
Matching now runs on the regex module, which accepts a per-search timeout that re has no equivalent for. The timeout is the actual bound: regex resolves many classic catastrophic patterns instantly, but not all of them, and `(a|aa)+$` and `(?:a|a)*$` still need it. The budget covers a whole tool call rather than a single search, because a pipeline builds one matcher per segment and a per-search budget would multiply by segment count, and because a per-line timeout would allow timeout multiplied by line count. It is carried in a context variable so one command shares it without threading a parameter through every handler, and it is charged only for time spent inside search(), so database round-trips and other coroutines cannot consume it. Exhausting it raises, and both entry points already render that as an error for the model to read.
Note for anyone tracking search behaviour: re and the regex module define \w, \W and \b differently on non-ASCII text. re follows str.isalnum(), the regex module follows UTS#18, so \w no longer matches superscripts and fractions such as the ones in Nd-adjacent categories, and now does match combining marks. POSIX classes like [[:alpha:]] are interpreted rather than read as a literal set, and \p{...} compiles instead of erroring. Results on ASCII content are unchanged.
regex was already installed as a transitive dependency of nltk, tiktoken and transformers. It is now declared directly, pinned in pyproject.toml and requirements.txt to the version the lockfile already resolves.
On latest `dev`, the chat that is currently open is indicated **only** by a background tint: `bg-black/[0.035]` in light mode and `dark:bg-white/[0.045]` in dark.
Against the page background that is **1.07:1** in light and **1.05:1** in dark. It is close to imperceptible for sighted users, and it carries no programmatic state at all, so assistive technology has no way to tell which entry in the list is the one being viewed.
Breaks WCAG 1.4.1 Use of Color (Level A), since the state is conveyed by colour alone, and 4.1.2 Name, Role, Value (Level A), since the state is not exposed.
Fix: set `aria-current="page"` on the chat link when it is the open chat, using the same `id === $chatId` condition that already drives the visual highlight, so the two cannot drift apart. `'page'` is the correct token because the trigger is a real navigation to `/c/{id}`.
This matches the existing pattern in `routes/(app)/workspace/+layout.svelte` and `chat/Placeholder/ChatList.svelte`, which already set `aria-current` for their active entries.
Verified that the bits-ui `LinkPreview.Trigger` forwards unknown attributes to the rendered anchor and does not set `aria-current` itself, so the attribute reaches the DOM.
This does not change the visual contrast of the highlight, which is worth addressing separately.
Severity: Serious. In a long chat list there is no reliable way to tell which chat is open.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, the sidebar folder row is a bare `<div>` carrying `on:click` (navigate into the folder) and `on:dblclick` (rename). It has **no `role`, no `tabindex` and no key handler**, so opening a folder is impossible from the keyboard.
The nested chevron `<button>` is focusable, but it only expands the folder in place, it does not navigate to it, so there is no keyboard route to the folder page at all.
Breaks WCAG 2.1.1 Keyboard (Level A) and 4.1.2 Name, Role, Value (Level A). The Svelte compiler already flags this file with `a11y_click_events_have_key_events`; after this change the component compiles with zero a11y warnings.
Fix: apply the row pattern already used elsewhere in this codebase (`workspace/Prompts.svelte`, `workspace/Knowledge.svelte`, `admin/Functions.svelte`), namely `role="button"`, `tabindex="0"` and a keydown handler for Enter and Space, with the same `e.currentTarget !== e.target` guard and the same `shouldIgnoreRowClick` helper those files use.
That guard matters more here than in the files it was copied from: the rename `<input>` is rendered **inside** this row, so without it typing a space in the rename field would be swallowed and navigate away, and Enter would both save the rename and navigate.
The navigation body is extracted to `openFolderHandler` because it now has two callers. The keyboard path calls it directly rather than through the 100ms `clickTimer`, which exists only to disambiguate single from double click and has no keyboard equivalent.
A dead `(e) => e.stopPropagation();` expression statement in the click handler is removed. It allocated an arrow function and discarded it without ever calling it.
The `…` folder menu is still `invisible group-hover:visible` and therefore unreachable, so rename, share, delete, export and new subfolder remain keyboard-inaccessible until that is addressed. That is fixed repo wide in a separate PR that replaces the `invisible group-hover:visible` pattern, so it is deliberately not touched here to avoid conflicting on the same line.
Folder reparenting by drag still has no keyboard alternative, which is a separate WCAG 2.5.7 issue needing a "Move" menu action.
Severity: Critical. Folders cannot be opened without a pointing device.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
Co-authored-by: Tim Baek <tim@openwebui.com>
* fix: fetch terminal system prompt per request with TTL cache
The system prompt was only fetched once in set_terminal_servers (startup
or connection save) with a 3s timeout, using a synthetic 'system' user.
That snapshot silently stays empty when the fetch races a cold-started
orchestrator instance, and goes stale when instances are reprovisioned
with a changed OPEN_TERMINAL_SYSTEM_PROMPT — recovering only after a
restart or a manual connection re-save.
- Fetch /system during get_terminal_tools with the user's own
credentials via a central TTL-cached method (5 min per server+user;
failures cached 60s so a dead instance doesn't stall every request),
falling back to the cached snapshot.
- Raise the fetch timeout from 3s to 30s so cold-provisioned instances
can answer.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KjnQJNKozp47vTB13pRyYs
* refac: fetch the terminal system prompt per request without a cache
Drop the module-level TTL cache and fetch the system prompt directly in
get_terminal_tools, gathered with the existing uncached per-request cwd
fetch that already follows this pattern. The fetch uses the user's own
credentials and falls back to the set_terminal_servers snapshot, so a
cold or unreachable instance degrades to the previous behaviour instead
of needing an error cache. Also restore the 3s timeout: on the request
path a 30s wait would stall chat completions, and a cold instance is
covered by the snapshot fallback until it warms up.
---------
Co-authored-by: Claude <noreply@anthropic.com>
* fix: block external resource loading in Vega chart rendering to prevent client-side SSRF
renderVegaVisualization renders vega/vega-lite chart specs that appear in untrusted chat content (shared chats, channel messages, assistant/RAG/tool output) by constructing a Vega View with no restricted loader, so a crafted spec could make a viewer's browser issue arbitrary outbound requests. There are two paths: data.url (and topojson/geo data) is fetched via loader.load at view construction, and image-mark urls are resolved via loader.sanitize and emitted as <image href> into the output SVG, fetched by the browser when the SVG is displayed. Both are client-side SSRF, and against same-origin or CORS-permissive targets allow reading the response back into the page. Pass a loader that rejects external resource loads on both paths, load throws and sanitize rejects http(s)/protocol-relative URIs, so rendered charts can only use inline data. Inline data.values charts are unaffected.
Co-authored-by: Zureno <Zureno@users.noreply.github.com>
* fix: resolve Vega image urls with the URL parser before blocking external loads
The previous scheme regex could be bypassed with encodings the browser URL parser
normalizes away, such as a leading tab or newline before the scheme and backslash
variants of protocol-relative urls like /\evil.com, which would still be emitted
into the rendered SVG and fetched externally on display. Resolve the uri against
document.baseURI with the browser's own URL parser and only allow data: uris and
same-origin results, so the check cannot diverge from what the browser would
actually fetch. Also shortens the explanatory comments.
---------
Co-authored-by: Zureno <Zureno@users.noreply.github.com>
update_chat_tags_by_id runs at the end of every completion when tag generation is enabled (the default). It loaded the full chat row including the multi-megabyte blob, mutated only meta.tags, committed, then refreshed the row, which re-fetched and re-parsed the entire blob a second time, and finally validated the whole thing into a ChatModel that its only caller (the auto-tagging handler) discards. add_chat_tag_by_id_and_user_id_and_tag_name had the same shape for a one-tag append, and orphan cleanup issued one COUNT query per removed tag.
Both tag writers now select only the meta column and issue a column-level UPDATE, never touching the blob; the single-tag path also skips the write entirely when the tag is already present. Orphan detection batches all per-tag counts into one round trip using one scalar subquery per tag with the exact same dialect-specific EXISTS filters as before; the existing single-tag count delegates to the batch helper so there is one implementation.
Benchmark (real SQLite DB, 200-message chat, ~600 KB blob):
| metric | before | after |
| --- | --- | --- |
| auto-tag update, 3 tags replaced | 12.85 ms | 7.36 ms |
The absolute saving grows with chat size since the blob no longer gets fetched, parsed, re-fetched and validated at all.
Functionally verified against a fresh database: tag replacement normalizes and filters the none placeholder, creates missing tag rows and leaves the blob untouched; orphaned tags are deleted while tags still referenced by other chats survive; single-tag add is idempotent; batch counts agree with the single count including unknown tags; unknown chat ids return None.
POST /knowledge/{id}/sync/cleanup verified write access to the knowledge base in the URL but then acted on the caller-supplied file_ids and dir_ids without checking they belong to that knowledge base. A user with write access to any knowledge base could pass another knowledge base's directory id to delete its directory subtree and knowledge_file associations, or another file's id to drop its file-{file_id} vector collection. Fetch each directory and skip any whose knowledge_id does not match the URL id (matching the explicit directory-delete endpoint), and gate the per-file vector cleanup on Knowledges.has_file(id, file_id) so a foreign file id cannot trigger collection deletion. Legitimate same-knowledge-base cleanup is unchanged.
Co-authored-by: whyiug <whyiug@users.noreply.github.com>
* fix: preserve system prompt across tool calls when memories are enabled
The native tool-call loop runs generate_chat_completion with
bypass_system_prompt=True, so the provider layer does not re-apply the
model's default system prompt on tool-call iterations. It relies instead
on metadata['system_prompt'], captured in process_chat_payload, to carry
the full system prompt forward and restore it after RAG injection.
That capture read the model default system prompt from
form_data['params']['system'], but apply_params_to_form_data had already
popped 'params' from form_data, so model_system_prompt was always empty.
metadata['system_prompt'] therefore only captured whatever was already
materialized in the messages. With memories enabled, that is the injected
<memory_context> system message, so tool-call requests were restored with
memory-only system content and the model's system prompt was dropped.
Without memories there was no system message to capture at all.
Capture the model default system prompt from form_data['params'] before
apply_params_to_form_data pops it, and use that value when building
metadata['system_prompt'].
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014Z4L51mvxDJCx1EDxP2vFN
* refactor: condense system prompt capture comment to a single line
Replace the four-line explanation above the model_system_prompt capture with a one-line note. The variable name and the surrounding code already convey what happens; the comment only needs to state why the capture sits before apply_params_to_form_data.
---------
Co-authored-by: Claude <noreply@anthropic.com>
The web-search-* namespace was the one collection namespace filter_accessible_collections admitted unconditionally for any non-admin user, on both read and write, unlike file-*, user-memory-* and knowledge bases which are owner-scoped. process_web_search now mints these ephemeral per-query collections as web-search-{user.id}-<hash>, and the access helper only admits web-search-{requester.id}-* names, so a web-search collection is readable and writable only by the user who created it (admins keep their bypass). The collections hold transient public web-search results and their names are non-enumerable query hashes, so there was no demonstrated cross-user access path; this removes the namespace exception so the per-user scoping the other namespaces enforce also covers web-search.
Co-authored-by: rexpository <rexpository@users.noreply.github.com>
The terminal proxy's system_oauth auth type read the OAuth access token from the client-supplied x-oauth-access-token request header and forwarded it verbatim as a Bearer token to the upstream terminal server, so an authenticated caller could substitute an arbitrary token for the one bound to their own session. Resolve the token server-side from the caller's OAuth session via oauth_manager.get_oauth_token(user.id, oauth_session_id), matching the openai.py proxy, so the forwarded token is always the one Open WebUI issued for the authenticated user and the client header is ignored.
Co-authored-by: brodmart <brodmart@users.noreply.github.com>
Previously, when web search retrieved pages successfully but saving them to the vector DB failed (for example an unreachable or misconfigured embedding endpoint), process_web_search swallowed the exception at debug log level and still returned status: True with the collection name. The chat then showed "Searched N sites" followed by "No sources found" at retrieval time, hiding the actual misconfiguration from the user and making the failure look like a search bug.
process_web_search now logs the failure at exception level and raises an HTTPException with an actionable message pointing at the embedding configuration in Admin Settings > Documents. chat_web_search_handler surfaces the detail of any HTTPException raised during the search in the emitted error status, so the real cause (embedding misconfiguration, search engine errors, no results) is shown in the chat UI instead of the generic "An error occurred while searching the web". Non-HTTP exceptions keep the generic message, so raw internal error strings are not exposed.
Ref #26750, #25038
The .msg branch routed to langchain's OutlookMessageLoader, which requires the extract_msg package. extract_msg pins beautifulsoup4<4.14, but we pin unstructured==0.22.31 (needs beautifulsoup4>=4.14.3) and beautifulsoup4==4.14.3, so extract_msg can never be installed alongside the current dependency set. As a result the .msg path could not function on any supported install: uploads failed at runtime with an ImportError, and adding the missing package broke the build with an unsatisfiable resolver error.
Switch to UnstructuredEmailLoader, which parses .msg through unstructured's partition_msg (backed by python-oxmsg). Both are already shipped, so .msg uploads work with no new dependency and no version conflict. Attachment partitioning is disabled to preserve the previous body-only extraction behaviour.
Fixes#26690
The backend rewrites its bundled static assets under open_webui/static on
startup. Under OpenShift's restricted SCC the container runs as a random UID
(member of GID 0), which cannot write to the root-owned static dir, so boot
logs fill with '[Errno 13] Permission denied: .../static/*'.
Give GID 0 the owner's permissions on that directory (chgrp 0 + chmod g=u),
the standard Red Hat arbitrary-UID idiom. Applied unconditionally since the
app writes there on every start; complements the opt-in USE_PERMISSION_HARDENING.
The legacy features block in process_chat_payload honoured client-supplied features.image_generation and features.web_search flags and dispatched to the image generation/edit provider and the web-search provider without re-checking the per-user permission that the direct /images routes and the native function-calling path enforce. A user denied features.image_generation or features.web_search could still trigger billable server-side image generation or web search via POST /api/chat/completions with params.function_calling set to legacy. Gate both branches on admin-or-has_permission before invoking chat_image_generation_handler / chat_web_search_handler, matching the existing code_interpreter gate, so a forged flag from an unpermitted user is ignored. Normal completions and permitted users are unaffected.
The chat completion entry point fetched the model row and then check_model_access immediately fetched the exact same row again. Inside the check, the direct grant lookup and every hop of the base-model chain each refetched the caller's group memberships, because neither call passed user_group_ids even though both AccessGrants.has_access and has_base_model_access already accept it.
check_model_access now takes an optional prefetched model_info (used only when its id matches the requested model, so stale callers cannot bypass the lookup) and resolves the caller's group ids once, sharing them across the direct check and the whole base-model chain. The group fetch is skipped entirely for the owner-with-no-base-chain case, which previously needed no groups either.
DB round trips for one completion-entry access check (non-owner model with one base-model hop):
| queries | before | after |
| --- | --- | --- |
| model row SELECTs | 3 | 2 |
| group membership SELECTs | 2 | 1 |
For deeper base-model chains the before column grows by one group SELECT per hop; the after column stays at one.
Functionally verified with stubbed model, group and grant accessors: owner fast path issues no group or grant queries; a non-owner with a base chain resolves groups once and passes the same set to every hop; a prefetched matching model_info skips the duplicate row fetch while a mismatched one is refetched; denial and unknown-model cases still raise; the arena path is unchanged.
The per-id model endpoint (GET /api/v1/models/model) strips params, the system
prompt and other curated model config, for callers who only have read access.
The list endpoint (GET /api/v1/models/list) did not: it returned each
read-accessible model's full params, so a read-shared model exposed its
params.system to non-owner read-grant holders.
Mirror the per-id behaviour: compute write_access per item and drop params
before serialising when the caller lacks write access (not the owner, not an
admin under BYPASS_ADMIN_ACCESS_CONTROL and holding no write grant). The
model-card list UI does not render params, so this does not change
functionality.
Co-authored-by: bogdancherniy11-sudo <229690748+bogdancherniy11-sudo@users.noreply.github.com>
* fix: do not expose tool source code to read-only users
The tool read endpoints build their responses from a content-bearing model via
model_dump() under ConfigDict(extra='allow'). ToolResponse deliberately omits
content (the Python source) and specs, but extra='allow' re-admits both, and the
get_tools defer_content flag was a no-op, so GET /tools/, GET /tools/list and GET
/tools/id/{id} returned a tool's full source to any caller with mere read access,
including any authenticated user for a publicly read-shared tool. Tool source
commonly embeds hard-coded credentials and internal URLs.
Strip content and specs for callers without write access across the three read
endpoints. Tool execution loads source server-side, so tool use is unaffected,
and writers still receive content where they did before. The duplicated
write-access check is extracted into a small helper.
Co-authored-by: bogdancherniy11-sudo <229690748+bogdancherniy11-sudo@users.noreply.github.com>
* fix: limit the tool source strip to the per-id endpoint
Upstream dev has since fixed the defer_content no-op in Tools.get_tools, so the list endpoints (GET /tools/ and GET /tools/list) no longer fetch tool source at all and the stripping added there is redundant. Stripping specs also broke the chat Available Tools modal, which lists a tool's functions from specs for every user who can use the tool.
Reduce the change to the one remaining leak: GET /tools/id/{id} builds its response from a full model_dump() and ConfigDict(extra='allow') re-admits content, so drop content there for callers without write access. Specs stay visible to read users as before and the helper functions are no longer needed.
---------
Co-authored-by: bogdancherniy11-sudo <229690748+bogdancherniy11-sudo@users.noreply.github.com>
convert_payload_openai_to_ollama deep-copied the entire request payload on every completion routed to an Ollama model, and again on every tool-call iteration. The cost of that copy scales with the number of messages and nested content parts in the history, so long chats pay the most, purely as CPU work before the request even leaves the server.
The function only ever mutates two things: it deletes keys on the top-level dict and on the nested options dict. convert_messages_openai_to_ollama already builds fresh message dicts. Shallow-copying exactly those two levels therefore preserves behavior while removing the whole-tree copy.
Benchmark (per conversion call):
| payload | before | after | speedup |
| --- | --- | --- | --- |
| 200-message text chat (~180 KB) | 0.22 ms | 0.057 ms | 4x |
| 20-message chat + 1 MB base64 image | 0.41 ms | 0.38 ms | 1.1x |
The image row barely moves because deepcopy shares immutable strings; the win comes from container-heavy histories, which are exactly the payloads that grow over a conversation's lifetime.
The output is byte-identical to the previous implementation (verified against it, including dict key order, root parameter hoisting, max_tokens remapping, stop handling and response_format precedence), and the caller's payload is left unmodified exactly as before.
* fix: restrict folder deletion to the owner or an admin
Deleting a folder cascades into the folder owner's chats, messages and the
entire subfolder subtree; the cascade is bound to the folder's owner, not the
caller. The delete handler only enforced owner/admin for root folders.
Subfolder deletion required merely write access, and a write grant on a shared
root folder is inherited by every descendant subfolder. A write-collaborator
could therefore permanently delete the owner's chats by deleting a subfolder of
a shared folder, data they do not own. With delete_contents=false the same path
force-moved the owner's chats out of the folder instead.
This also contradicted the documented sharing model: only the owner or an admin
may delete a shared folder, and write access covers adding and editing chats and
subfolders, not removing the folder.
Because any folder deletion cascades into the owner's data, restrict it to the
owner or an admin for root and subfolders alike, replacing the root/subfolder
split with a single check. Owners and admins are unaffected, and a
write-collaborator can still create, rename and add to shared folders and delete
subfolders they own.
Co-authored-by: legobattman <302282032+legobattman@users.noreply.github.com>
* style: condense the folder deletion authorization comment
Shorten the multi-line comment above the owner-or-admin check to a single line stating why deletion is restricted. The full rationale lives in the pull request description and does not need to be narrated in the code.
---------
Co-authored-by: legobattman <302282032+legobattman@users.noreply.github.com>
KatexRenderer rendered the raw math source through {@html} whenever renderToString threw. throwOnError only suppresses KaTeX ParseError, so a RangeError (maximum call stack size exceeded, reachable with deeply-nested brace input) escaped into the catch and re-exposed the unescaped source. Because the math tokenizer captures everything between the delimiters verbatim, that source can carry an HTML/JS payload which then executed in the viewer's browser on the application origin, a stored, cross-user XSS reachable through normal chat/channel/shared-chat rendering. Escape the fallback so the source is shown as text and is never injected as HTML. Valid math is unaffected, it still renders through the success path.
Co-authored-by: maxntv <maxntv@users.noreply.github.com>
ConfirmDialog trapped focus and closed on Escape but its container was a plain
div, so screen readers did not announce it as a modal dialog. Its text input
also had only a placeholder, giving no persistent accessible name. Add
role=dialog / aria-modal / aria-label / tabindex to the dialog surface and an
aria-label to the textarea.
Relates to #2790
Co-authored-by: Tim Baek <tim@openwebui.com>
Adds a multiselect input type for Valves and UserValves so plugin authors can let users pick multiple values from static or runtime-resolved options instead of maintaining comma-separated text fields with hardcoded allowed-value lists in the description.
ENABLED_ITEMS: list[str] = Field(
default=["foo"],
json_schema_extra={"input": {"type": "multiselect", "options": "get_item_options"}},
)
@classmethod
def get_item_options(cls):
return [{"value": "foo", "label": "Foo"}, {"value": "bar", "label": "Bar"}]
Options accept the same shapes as the existing select input: either a static list (strings or {value, label} dicts) or a classmethod name resolved at request time (including __user__ context for UserValves). No backend changes are needed because resolve_valves_schema_options already resolves options independently of the input type.
The new MultiSelect component follows the existing Select portal dropdown pattern and renders checkbox rows that stay open while toggling, with the selected labels shown in the trigger. Values bind as a real string array end to end: the array-to-comma-string conversions in the chat controls valves panel and the valves modal are skipped for multiselect fields, so the stored valve is a native list[str] validated by Pydantic.
Requested in #26848.
_sanitize_proxy_path decodes the path and then relies on posixpath.normpath plus a leading '..' check. posixpath splits on '/' only, so a backslash run is treated as part of a single path component: 'foo/..\..\etc' normalizes to itself, does not start with '..' and is forwarded unchanged, reaching the upstream as '/foo/..%5C..%5Cetc'. An upstream that treats the backslash as a separator would resolve those '..' sequences.
Reject any path containing a backslash after decoding, matching the existing fail-closed behaviour for paths that are still encoded past the decode cap. A backslash is not meaningful in the upstream API paths this route proxies, so legitimate requests are unaffected.
Co-authored-by: babakizo420 <babakizo420@users.noreply.github.com>
stream_wrapper without a content handler iterates aiohttp's response.content, which reads line by line: every line costs a buffer scan, a slice, a bytes concat, a generator resume and its own ASGI response message. A typical SSE event is two lines (the data line and the blank separator), so every upstream token event became two yields and two transport writes even on routes where the body is never inspected.
stream_wrapper now takes passthrough=True, which iterates response.content.iter_any(): the exact same bytes, one yield per network read, no line scanning. It is applied only to routes no internal consumer parses line-by-line: the ollama pull/push/create/generate proxies and its v1 completions, chat completions, messages and responses endpoints, plus the openai embeddings, responses and catch-all proxies. The two internally consumed chat routes keep line iteration, which the streaming middleware and the Ollama-to-OpenAI converter require; the ollama send_request signature documents that constraint.
Benchmark (local aiohttp SSE server, 500 events, consumed through stream_wrapper):
| metric | before (readline) | after (iter_any) |
| --- | --- | --- |
| stream consumption time | 1.46 ms | 0.62 ms |
| generator yields + response writes per stream | 1000 | 1 |
The single yield is a loopback artifact (the whole body arrives in one buffered read); over a real network it becomes one yield per TCP read instead of two per SSE event.
Functionally verified: line mode and passthrough mode produce byte-identical output for the same stream, and passthrough always yields fewer, larger chunks.
_process_picture_url validated the picture URL with validate_url() but then fetched it with a plain aiohttp session that resolves the hostname again at connect time, leaving a DNS-rebinding TOCTOU window (the same gap already closed for the RAG loader, the content probe, the image fetches and webhook delivery). Routing the fetch through get_ssrf_safe_session() pins the connect-time resolution via _SSRFSafeResolver and rejects non-global addresses, so a rebinding host can no longer redirect the fetch to loopback, RFC1918 or cloud-metadata endpoints. It also stops the forwarded OAuth access_token from leaking to a rebound internal target.
* chore: drop redundant background repaints so surfaces inherit their parent
Four spots repaint the exact color their parent surface already provides
(bg-white / dark:bg-gray-900 rows inside same-colored pages and modals,
and the selectClass dark repaint inside the connection modals — the
sibling input const is already fully transparent). Visually identical in
stock light and dark; removing them lets instance theming show through
instead of leaving opaque boxes:
- .tiptap tr (app.css) — table rows in notes/editors
- Edit User Group Users tab body rows (common Modal surface)
- AddToolServerModal + AddTerminalServerModal selectClass dark repaint
The matching repaints inside the ModelUsage/UserUsage components are
not part of this change — those files were dead code and were removed
entirely in #27574.
* chore: catch remaining redundant surface repaints missed in the first pass
Same rule as the previous commit — every one of these repaints the exact
color its parent surface already provides, so removal is stock-identical
in light and dark while letting instance theming show through:
- Analytics Dashboard's inline Model Usage / User Activity row markup
(the Analytics tab renders these tables from Dashboard.svelte itself;
the unreferenced ModelUsage/UserUsage component files were removed
in #27574)
- Evaluations Feedbacks + Leaderboard body rows (settings modal surface)
- admin UserList body rows (app page surface)
- chat markdown tables (MarkdownTokens): thead and body rows — unlike
the tiptap header (gray-850 contrast, untouched), this thead painted
the page's own color
- CitationsModal source rows (common Modal surface)
- AddConnectionModal selectClass dark repaint — third copy of the same
const already fixed in AddToolServerModal / AddTerminalServerModal
get_ollama_versions was the only Ollama route besides the static health check without an authentication dependency, so an anonymous caller could read the configured backend's version string and, by walking url_idx until the lookup raised, count the configured backends.
Nothing depends on the route being public. The frontend wrapper takes a token and sends it on every call, and its three call sites (admin model management, the model selector and the About panel) all pass an authenticated token, so the client already treats this as an authenticated route. Add the same get_verified_user dependency the sibling routes carry.
Co-authored-by: Grg0rry <Grg0rry@users.noreply.github.com>
Nothing in the tree imports either component; the admin Analytics tab
renders its own inline copies of both tables directly from
Dashboard.svelte. Both files landed with the dashboard in a4ad34841
(feat: analytics frontend dashboard) but were never wired into it.
The remaining name matches elsewhere (the getUserUsage API and
UserUsage* types in src/lib/apis/users/index.ts, consumed by
chat/Settings/Usage.svelte, plus the backend usage endpoints) belong to
the unrelated per-user usage feature and are untouched.
* feat: add default upload mode setting
Add user setting to configure the default upload mode for files, allowing users to choose between "Using Entire Document" (full context) and "Using Focused Retrieval" (RAG processing) as the default behavior.
* i18n: sync locale catalogs for the new upload mode strings
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* chore: re-trigger CI (previous run hit the pre-existing Node heap OOM, see #27254)
* fix: apply the default upload mode at upload time so the payload carries it
The previous approach only pre-set the modal toggle's visual state on
mount; item.context is written solely by the Switch's on:change, so the
sent files kept context: undefined and the backend never saw 'full'. It
also showed a misleading ON state for legacy context-less files, since
FileItemModal mounts with every FileItem chip render.
Stamp context on the fileItem in uploadFileHandler instead (before
...itemData, so callers passing an explicit context still win) and
revert the FileItemModal hunk — the modal already renders from
item.context alone.
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
aiohttp resolves every hostname with ThreadedResolver unless the aiodns package is importable, and ThreadedResolver runs socket.getaddrinfo on asyncio's default ThreadPoolExecutor. That executor is capped at min(32, cpu_count + 4) threads and is shared with every other piece of blocking work posted to it, so DNS is currently a bounded blocking resource sitting in front of every model call, every web search fetch, every RAG page load and every tool call. In plain terms: once that pool is busy, requests wait on name lookups that should never have occupied a thread at all.
This is a dependency-only change. aiohttp sets `DefaultResolver = AsyncResolver` as soon as aiodns is importable (aiohttp/resolver.py), so resolution moves onto the event loop via c-ares with zero application code touched. That is deliberate rather than lazy: there are 50 `aiohttp.ClientSession(...)` construction sites in the backend, most building a fresh default connector per call, and the alternative of passing `resolver=aiohttp.AsyncResolver()` explicitly would mean touching all of them and re-touching every future one. The shared pool in `utils/session_pool.py` does set `ttl_dns_cache`, but that only helps the shared pool. Every per-request session, including `SafeWebBaseLoader._fetch()` which builds a new session per URL, starts with a cold DNS cache and resolves from scratch.
It also unbreaks a code path that is dead today. `backend/open_webui/retrieval/loaders/mistral.py:480` constructs `aiohttp.AsyncResolver()` unconditionally, and `AsyncResolver.__init__` raises `RuntimeError("Resolver requires aiodns library")` when aiodns is absent, so the Mistral OCR content extraction engine fails on a stock install. This supplies the dependency that line already assumes. Once it is present that kwarg is redundant, since it now names the default, and dropping it is a reasonable follow-up. Reproduced by blocking the aiodns import:
```
aiodns importable: False
DefaultResolver: ThreadedResolver
AsyncResolver(): RuntimeError: Resolver requires aiodns library
```
## Benchmarks
Both sides run the real aiohttp resolver classes. The DNS wire time is replaced by an identical fixed 50ms delay on both sides, so the only variable measured is where that delay is spent. 24 cores, so the default executor holds 28 threads. `exec_max` is the worst latency an unrelated `run_in_executor` job suffered while the lookups were in flight.
Concurrent lookups, wall time:
| concurrent lookups | ThreadedResolver | AsyncResolver | speedup | exec_max before | exec_max after |
|---|---|---|---|---|---|
| 16 | 54.3ms | 41.0ms | 1.3x | 2.1ms | 1.9ms |
| 32 | 101.9ms | 50.6ms | 2.0x | 36.6ms | 1.8ms |
| 64 | 152.5ms | 43.2ms | 3.5x | 88.4ms | 2.0ms |
| 128 | 254.9ms | 44.5ms | 5.7x | 190.3ms | 2.2ms |
| 256 | 508.2ms | 50.7ms | 10.0x | 443.8ms | 2.4ms |
| 512 | 965.2ms | 47.8ms | 20.2x | 900.2ms | 2.6ms |
ThreadedResolver scales linearly with concurrency because it can only run 28 lookups at a time. AsyncResolver stays flat at roughly the cost of one lookup.
The reverse direction is worse and is not hypothetical. Open WebUI already posts long blocking jobs to that same executor (`retrieval/vector/dbs/pinecone.py:323` batch upserts, `retrieval/loaders/youtube.py:156` transcript loads). With 28 such jobs holding the pool, a single DNS lookup waits for them to finish:
| | one DNS lookup |
|---|---|
| ThreadedResolver | 1989.7ms |
| AsyncResolver | 58.9ms |
A Pinecone bulk upsert currently stalls name resolution for every other user on the instance. After this change it cannot.
At low concurrency on a real network the two are equivalent, as expected: 8 concurrent lookups against disjoint cold hostname sets landed within noise of each other in both directions.
## Behaviour verification
Checked against Open WebUI's own code, not in isolation:
- c-ares reads the system hosts file. Verified against a machine whose hosts file maps `adobe.io` to `0.0.0.0`, an address real DNS never returns for that name: c-ares returned `0.0.0.0`. `host.docker.internal`, compose `extra_hosts` and Kubernetes `hostAliases` keep working.
- `_SSRFSafeResolver` subclasses `aiohttp.resolver.DefaultResolver`, so this change swaps its base class from ThreadedResolver to AsyncResolver at runtime. It still resolves public hosts, still returns entries with the `host`/`port` keys the SSRF check reads, and still raises on a private address: resolving `localhost` raised `ValueError: The URL you provided is invalid.`
- A real fetch through `get_ssrf_safe_session()` returned 200.
- NXDOMAIN still surfaces as `aiohttp.ClientError` (`ClientConnectorDNSError`), not a c-ares specific exception, so existing error handling is unaffected.
Known limit: c-ares reads `/etc/resolv.conf` and the hosts file but not the rest of `nsswitch.conf`. Names served only by an NSS module, such as `.local` via avahi/mDNS, NIS/LDAP backends or Windows NBNS, will resolve differently or not at all. On a multi-homed test machine the local hostname returned two addresses through the system resolver and one through c-ares. Deployments pointing Open WebUI at an mDNS or NetBIOS hostname are the group affected. Resolver failures also arrive as plain `OSError` rather than `socket.gaierror`, which no code in this repo catches today.
Custom per-connection headers can now forward the user's groups to
upstream backends via two new template placeholders:
- {{USER_GROUPS}}: comma-separated group names
- {{USER_GROUP_IDS}}: comma-separated group ids
The group lookup is async, so get_custom_headers becomes an async
wrapper around the sync template substitution (parse_custom_headers)
and fetches groups lazily — only when a header value actually
references a groups placeholder. The external document loader path
runs in a worker thread without an event loop, so Loader.aload
prefetches the groups before offloading and passes them through to
ExternalDocumentLoader.
Claude-Session: https://claude.ai/code/session_01EbBEfTyu8fFJmC13rnQthT
Co-authored-by: Claude <noreply@anthropic.com>
Chat search built each result row with ChatTitleIdResponse(**chat.model_dump(), ...), which recursively copies the entire chat blob per row only for the constructor to ignore everything except id, title and timestamps: a 60-row search page deep-copied up to 60 full conversations. The folder listing, archived and export endpoints and every single-chat response did the same dump-and-revalidate dance via ChatResponse(**chat.model_dump()).
Search rows are now built from the five fields the response actually has (the snippet helper receives the blob by reference as before), and all 18 ChatResponse constructions use ChatResponse.model_validate(chat, from_attributes=True), which reads the fields off the already-validated ChatModel without copying the blob.
Benchmark (~500 KB chat blob):
| metric | before | after |
| --- | --- | --- |
| search result row | 0.05 ms | 0.003 ms |
| ChatResponse construction | 0.05 ms | 0.003 ms |
| per search page (60 rows) | 3 ms | 0.2 ms |
Beyond CPU, each converted row also stops materializing a second full copy of the conversation in memory while the page is being built.
Functionally verified: both construction styles produce identical model_dump() output for ChatResponse (including defaulted fields absent on ChatModel) and for search rows including the snippet.
DELETE /api/v1/chats/{id} called stop_item_tasks(id) before checking the
caller's chat.delete permission or ownership of the target chat. An
authenticated user who knew another user's chat id could therefore cancel that
chat's in-flight generation (streaming response, title or tag generation) even
though the deletion was then rejected. The chat id is discoverable through
legitimate read-only access to a shared chat or folder.
Reorder the handler to authorize first (admin, or owner holding chat.delete) and
only then cancel tasks and delete, matching the dedicated task-stop endpoint.
Legitimate deletions are unchanged; an unauthorized caller now returns 404 or 401
before any cancellation. The duplicated tag-cleanup and event-publish blocks are
merged.
Co-authored-by: GabrielGomesAL <193945687+GabrielGomesAL@users.noreply.github.com>
Completes the contrast set after #27555, #27554 and the gray-500 branch, which between them cover text-gray-400, text-gray-500, dark:text-gray-600 and placeholders. This is everything still under 4.5:1 after those.
The grey scale in src/tailwind.css is achromatic oklch(L 0 0), so relative luminance is exactly L³. What is left:
- text-gray-300 dark:text-gray-700, the lightest muted pair, at 1.58:1 in light and 2.14:1 in dark. Used for inactive tab labels across the admin, workspace and playground layouts, breadcrumb separators and empty state hints. It resolves to gray-600 in light and gray-400 in dark.
- text-gray-400/70 on the embedded chat history dropdown icon, 2.07:1 against the 3:1 that WCAG 1.4.11 requires of icons.
- Hover states that land lighter than the new resting colour. Once the resting state is gray-600, an element hovering to gray-500 gets less readable on interaction rather than more, so hover and group-hover targets of gray-500 resolve to gray-800. Sidebar/Section.svelte and the citation modal links are the sites this affects.
- The autocompletion ghost text in src/app.css, hardcoded #a0a0a0, 2.65:1 in light. The dark canvas already passes.
- The shimmer used for loading text, a #b4b4b4 gradient clipped to the glyphs at 2.10:1 in light. There is no solid colour to raise, so with the setting on it renders as flat gray-700 text instead.
Everything above is gated on the existing High Contrast Mode setting and changes nothing when it is off. No markup is touched, so this is src/app.css only.
Deliberately left alone: disabled: variants, since WCAG 1.4.3 exempts inactive components; the FileNav breadcrumb ancestors, which are non-clickable; decorative folder icons; and text-gray-100, dark:text-gray-800 and dark:text-gray-900, which are inverse text on filled buttons and already high contrast against their own backgrounds. The ad-hoc dark:text-gray-800 pairs in ChannelModal.svelte and automations/+layout.svelte stay as they are; dark:text-gray-800 doubles as the inverse text on the white buttons in Message.svelte, ResponseMessage.svelte and UserMessage.svelte, so it cannot be remapped in CSS without breaking those.
Not fixed here, and a genuine follow-up: .hljs-comment in src/app.css is #616161, roughly 3:1 on the dark code background. It sits outside the Tailwind grey scale and needs a highlight.js theme override rather than a utility remap.
Verified in a browser against Tailwind's emitted rules and layer order: with the setting on, the lightest pair resolves to gray-600 in light and gray-400 in dark, the hover and group-hover targets to gray-800, ghost text to gray-600 and the shimmer to solid gray-700, while inverse button text and every dark hover variant stay where they are; with the setting off nothing changes in either theme.
On latest `dev`, the `Select` trigger sets `aria-label={placeholder}`. In the accessible name computation `aria-label` is evaluated before the element's contents, so on a button that renders visible text it **replaces** that text instead of adding to it.
The trigger's content is `selectedLabel`, which resolves to the selected item's label and only falls back to `placeholder` when nothing is selected. So a control visually reading "Week" is exposed to assistive technology as "Select view", permanently, no matter what is selected. The dropdown items expose no `aria-selected` either, the current one is marked with a check icon only, so there is no path by which a screen reader user can find out what the control is set to.
Breaks WCAG 2.5.3 Label in Name (Level A), because the accessible name does not contain the visible label, so voice control cannot target the control by what it says on screen. Also 4.1.2 Name, Role, Value (Level A), because the value is never exposed.
Fix: drop the overriding `aria-label` so the name is computed from the visible text, and expose `aria-expanded` so the open state is conveyed. All 7 call sites plus the `DropdownOptions` wrapper override `slot="trigger"` and every one of them renders `selectedLabel` (or `placeholder`) as text, so no trigger is left unnamed. The two call sites that pass no `placeholder` were already emitting an empty `aria-label`, which is skipped by the name computation, so they are unaffected.
`aria-haspopup` is deliberately not added: the popup is a plain `DropdownMenu` of buttons with no `listbox` role, so claiming one would misdescribe it.
`placeholder` is still used, it drives the `selectedLabel` fallback and `TagSelector` renders it directly.
Severity: Serious. Affects every custom select in Admin Settings, Workspace and the calendar and automations pages.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, the five sortable column headers in the admin Users table are click handling `<th>` elements:
```svelte
<th scope="col" class="px-2.5 py-1.5 font-normal cursor-pointer select-none" on:click={() => setSortKey('name')}>
```
A `<th>` is not interactive. There is no `<button>`, no `tabindex`, no `role` and no key handler, so **sorting the user list is impossible without a mouse**. The sort direction is also conveyed only by an 8×8 pixel chevron, with no programmatic state, so assistive technology cannot report which column is sorted or in which direction.
Breaks WCAG 2.1.1 Keyboard (Level A) and 4.1.2 Name, Role, Value (Level A).
Fix: move the click handler onto a real `<button>` inside the header, which brings native focus, Enter and Space activation and the correct role, and add `aria-sort` to the `<th>`, which already carries `scope="col"` and therefore the implicit `columnheader` role. Only the active column reports a direction, since `orderBy` is a single value; the non sortable actions column deliberately gets no `aria-sort` at all rather than `none`, so it is not advertised as sortable.
The cell padding moves from the `<th>` onto the button so the whole header stays clickable. Left on the `<th>`, the padding ring would have become a dead zone, shrinking the hit target and flipping the cursor at an invisible boundary inside the header.
`cursor-pointer` is dropped from the `<th>` because `src/tailwind.css` already applies it to every `button`.
The repeated `aria-sort` ternary is extracted to a small `sortState` helper rather than pasted five times.
The same mouse only `<th on:click>` pattern still exists in the Analytics, Evaluations and Groups tables and is not touched here.
Severity: Serious. A core admin function is unreachable without a pointing device.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
`_normalize_token_expiry()` derived the session expiry from the access token alone, and that value is what `oauth_session.expires_at` stores and what both `get_oauth_token()` implementations check to decide whether to refresh five minutes ahead. Providers that issue a shorter-lived id_token than access token (Microsoft Entra ID: roughly 60 minutes against 75) therefore left a window where the session still looked valid while the id_token had already expired, so pipes and tools reading `__oauth_token__["id_token"]` forwarded a dead JWT and downstream services rejected it with 401.
The stored expiry is now capped at the id_token's `exp` claim whenever that JWT expires first, which moves the refresh ahead of the earliest expiring token in the set. This is applied in the single function every session write already passes through, so it covers both the SSO manager and the MCP client manager on their callback and refresh paths alike. Sessions without an id_token, with an opaque one, or with no `exp` claim are unaffected.
Fixes#27066
When an audio file is split into multiple chunks for transcription, transcribe() collected the per-chunk results with asyncio.as_completed(), which yields results in completion order rather than submission order. Whenever a later chunk finished transcribing before an earlier one, the assembled transcript was scrambled, for example the second half of a recording appearing before the first, and the stored file content plus everything downstream (file preview, full-context retrieval) read out of chronological order. This change awaits the chunk tasks with asyncio.gather() instead, which runs them just as concurrently but returns the results in the order the tasks were created, i.e. chunk_paths order. The existing error handling and chunk cleanup are unchanged: an HTTPException from a chunk is re-raised as is and any other error is wrapped in a 500. Fixes#27143
* feat: prevent duplicate auth form submissions while one is pending
When a sign in, sign up or LDAP request is slow, the auth form can be submitted again and every extra click or Enter press starts another concurrent authentication request. The form never tracked a pending state, so submitHandler dispatched a new API call on every submit event. This adds a submitting flag that makes submitHandler ignore re-entrant submits, disables both submit buttons with a dimmed style while a request is in flight and resets the flag in a finally block so the form recovers after a failed attempt. Guarding submitHandler covers button clicks and Enter key submits for sign in, sign up and LDAP alike since all of them flow through the single form submit handler.
Fixes#27264
* feat: show a spinner while an auth request is pending
Disabling the submit button stops a second submission but gives no positive sign that the first one is still running, so on a slow identity provider the form looks unresponsive rather than busy. Both submit buttons now render the existing Spinner next to their label while submitting is set, following the same in-button pattern used by the workspace editors.
When a filter's outlet() modified the structured assistant output in place, the change was shown immediately but lost after reload. outlet_filter_handler built its outlet payload with a shallow reference to the message's output list from messages_map, so the filter mutated the stored baseline itself and the subsequent output comparison compared the object against itself, never detecting a change and never persisting it. The same aliasing corrupted originalContent for messages whose text lives only in output. Deepcopy the output when building the outlet payload so messages_map stays a pristine pre-filter baseline and the existing change detection persists outlet-modified output through the existing upsert path. Fixes#27017.
* fix: show download preparation toast and prevent duplicate zip jobs
Downloading a file or folder from the file navigator gave no feedback while the server prepared the response, which can take 30 seconds or more for large folders that are zipped server-side. Users assumed the click did nothing and pressed Download again, starting additional zip jobs on the server. Both downloadFile and bulkDownload, the two functions every download control funnels through, now show a persistent "Preparing download..." loading toast while the request is in flight and dismiss it once the download starts or fails. A shared downloading flag makes repeated clicks no-ops until the current download finishes, so a single click starts exactly one server-side job. Fixes#27055
* fix: report terminal download failures instead of dismissing the toast
A failed download dismissed the preparation toast without saying anything, which reads as the download silently disappearing. Both download paths now report the failure. The two helpers also declare a nullable return but could still reject once the response body started streaming, so an interrupted transfer escaped as an unhandled rejection and left the same silent dismissal. They now return null in that case, which also stops an interrupted preview from leaving its spinner running.
Add self-hosted OpenSERP as a web search engine option. OpenSERP
provides browser-rendered search across Google, Bing, Yandex, Baidu,
DuckDuckGo, and Ecosia with no API keys required.
- New module: retrieval/web/openserp.py (async, uses aiohttp session pool)
- Config: OPENSERP_BASE_URL env var (defaults to http://localhost:7070)
- Routing: search_web() dispatch for 'openserp' engine
- Follows existing patterns (searxng, brave)
Co-authored-by: crustopher-lgtm <crustopher-lgtm@users.noreply.github.com>
Every other navigation entry in the user menu (Settings, Admin Panel,
Archived Chats, Workspace, Notes, Automations, Playground, Sign Out)
collapses the sidebar on mobile after navigating, but the Calendar entry
was missing this handling, leaving the sidebar open over the Calendar
page on mobile. Add the same mobile guard used by the sibling entries.
When the same model is selected multiple times in a side-by-side chat,
each response is created with a distinct modelIdx (0,1,2,3) that
identifies its column. The backend now owns message persistence, but it
built the assistant placeholders without modelIdx, so the field was
never saved. On reload MultiResponseMessages groups responses by
modelIdx and falls back to grouping by model id when modelIdx is
missing; with duplicate models that fallback lumps every response into
each column, so all columns render the first response (and show a bogus
'1/N' pager).
Send modelIdx with each message_ids entry from the frontend and persist
it on the assistant placeholders in both the new-chat and existing-chat
paths. The message_ids list is now forwarded for every send (not just
multi-model ones) so single-column regenerations in a duplicate-model
chat also keep their column identity across reloads.
The automation create/edit modal footer laid the schedule/model dropdowns
and the Cancel/Create actions in a single justify-between row. On narrow
modals the fixed-width actions left too little room for the dropdowns,
which wrapped to two stacked lines with Cancel squeezed in the middle.
Stack the footer vertically on mobile (dropdowns row, then a right-aligned
actions row) and restore the horizontal layout at sm and up.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
The OAuth / OIDC section in Admin Settings > Authentication had no
enable/disable switch, unlike the LDAP section above it. Add one that
persists via the existing Save flow and actually gates OAuth sign-in,
mirroring how the LDAP toggle works.
- config: new ENABLE_OAUTH persistent config ('oauth.enable'), defaulting
to True so existing deployments with a provider configured keep working.
- oauth: expose ENABLE_OAUTH via the OAuth runtime config and reject the
login and callback handlers with 404 when it is disabled.
- /api/config: report no OAuth providers when disabled so the login page
hides the OAuth buttons (and cannot auto-redirect), without clearing the
admin's provider configuration.
- auths: expose ENABLE_OAUTH through the admin OAuth config get/update
endpoints (OAuthConfigForm + OAUTH_CONFIG_KEYS).
- Authentication.svelte: bind the OAuth / OIDC header Switch to the
persisted oauthConfig.ENABLE_OAUTH and collapse the section when off,
matching the LDAP header (size, weight, alignment).
Follow-up to #27495, reopened as a high contrast mode change. Builds on the `high-contrast` class landed in #27555.
Muted UI text is written as text-gray-400 dark:text-gray-600. The grey scale in src/tailwind.css is achromatic oklch(L 0 0), so relative luminance is exactly L³: text-gray-400 is 2.07:1 on white and dark:text-gray-600 is 3.12:1 on #171717. The pair is effectively inverted, and both halves fail the 4.5:1 required by WCAG 1.4.3, with the light value also failing the 3:1 required of icons under 1.4.11. This is the text used for settings and admin section headings, field descriptions, sidebar labels, timestamps and counters, all at 10px to 12px, so the large-text exemption does not apply.
Rather than rewriting the class literal at 251 sites, the remap is two CSS rules that only apply when the existing High Contrast Mode setting is on, so the default theme is untouched:
- text-gray-400 resolves to gray-600 (5.75:1) in light mode
- dark:text-gray-600 resolves to gray-400 (8.65:1) in dark mode
dark:text-gray-500 already passes at 6.46:1 and is left alone.
The rules live in `@layer utilities` and use `:where()` to stay at low specificity: they outrank the base utility but lose to `hover:` and `dark:hover:` variants, so hover feedback keeps working. Verified in a browser against Tailwind's emitted rules and layer order: with the setting on, resting text resolves to gray-600 in light and gray-400 in dark, hover still resolves to its own value, and with the setting off nothing changes in either theme.
One component change is required alongside it. In admin/Settings/Audio.svelte the help text puts links inside the muted block via `[&_a]:text-gray-600`; once the surrounding prose resolves to gray-600 the link becomes the same colour as the text it sits in, and it has no resting underline, which would be a new WCAG 1.4.1 failure. The link moves to gray-900. This is the only place in the codebase where a link colour is nested inside muted text.
Letting the variants win has one edge: a few elements hover to a grey lighter than their new resting colour, so hovering would have lowered contrast instead of raising it. A light-mode hover landing on gray-500 now resolves to gray-800, which keeps the hover darker than the gray-600 resting state. Sidebar/Section.svelte is the site this branch would otherwise break.
Not covered here: text-gray-500 dark:text-gray-400 (2.77:1 in light), which is a separate branch.
With ENABLE_PERSISTENT_CONFIG=False the admin's model order is reset on every restart because ui.model_order_list falls back to its DEFAULT_CONFIG default, and unlike every other Models setting (DEFAULT_MODELS, DEFAULT_PINNED_MODELS, DEFAULT_MODEL_METADATA and DEFAULT_MODEL_PARAMS) that default was hardcoded to an empty list with no environment variable to source it from. This adds a MODEL_ORDER_LIST environment variable parsed as a JSON array using the same guarded pattern as the neighbouring DEFAULT_MODEL_METADATA and DEFAULT_MODEL_PARAMS defaults, falling back to an empty list on parse errors. Behaviour when the variable is unset is unchanged.
Fixes#27206
`_detect_text_encoding()` hands the complete file to `chardet.detect()`. chardet is pure Python and costs roughly 1.3 seconds per megabyte, so uploading a large non-UTF-8 text file stalls for seconds inside encoding detection alone. A 4 MiB Shift-JIS file spends 6.4 seconds there. The UTF-8 fast path above it means only non-UTF-8 files reach this, which in practice are exactly the CJK documents the surrounding code was written to handle, so the slow case and the case that matters are the same case.
Detection does not need the whole file. It needs the bytes that are actually not UTF-8, and `UnicodeDecodeError.start` from the fast-path decode already says where those begin, so this samples a 256 KiB window around that offset.
Two things make that safe rather than merely fast.
Centring the window on the first non-UTF-8 byte instead of the file head is what keeps the common case correct. A plain head sample makes chardet report ascii for a file that is ASCII for its first few hundred KiB and only turns CJK later, and the method then falls through to latin-1 instead of the right codec.
The window still cannot help when a stray byte, a pasted Windows-1252 artifact for example, sits hundreds of KiB ahead of the real payload: the sample is then almost pure ASCII and carries no signal. So when the sample holds almost no non-ASCII bytes and is a strict subset of the file, detection falls back to the whole buffer. That case pays the old cost, which is the right trade, because it is precisely the case where sampling would otherwise be wrong. Without this guard a Cyrillic document with a stray leading byte was detected as ISO-8859-1 rather than windows-1251, which is silent mojibake.
Measured, with the encoding returned identical in every case:
| file | before | after |
|---|---|---|
| shift_jis 4 MiB | 6402ms | 755ms |
| gb18030 4 MiB | 3199ms | 449ms |
| big5 4 MiB | 2926ms | 413ms |
| euc-jp 4 MiB | 2456ms | 413ms |
| euc-kr 4 MiB | 2382ms | 468ms |
| latin-1 4 MiB | 1902ms | 394ms |
| gb18030 1 MiB | 807ms | 376ms |
| ascii head then gb18030 tail | 533ms | 294ms |
| stray byte then cp1251 payload | 496ms | 1051ms |
| any UTF-8 file | 8ms | 0ms |
29 cases, all returning an identical encoding before and after: six encodings at 100 KiB, 1 MiB and 4 MiB, three layouts where the non-UTF-8 bytes only begin beyond the window, four where a stray byte is separated from the payload, plus plain UTF-8, UTF-8 CJK and an empty file. The stray-byte rows are slower than before because they scan twice, once over the window and once over the whole buffer. They are the pathological shape, and correctness wins there.
The residual time is now the decode-and-validate loop below, which walks the file once per candidate codec, and `_has_cjk_characters`, which is a per-character Python loop over the decoded text. Both are the same "full scan for a detection decision" pattern and could take a bounded prefix too. That is left alone here.
`alazy_load()` builds every BeautifulSoup tree inline in an async function, so a web search that pulls in ten pages stops the entire worker for the whole time it spends parsing. Nothing else on that worker runs during it: not other users' token streams, not health checks, not socket.io traffic. Parsing is CPU work and it belongs in a thread.
Measured over 37 real pages, 13.5 MiB total, with a 5ms ticker sampling event-loop lag:
| | wall | worst loop stall | ticker fired |
|---|---|---|---|
| inline, html.parser (today) | 1793.8ms | 1788.8ms | 1 time |
| offloaded, html.parser | 1872.9ms | 82.9ms | 88 times |
| inline, lxml | 1346.7ms | 1341.8ms | 1 time |
| offloaded, lxml | 1445.4ms | 37.0ms | 118 times |
Today the loop is not merely slow during a batch, it is gone: a 5ms timer fired exactly once across 1.8 seconds. After the change it fires normally and the worst single stall drops by a factor of 20 to 36. The cost is 4 to 7 percent more wall time for the batch itself, from the thread handoffs, which is the right trade for a server handling more than one user.
Three details behind the shape of the change:
`get_text()` is only 2 percent of the cost (34ms against 1706ms of parsing over the corpus), so the whole per-page unit moves into the thread rather than the parse alone. Splitting them measured worse on both axes.
The offload is per page, not per batch. Handing the whole batch to one thread measured worse than either (2081ms wall, 235ms worst stall), so the loop is yielded to between pages.
The metadata block in `alazy_load()` was a duplicate of the module-level `extract_metadata()`, field for field, and `lazy_load()` was already using the shared helper. The new helper calls it too, which is why the diff removes more lines than it adds. The `ascrape_all()` override goes with it: it was a verbatim copy of the inherited implementation and `alazy_load()` was its only caller, so anything still calling it now gets the identical parent method, which resolves `self._unpack_fetch_results` to the override this class keeps.
Verified by feeding the real loader a 37 page corpus and comparing every resulting Document against the implementation this replaces:
```
PASS one Document per url (37)
PASS every Document identical to the pre-change implementation (0 differ)
PASS parsing ran off the main thread
PASS event loop kept running during parsing (90 ticks)
```
Both `page_content` and `metadata` are byte-identical on all 37 pages. This is independent of the parser in use and composes with switching the default parser to lxml: that change makes the stalls shorter, this one takes them off the loop.
Regenerated with the current uv so `uv lock --check` passes again. The committed lockfile was written by an older uv and every run since has reported it as needing an update, which makes it impossible to tell real drift apart from format drift.
No resolved dependency changes: all 354 packages keep their versions, and no existing artifact URL or hash changes. The diff is almost entirely `upload-time` annotations added per artifact. The remaining changes are the `revision = 3` format marker, a `provides-extras` entry on the project stanza, additional GraalPy wheel URLs for already-locked versions of jiter, pybase64, pydantic-core and ujson, and the removal of the hardcoded `version = "0.10.2"` from the project's own stanza, which was stale metadata since pyproject.toml declares `dynamic = ["version"]`.
Nothing that gets installed changes: the Dockerfile installs from backend/requirements.txt and no workflow runs `uv lock` or `uv sync`.
OpenAPI specs with circular schema references, such as Mealie's where Recipe and RecipeCategory reference each other through properties and array items, crashed convert_openapi_to_tool_payload with a RecursionError, so the tool server produced no specs and the integration never appeared in the model or tool selection. resolve_schema already had a visited-set guard against circular references, but the recursive calls for properties and items dropped the set, so cycles running through those edges were never detected. This threads the visited set through those calls and passes a per-path copy when following a $ref, so only true ancestor cycles are pruned to an empty schema while sibling references to the same schema still resolve fully. Fixes#27239.
Every page pulled in by web search and web RAG is parsed with BeautifulSoup's `html.parser`, a pure-Python parser. It is the slowest option bs4 offers, and it is being handed 300 KiB to 1.5 MiB documents, several per query. `SafeWebBaseLoader` inherits `default_parser = "html.parser"` from langchain's `WebBaseLoader` and never overrides it, so this is an upstream default carried by accident, not a decision anyone made for Open WebUI.
`default_parser` is the single chokepoint for both the sync `_scrape()` path and the async `ascrape_all()` path, so one `setdefault` covers everything and an explicit caller override still wins.
lxml is already in the tree as a transitive hard dependency of ddgs, python-pptx and unstructured, so nothing new enters the image and `uv.lock` already resolves it at 6.1.1. The pin makes it explicit and closes a latent failure: bs4's `"xml"` feature, already used for `.xml` URLs in `_unpack_fetch_results()`, requires lxml and would raise `FeatureNotFound` the day that transitive dependency moves.
## Benchmarks
37 real pages, 13.8 MiB of HTML, median of 5 runs each. The timed operation is `BeautifulSoup(html, parser)` plus `get_text()` plus `extract_metadata()`, which is exactly what the loader does per page. bs4 4.14.3, lxml 6.1.1, CPython 3.12.
| | html.parser | lxml | |
|---|---|---|---|
| 37 pages, 13.8 MiB total | 1611.0ms | 1151.3ms | 1.4x faster, 460ms saved |
Largest pages:
| page | size | html.parser | lxml | speedup |
|---|---|---|---|---|
| pypi.org/project/aiohttp/ | 1259 KiB | 243.75ms | 180.51ms | 1.4x |
| gnu.org/software/bash/manual/bash.html | 1017 KiB | 257.97ms | 178.99ms | 1.4x |
| rfc-editor.org/rfc/rfc9110.html | 1157 KiB | 205.94ms | 154.87ms | 1.3x |
| docs.aiohttp.org/en/stable/client_reference.html | 403 KiB | 108.62ms | 84.93ms | 1.3x |
| ollama.com/library | 779 KiB | 117.55ms | 73.64ms | 1.6x |
| theregister.com | 1052 KiB | 88.32ms | 60.12ms | 1.5x |
| kubernetes.io/docs/concepts/services-networking/service/ | 563 KiB | 72.18ms | 43.43ms | 1.7x |
| docs.python.org/3/library/socket.html | 301 KiB | 71.88ms | 49.04ms | 1.5x |
Ranges from 1.1x to 1.7x, and the win grows with page size. A ten result web search sheds roughly 125ms of parsing. Because the async path builds its soups inline in `_unpack_fetch_results()`, that is 125ms the event loop spends parsing HTML instead of serving other users' streams. Pages under about 10 KiB are marginally slower under lxml due to fixed setup cost, which is worth nothing either way.
## Output verification
The risk in changing parser is silently different extracted text, so that was measured rather than assumed. Across all 37 real pages:
- **Zero characters of text were lost.** Every diff opcode against html.parser output was an insertion. Not one page dropped content under lxml.
- 659 characters were added, all on one page (docs.docker.com), where an inline Alpine.js `@click` handler containing a regex confuses libxml2's attribute handling and leaks a 73-character JS fragment into the text nine times. That is 659 characters of script noise in 27,206 characters of extracted text, with no content affected.
- Metadata (`title`, `description`, `language`) was identical on 35 of 37 pages. The two exceptions are 141-byte Wikipedia bot-block stubs with no `<html>` element, where lxml's fragment auto-wrapping adds `language: "No language found."`. Both parsers extract the same text from them.
Large documents were checked separately because libxml2 carries internal size caps. A 12 MiB single text node, 12 MiB spread across 400k nodes, a 3 MiB attribute value and 50k sibling elements with a trailing marker all produced byte-identical text under both parsers, with no truncation.
Malformed markup was checked too. lxml and html.parser diverge on unterminated comments, bare CDATA and duplicated `<html>` elements, all cases where both parsers are guessing and neither is correct. None of those shapes appeared in the 37 page corpus.
`backend/open_webui/env.py:184` also uses `html.parser`, on the local CHANGELOG at import time. That is trivial input on a startup path and is deliberately left alone.
The European Portuguese (pt-PT) catalogue mixes Brazilian and European forms
for the same concepts. In every case below the file ALREADY uses the European
term elsewhere, so these read as inconsistencies within the locale.
Corrects 74 string values. No keys are added, removed, renamed or reordered,
and no empty strings are filled (982 before, 982 after).
Genuinely Brazilian forms (Priberam tags these [Brasil] explicitly):
arquivo -> ficheiro (file sense only; Priberam: "Equivalente no portugues
de Portugal: ficheiro")
usuario -> utilizador (Priberam sense 3: "[Brasil] ... = UTILIZADOR")
acessar -> aceder (Priberam: entire entry tagged [Brasil])
equipe -> equipa (Priberam: "Grafia em Portugal: equipa")
midia -> multimedia (Priberam: every sense [Brasil])
"em um" -> "num"
"esta digitando" -> "esta a digitar": European Portuguese prefers
estar a + infinitive over estar + gerund (Microsoft pt-PT Style Guide p.34)
European UI convention, where the file is inconsistent with itself. These
words are NOT Brazilian; the change is for consistency, not correctness:
senha -> palavra-passe (0 vs 12 in this file) salvar -> guardar (3 vs 19)
compartilhar -> partilhar (6 vs 29) excluir -> eliminar (4 vs 19)
conexao -> ligacao (3 vs 27) gerenciamento -> gestao
desenvolvedor -> programador
Settings -> "Definicoes" (Settings label sense only, 15 vs 3)
the explicit pronoun "voce" -> European formal verb forms
Outright defects:
"Confirme sua a nova palavra-passe" -> "Confirme a sua nova palavra-passe"
"e sua especificacao" -> "e a sua especificacao" (missing article)
"Compartilhamento Publico" -> "Partilha Publica" (gender agreement)
"Archive" -> "Arquivar", not "Arquivo". All four call sites are button
labels/tooltips invoking archiveChatHandler(), so this is an action and
needs the verb, matching its siblings "Archive All" -> "Arquivar Tudo".
Deliberately NOT changed:
- 12 strings whose English key says "Config"/"configuration". "Configuracao"
is correct European Portuguese; only the Settings label becomes
"Definicoes".
- "You" -> "Voce" (the chat label), the string referencing that label, and
the default system prompt.
- "confiavel" is left alone: Priberam gives it no regional label and
professionally translated pt-PT (GNOME) ships it.
The migration to joserfc completed the job but left the old dependency pinned. `python-jose` now has zero imports anywhere in the backend: the only `jose` references left are `joserfc` in `utils/oauth.py`, and a repo-wide search for `from jose`, `import jose` or `python_jose` returns nothing outside the three pin files.
Removing it also removes `ecdsa` and `rsa` from the image, which were pulled in only by python-jose. `uv lock` confirms that: it drops exactly those three packages and nothing else, because google-auth 2.55 depends on cryptography and pyasn1-modules rather than rsa. That is worth having beyond the size saving, since `ecdsa` ships a documented Minerva-style timing side-channel in its P-256 signing path that upstream has declined to fix, so keeping it in the image means shipping a flagged crypto library that nothing calls.
Verified by blocking the `jose` module at import time and importing the backend anyway:
```
PASS import open_webui.utils.auth
PASS import open_webui.utils.oauth
PASS import open_webui.main
jose in sys.modules: False
PASS create_token/decode_token round trip
```
One user-visible consequence worth stating: Tools and Functions run in the same interpreter, so a third-party plugin that imports `jose` directly stops working after this. Nothing in Open WebUI itself does, and PyJWT remains a dependency, but a plugin relying on a library the application never declared for that purpose is the only thing this can break.
`uv.lock` was edited surgically rather than regenerated, to avoid the unrelated whole-file churn a newer uv version introduces. The result was diffed against real `uv lock` output and matches it exactly apart from that version's cosmetic fields.
The events:chat socket handler called the ownership-checked update for last_read_at, discarded the boolean it returns, and then cancelled the chat's pending timers regardless of the answer. cancel_timers_for_chat selected on the internal marker, the type, the parent chat id and the status, and never on the owner, so it matched rows belonging to any user. An authenticated user who knew another user's chat id could mark that chat read over their own socket session and silently cancel the owner's pending timers, and the owner got no notification: the scheduled action simply never fired.
The missing owner predicate also cut the other way in ordinary use. Because the query matched every timer sharing a parent chat id, one user reading a chat cancelled the timers of anyone else holding one on the same chat, so this was collateral damage as much as an attack.
cancel_timers_for_chat now requires a user_id and filters on it, which is the durable fix, and the socket handler returns early unless the ownership-checked update reports that the caller owns the chat. The parameter is required rather than defaulted so a later caller cannot reintroduce the unscoped query by omission. Both existing call sites already know the acting user. Timer rows are created with the same owner as the parent chat and the execution path already refuses to run one whose owner does not match, so scoping the cancellation the same way cannot strand a timer that would otherwise have fired.
One behaviour change worth noting: an administrator posting into another user's chat no longer cancels that user's chat.user_message timers, because the acting user is the administrator. The timer fires instead of being cancelled, which is the safe direction.
* refac: use MilvusClient instead of deprecated ORM-style PyMilvus APIs
PyMilvus 2.6 emits a PyMilvusDeprecationWarning for every ORM-style call (`connections.connect`, `utility.*`, `Collection` and its methods) and will remove those APIs in PyMilvus 3.1. Both Milvus backends still used them, so a running instance floods its logs with deprecation warnings during indexing and retrieval, and would break outright once PyMilvus 3.1 lands.
Both vector clients now go through `MilvusClient`:
- `milvus_multitenancy.py`: collection creation, index creation, has_collection, insert, search, query iteration, delete and reset.
- `milvus.py`: the remaining ORM calls in `query()` (`connections.connect`, `Collection(...).load()`, `Collection.query_iterator`), plus the now-unused `FieldSchema` import.
Behaviour is unchanged: same schema, same index parameters and the same two-step scalar-index fallback, same filter expressions, same result shapes. Verified against embedded Milvus (milvus-lite, pymilvus 2.6.14) with a functional harness over both clients: insert, get, query by string/int/bool metadata filters, vector search, tenant isolation, oversized-text truncation, delete by id and by filter, delete_collection and reset all return identical results before and after, while the deprecation warnings drop from 57 to 0 for the multi-tenancy client and from 16 to 0 for the standard one.
One Milvus Lite nuance worth recording: `MilvusClient` sends index build parameters (`M`, `efConstruction`, `nlist`) as flat keys rather than as a nested `params` blob. A Milvus server accepts both forms, Milvus Lite only reads the nested one, so those tuning values are ignored on Lite. `MilvusClient` offers no way to send the nested form, and `milvus.py` already built its index parameters this way, so both backends are now consistent.
Fixes#26978
* refac: correct the Milvus scalar-index comment
The comment claimed that embedded Milvus Lite requires an explicit scalar index type. It does not: Milvus Lite rejects `create_index` on a VARCHAR field outright ("create_index only supports vector fields"), for every index type and with or without a metric type, so neither the parameterless call nor the explicit INVERTED fallback can succeed there. Filtered queries on `resource_id` still work on Lite, just unindexed.
Only the accurate half is kept, which is the reason the parameterless call is deliberate rather than an omission.
Follow-up to #27497, reopened as a high contrast mode change, and the third and last of the contrast set after #27554 and #27555.
The grey scale in src/tailwind.css is achromatic oklch(L 0 0), so relative luminance is exactly L³ and text-gray-500 is 2.77:1 on white, against the 4.5:1 required by WCAG 1.4.3 and the 3:1 required of icons by 1.4.11. Around 290 sites use text-gray-500 dark:text-gray-400 for secondary labels, descriptions, counters and icons. Dark mode already passes at 6.46:1 and is left alone.
Rather than rewriting the class literal at every site, the remap is two CSS rules that only apply when the existing High Contrast Mode setting is on, so the default theme is untouched. Light mode resolves to gray-600 (5.75:1).
The split is not cosmetic. The `text-gray-500` utility is overridden inside `@layer utilities` with `:where()` so the rule sits below `hover:text-gray-*` and `dark:hover:text-gray-*` in specificity and hover feedback keeps working. The `.app-muted`, `.app-icon-muted` and `.tiptap table` classes in src/app.css are declared unlayered, which means no layered rule can reach them, so their override is unlayered too. They are `@apply text-gray-500 dark:text-gray-400` and fail identically, so leaving them out would have left the slash command menu and tiptap tables below 4.5:1 with the setting on.
Verified in a browser against Tailwind's emitted rules and layer order: with the setting on, the utility, .app-muted and .tiptap table all resolve to gray-600 in light while a focusable element carrying hover:text-gray-700 still resolves to gray-700 on interaction; dark mode and the setting-off case are unchanged in both themes.
One site is deliberately left out: EmbeddedChatHistoryDropdown.svelte uses text-gray-500/70, a separate class token that the selector does not match.
Follow-up to #27496, reopened as a high contrast mode change.
Placeholder text is the lowest contrast text in the product. The grey scale in src/tailwind.css is achromatic oklch(L 0 0), so relative luminance is exactly L³, and placeholder:text-gray-300 is 1.58:1 on white against the 4.5:1 required by WCAG 1.4.3. The large text exemption does not apply, the largest of these is text-lg. Placeholders are frequently the only format hint a field gives, for example admin/Settings/General.svelte uses e.g.) "http://localhost:3000".
Rather than deleting the ~200 per-component placeholder utilities and rewriting the base rule, the remap now happens in two CSS rules that only apply when the existing High Contrast Mode setting is on, so the default theme is untouched:
- placeholders resolve to gray-600 (5.75:1) in light mode
- placeholders resolve to gray-500 (6.46:1) in dark mode
The rules sit in `@layer utilities` and are anchored on `input`/`textarea`, which puts them above both the base rule in src/tailwind.css and every per-component `placeholder:text-*` utility, so no call site has to change. Placeholders stay distinguishable from real input values, which are text-gray-700 (8.46:1) in light and dark:text-gray-300 (11.39:1) in dark.
The chat composer placeholder is a tiptap ::before, so neither the base rule nor any utility reaches it. It is hardcoded #676767, which is 5.66:1 in light but only 3.17:1 on the dark canvas, so only the dark side is remapped, to gray-500. That rule stays outside the layer because the rule it overrides is unlayered too.
The root layout toggles a `high-contrast` class on documentElement from `$settings.highContrastMode`, alongside the existing theme classes, so every route is covered and the class is removed again when the setting is turned off.
Verified in a browser against Tailwind's emitted rules and layer order, on inputs both with and without per-component placeholder utilities: with the setting on, placeholders resolve to gray-600 in light and gray-500 in dark, the composer placeholder resolves to gray-500 in dark even with the prefers-color-scheme override treated as unconditional, and with the setting off nothing changes in either theme. Also checked against the oled-dark theme (gray-500 on #000 is 7.57:1) and the dark:bg-white/[0.03] input surface (6.01:1).
Note: the `high-contrast` class toggle is the same hunk as in the muted text contrast branch. Whichever lands first, the other rebases cleanly by dropping it.
The `create_automation` and `update_automation` builtin tools wrote straight to `Automations.insert` / `Automations.update_by_id`, skipping the limit checks that `/api/v1/automations/create` and `/api/v1/automations/{id}/update` run through `check_automation_limits`. A non-admin user could therefore ask the model to create automations indefinitely, ignoring `AUTOMATION_MAX_COUNT`, and could schedule them below `AUTOMATION_MIN_INTERVAL`, on both create and update.
Both tools now call the same `check_automation_limits` helper the routers use, so the limits and the admin bypass cannot drift between the chat path and the HTTP path. A rejection is returned to the model as a plain error message instead of raising. `update_automation` also gained the missing user lookup guard, since the helper needs the user's role.
The `automations.enable` toggle and the `features.automations` user permission were already enforced when the tool set is assembled, so they are unaffected.
Fixes#27121
Uvicorn's `--ws auto` selected its `websockets_impl` protocol on 0.41.0, which is built on `websockets.legacy`. That module raises `AssertionError` in `_drain_helper` during keepalive pings and kills the websocket connection. Each crash runs the Socket.IO `disconnect` handler and drops the session from `SESSION_POOL`, so every subsequent server-to-browser call fails. The most visible symptom is the Pyodide code execution tool, which reaches the browser through `sio.call('events', ...)` and returns `{"stderr": "Client session disconnected."}` on every run.
Uvicorn 0.50.0 changed `--ws auto` to select the sans-io implementation whenever websockets is installed, and deprecated the legacy one. Bumping the pin therefore fixes this on every launch path at once, without adding a `--ws` flag to the startup scripts. Doing nothing is not stable either: websockets is unpinned apart from uvicorn's own `>=13.0` floor, and `websockets.legacy` is removed outright in websockets 17, which turns the current AssertionError into an ImportError on a fresh install.
Bumping to 0.51.0 rather than the minimum 0.50.0 also picks up the sans-io keepalive pings added in 0.44.0, so raw websocket endpoints keep the idle-timeout behaviour they have today behind a reverse proxy. Uvicorn 0.51.0 drops colorama from its `standard` extra and raises the httptools floor to 0.8.0, which the lockfile already satisfies.
Verified on the bumped pin: the backend boots, `/health` returns 200, `--ws auto` resolves to `WebSocketsSansIOProtocol`, a Socket.IO client completes a websocket handshake against the running app, and a bidirectional `sio.call` round trip succeeds. The unit test suite reports an identical 2273 passed / 7 failed on 0.41.0 and 0.51.0, with the 7 failures unrelated to uvicorn.
Fixes#27550
`from datetime import datetime, time, timedelta` shadows the `time` module, so `RateLimitMixin._sync_wait_for_rate_limit` calls `datetime.time.sleep` and raises `AttributeError: type object 'datetime.time' has no attribute 'sleep'` whenever it actually has to wait.
Every synchronous loader path that paces requests hits this. `SafeFireCrawlLoader.lazy_load` calls the limiter directly, and Tavily, Microsoft Web IQ and Playwright reach it through `_safe_process_url_sync`. The exception is raised inside their per-URL `try`, so with `continue_on_failure=True` (the default) the URL is logged as a per-URL failure and dropped instead of being scraped. This is live by default: `WEB_LOADER_CONCURRENT_REQUESTS` is passed as `requests_per_second` and defaults to 10, so any URL whose predecessor finished within 100ms takes the sleep branch and is lost. Tavily and Microsoft Web IQ report it as "SSL verification failed", which points at the wrong cause.
`_wait_for_rate_limit` uses `asyncio.sleep` and is unaffected, but `SafeMicrosoftWebIQLoader.alazy_load` runs `lazy_load` in a threadpool, so its async entry point is affected too.
`datetime.time` is not used anywhere in the file, so importing the `time` module instead is enough.
The per-URL `continue` half of #26079 landed in 6f8221df5, which also added the `_sync_wait_for_rate_limit()` call to the Firecrawl loop. This makes that call work rather than throw.
Fixes#26079
Both routes read `chat_id` from the request body and passed it into `get_event_emitter` without checking the caller owns that chat. The emitter persists through `upsert_message_to_chat_by_id_and_message_id`, which resolves by primary key and takes no owner argument, so an invoked filter or action wrote into whichever chat the caller named. `/api/chat/completions` already performs this check; these two routes did not.
Adds `verify_chat_ownership`, called at the top of both handlers. It runs before the existing try block because the `except Exception` there catches HTTPException and would rewrite the 404 into a 400. Admins are exempt, matching the completions path, so deliberate cross-user operations keep working.
`local:` chat ids are allowed through: they are per-socket, the emitter suppresses database writes for them, and the socket emit targets the caller's own room. `channel:` chat ids are rejected instead. They reach the channel emitter, whose write only checks that the message belongs to the channel and never that the caller may write it, and the membership and write-access gate for channels exists solely on `/api/chat/completions`. No caller sends a `channel:` id to these two routes: the only frontend callers are in the regular chat UI, and the backend channel path dispatches through the completions handler.
Co-authored-by: manus-use <213290975+manus-use@users.noreply.github.com>
`admin/Users/Groups/Permissions.svelte` contains **64** `<Switch>` instances and not one of them passes `ariaLabel`, `ariaLabelledbyId` or `id`. bits-ui renders the switch as a `<button role="switch">` whose subtree is a text free thumb, so all 64 have **no accessible name**. The visible label is a sibling `<div>` with no association to the control.
This is the worst remaining case in the admin area: 64 toggles in one dialog, many with near identical adjacent labels (Import Models / Export Models / Import Prompts / Export Prompts / Import Tools / Export Tools). A screen reader user hears 64 consecutive "switch, on" and "switch, off" with no way to tell which permission is which.
Breaks WCAG 4.1.2 Name, Role, Value (Level A).
Fix: pass the row's own label to each switch. The `ariaLabel` expression is the **same `$i18n.t()` key** as the visible text two lines above it, so the accessible name equals the visible label in every locale, which also satisfies 2.5.3 Label in Name and keeps voice control working.
`ariaLabel` rather than `ariaLabelledbyId`, which is what `chat/Settings/Interface.svelte` uses for the same row shape. The difference is that `Interface.svelte` is a singleton, whereas this component is rendered from `EditGroupModal`, which is instantiated in three places including once per group in `GroupItem.svelte`. Only one can be visible today, but nothing enforces that, and 64 hardcoded ids would fail silently the day two coexist, since `aria-labelledby` resolves to the first matching id. `aria-label` has no such failure mode and needs half the edits.
All 64 mappings were checked individually rather than assumed. The nearest preceding label is the correct one in every case, including the three rows wrapped in a `<Tooltip>` (whose `content` attribute precedes the label in source order) and the ~60 `{#if}` / `{:else if}` explanatory strings (which always follow their switch). All 64 resulting labels are distinct.
The nested sub toggles are unambiguous on their own because upstream already labelled them fully ("Import Models" rather than "Import"), so no extra scoping is needed.
Two known follow ups, deliberately not bundled:
- The warning tooltips on Tools Access, Skills Access and Automations ("Warning: Enabling this will allow users to upload arbitrary code on the server.") are attached to a non focusable wrapper `<div>`, so keyboard and screen reader users never receive them. That needs a change in `common/Tooltip.svelte` or an `ariaDescribedbyId` on `Switch`, not a naming change.
- This file is a ~14 line block repeated 64 times where only the label and permission key vary, and it wants a shared `PermissionRow` component. Extracting it here would bundle a large structural refactor into an accessibility fix and make the diff unreviewable against the claim, so it is left alone.
The diff is +164/−65 rather than 64 changed lines, because 33 of the switches exceed the 100 column print width and Prettier reflows them to the multi line form. The file is Prettier clean and compiles with no new warnings.
Severity: Serious.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, each sidebar section header in `Sidebar/Section.svelte` is a real `<button>` carrying `aria-expanded` and `aria-controls`, but it has **no activation handler**. The toggle comes only from `on:pointerup` on the wrapper inside `common/Collapsible.svelte`.
Keyboard activation dispatches a synthetic `click`, never `pointerup`, and that wrapper's own `on:click` handler calls `stopPropagation()`. So pressing Enter or Space on the header does nothing at all, while `aria-expanded` tells assistive technology this is a working disclosure control.
This affects every section in the sidebar: Models, Notes, Channels, Folders and Chats. Section state is persisted to `localStorage`, so a user whose section was collapsed on a previous visit has no keyboard way to open it again, and the content stays unreachable.
Breaks WCAG 2.1.1 Keyboard (Level A), and 4.1.2 Name, Role, Value (Level A), because the exposed expanded state belongs to a control that cannot be operated.
Fix: handle activation on the header button itself, where focus actually lands, and stop the now duplicate pointer path so a mouse click does not toggle twice. The existing inline `onChange` body is extracted to `setOpen` so the `change` dispatch and the `localStorage` write stay in one place and fire exactly once per toggle in both input modes. The adjacent "+" (`onAdd`) button already stops both `pointerup` and `click`, so it still does not toggle the section.
`Collapsible`'s wrapper cannot simply become a `<button>` instead, because its slot receives buttons from this component and others, so the fix belongs here.
`common/Folder.svelte` and `Sidebar/RecursiveFolder.svelte` have the same latent defect and are not touched by this PR.
Severity: Critical. Sidebar navigation cannot be expanded without a mouse.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, `SensitiveInput` defaults to `export let id = 'password-input'`. The id is used both for the input itself and as the `for` target of the screen reader label rendered just above it.
There are 80 `<SensitiveInput>` usages in `src/` and only 4 pass an explicit id, so the remaining 76 all render `id="password-input"` together with `<label for="password-input">`. These collide on the same page in completely ordinary configurations: `admin/Settings/Audio.svelte` renders 4 at once with `STT_ENGINE === 'openai'` and 4 more with `TTS_ENGINE === 'openai'`, `admin/Settings/Documents.svelte` has 11, and `admin/Settings/WebSearch.svelte` has 33.
`for` resolves to the first matching element, so every label after the first points at the wrong input. In practice a screen reader user tabbing to the OpenAI TTS API key field hears the label belonging to the STT key field from a different section, and every one of those fields announces the same name. Browser password managers and any `getElementById` lookup collapse onto the first element the same way.
Breaks WCAG 1.3.1 Info and Relationships (Level A), because the programmatic label/field relationship is wrong, and 4.1.2 Name, Role, Value (Level A), because the fields do not expose their correct accessible name.
Fix: default the id to a per instance unique value. A Svelte prop default is evaluated per component instance, so each `SensitiveInput` gets its own stable id, and the 4 call sites that pass an explicit id are unaffected. `uuid` is already a direct dependency and `import { v4 as uuidv4 } from 'uuid'` is the existing pattern in the codebase, including `common/Collapsible.svelte`, which already generates a DOM id this way.
Note for self hosted setups: a `#password-input` selector in `static/custom.css` would stop matching. That selector already matched up to 8 elements at once on the Audio settings page, so it was never a reliable hook.
Severity: Serious. Every API key field in Admin Settings is mislabelled for assistive technology.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, the `title !== null` branch of `Collapsible` renders its header as a bare `<div>` whose only handler is `on:pointerup`, with the two Svelte a11y warnings suppressed above it.
`pointerup` is never dispatched by keyboard activation, and the `<div>` has no `role`, no `tabindex` and no `aria-expanded`. The header is therefore not focusable, not activatable and not announced as a control. This is the header for "Thinking..." / "Thought for N seconds", "Analyzing..." / "Analyzed", and every `<details>` block rendered from model output, via `Messages/Markdown/MarkdownTokens.svelte`, `Messages/StructuredOutputRenderer.svelte` and `chat/Controls/Controls.svelte`.
In practice a keyboard or screen reader user cannot expand any model reasoning trace, tool call detail or code interpreter block, and a screen reader reads the header as static text with no hint that anything is collapsed behind it.
Breaks WCAG 2.1.1 Keyboard (Level A), since the disclosure has no keyboard operation at all, and 4.1.2 Name, Role, Value (Level A), since it exposes neither a button role nor its expanded state.
Fix: render that header as a real `<button type="button">` with `aria-expanded` and the native `disabled` attribute, and toggle on `click`, which fires for both pointer and keyboard activation. This branch contains no `<slot />` and no interactive descendants, so a button is valid here. `block text-start` keeps the previous box and alignment behaviour, since a `<button>` otherwise defaults to `inline-block` and centred text. `disabled:cursor-default` replaces the old `{disabled ? '' : 'cursor-pointer'}` ternary, which became a no-op once this was a button, because `src/tailwind.css` applies `cursor-pointer` to every `button`. Verified in a browser that display, text alignment and rendered height match the previous `<div>`, and that a disabled header no longer shows a pointer cursor.
Switching from `pointerup` to `click` also means the header no longer toggles on right click, or when a drag starts outside it and ends inside.
The `{:else}` branch is deliberately left alone. Its `<slot />` receives buttons from `Sidebar/Section.svelte`, `common/Folder.svelte` and `Sidebar/RecursiveFolder.svelte`, so it cannot legally become a `<button>` and needs a different fix.
Severity: Critical. Model reasoning output is entirely unreachable without a mouse.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, `ConfirmDialog` registers `handleKeyDown` on `window` and treats **every** Enter keypress as Confirm, calling `event.preventDefault()` first so the focused control never gets its native activation.
The dialog also activates a focus trap with no `initialFocus`, so focus-trap falls back to the first tabbable node, which is the **Cancel** button. So the dialog opens with Cancel focused, and pressing Enter runs Confirm.
This is the confirm surface for Delete chat, Delete folder, Delete model, Delete knowledge base and ~40 other call sites. A keyboard user who tabs to Cancel and presses Enter deletes the thing they were trying to keep. Screen reader users are hit hardest, since they cannot see which button they are on and the control that means "back out safely" performs the irreversible action instead.
Two related paths have the same cause: Enter in the `input=true` textarea submits instead of inserting a newline, and a markdown link inside `message` (reachable via `eventConfirmationMessage` from tool `__event_call__` payloads, and via `web_search_confirmation_content`) becomes the first tabbable node, so Enter on that link confirms instead of following it.
Breaks WCAG 3.2.2 On Input (Level A): changing the focused control changes what the Enter key does, and activating a control performs a different action than the one it is labelled with. Also 2.1.1 Keyboard (Level A), since Cancel has no working keyboard activation.
Fix: let the focused control act on Enter itself, and only fall back to Confirm otherwise. Uses the same `target instanceof Element && target.closest(...)` guard already used in `Functions.svelte`, `Knowledge.svelte`, `Models.svelte`, `Prompts.svelte`, `Skills.svelte` and `Tools.svelte`. `select` is deliberately not in the list, because a native `select` does not act on Enter and excluding it would silently break confirm for the `inputType === 'select'` variant. Two stray `console.log` calls in the same function are removed.
Behaviour after this change: Enter on Cancel cancels, Enter on Confirm confirms, Enter in the textarea inserts a newline, Enter on a link follows it, and Enter anywhere else still confirms as before.
Severity: Critical. Silent, unrecoverable data loss triggered by the most ordinary keyboard interaction there is.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, `common/Checkbox.svelte` renders a `<button type="button">` containing only `aria-hidden="true"` SVGs. It has no `role`, no `aria-checked` and no accessible name, and the component has no `$$restProps` spread, so a caller cannot supply a name either.
Assistive technology announces every one of these as an unnamed "button". A screen reader user cannot tell that the control is a checkbox, cannot tell whether it is on or off, and cannot tell what it toggles. The visible label is always an unassociated sibling element, for example `Capabilities.svelte` puts it in a preceding `<div>` with no `id`, and `Groups/Users.svelte` puts it in a different table cell from the checkbox.
Breaks WCAG 4.1.2 Name, Role, Value (Level A) on all three counts at once.
Fix: expose `role="checkbox"` and `aria-checked` on the control, add an `ariaLabel` prop, and pass the label text that is already in scope at each call site. `aria-checked` mirrors the component's existing icon logic exactly, so the indeterminate dash reports `mixed` rather than `false`. The `ariaLabel={ariaLabel || undefined}` shape matches the sibling `common/Switch.svelte`. Every label expression is the same one that renders the visible text next to the checkbox, so the accessible name always matches what is on screen.
Three call sites are deliberately left out of this PR, because they nest `Checkbox` inside another `<button>`, which is invalid HTML and independently broken:
- `workspace/Knowledge/KnowledgeBase.svelte` — the Checkbox's `on:change` sets `includeContent = true` and then the same click bubbles to the outer button, which flips it back with `includeContent = !includeContent`. Clicking the checkbox square is a no-op today, only the text label works. Giving it a confident name would advertise a control that does nothing.
- `workspace/common/MemberSelector.svelte` (two instances) — the inner Checkbox has no `on:change` at all and only works because its click bubbles to the row button. Naming it would create two focusable controls per row with the same name.
Both need the nesting resolved first, so that the row button carries the checkbox semantics. That is a behavioural fix and belongs in its own PR.
Severity: Serious. Affects model capabilities, default features, builtin tools, tool/filter/skill/action selectors and group membership.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
On latest `dev`, `RichTextInput` passes only `attributes: { id }` to tiptap, so the rendered contenteditable has an implicit `textbox` role and **no accessible name at all**.
The only label is the tiptap placeholder, which renders as CSS generated content in `src/app.css` via `content: attr(data-placeholder)`. Generated content never becomes an element's accessible name, so assistive technology announces the field as "edit text, blank".
This is the chat composer, the channel and thread composers, and the note editor, so it is the most used control in the product.
Breaks WCAG 4.1.2 Name, Role, Value (Level A), and 3.3.2 Labels or Instructions (Level A), since the only instruction is invisible to assistive technology.
Fix: expose the placeholder as `aria-label` on the editor element.
`attributes` is passed as a **function** rather than an object literal. The object form is evaluated once when the `Editor` is constructed and never rebuilt, but `placeholder` is deliberately runtime mutable: `channel/MessageInput.svelte` and `channel/Thread.svelte` swap it between "You do not have permission to send messages in this thread." and "Reply to thread..." once `channel` resolves, and it also changes when the interface language changes. With the object form the field would have been permanently named with whatever string happened to be set at mount, which for a channel the user *can* write to is the no-permission message. That would be worse than no name at all. ProseMirror supports the function form and re-evaluates it on every state update, and the component's existing `setPlaceholder` already dispatches an empty transaction, so the label now tracks the visible placeholder. It binds to `_placeholder`, the same value that feeds the visible text, so the two cannot diverge.
`aria-multiline` is deliberately not set. It is only valid on an explicit `textbox`/`searchbox` role, and adding `role="textbox"` would flatten the editor's inner structure so headings, lists and links inside rich text stop being exposed.
Severity: Critical. The application's primary input announces as an unnamed edit field.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
`common/Switch.svelte` already accepts `id`, `ariaLabel` and `ariaLabelledbyId`, but **not one of the 148 `<Switch>` instances under `src/lib/components/admin/` passes any of them**.
`admin/Settings/AdminSettingRow.svelte` renders the row label as a plain `<div>` and the control in a **sibling** slot, so there is nothing tying them together. bits-ui renders the switch as a `<button role="switch">` whose subtree is a text free thumb, so it has no accessible name from any source.
A screen reader user working through Admin Settings hears a long run of "switch, on" and "switch, off" with no indication of what any of them controls.
Breaks WCAG 4.1.2 Name, Role, Value (Level A) and 1.3.1 Info and Relationships (Level A).
Fix: `AdminSettingRow` mints a per instance id, puts it on the label element, and hands it to the default slot, so each row's switch can point at the label that is already rendered next to it. This is the pattern `chat/Settings/Interface.svelte` already uses by hand in 45 places, hoisted into the shared row component so call sites stop hand authoring ids.
`aria-labelledby` rather than a wrapping `<label>`: per HTML-AAM a `<button>` takes its name from `aria-labelledby`, then `aria-label`, then its own subtree, never from an associated `<label>`. `chat/Settings/Subagents.svelte` already wraps two switches in a `<label>` and they are still unnamed, which is the same trap. Using the existing label element also guarantees the accessible name is byte identical to the visible text, which keeps voice control working.
The `description` paragraph deliberately sits outside the referenced element, so verbose help text is not pulled into the name.
Scope: this covers the **72** switches that live inside an `AdminSettingRow`, which is every switch that flows through the shared row component. There are no rows containing more than one switch, so nothing is silently skipped.
The remaining 76 admin switches are not in this component and are not touched. 64 of them are in `admin/Users/Groups/Permissions.svelte`, which hand rolls its own row markup, and the other 12 are per entity toggles in lists and dropdowns where the label is a dynamic row name. `Permissions.svelte` is the worst remaining case, 64 toggles with near identical adjacent labels, and it needs either its own labelling pass or a conversion to `AdminSettingRow` that changes its visual styling. Either way that is not an accessibility only diff and belongs in its own PR.
All 12 touched files compile with the Svelte compiler with no new warnings and are Prettier clean.
Severity: Serious. Admin Settings is unusable with a screen reader.
### 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
-->
- [x] 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.
`SafePlaywrightURLLoader` opened a new Playwright page for every URL and never closed it, and it only closed the browser after the URL loop finished normally. Pages therefore piled up for the whole batch, and any early exit (a raised error with `continue_on_failure=False`, or the caller abandoning/cancelling the generator mid-search) skipped `browser.close()` entirely.
With `PLAYWRIGHT_WS_URL` pointing at a remote Playwright server this leaks sessions on that server: navigation and route timeouts on slow or bot-protected pages leave pages and browser connections open until the server is restarted, which degrades every later web search.
Both `lazy_load()` and `alazy_load()` now scope the page to the per-URL loop body and the browser to the whole loop using their context managers, so each page is closed as soon as its URL is done and the browser is closed on success, on failure, and on cancellation. Closing a page also disposes the context implicitly created by `new_page()`. Exception handling is unchanged: a close error raised while `continue_on_failure` is set is still caught, logged, and the loop continues.
Fixes#25880
When `RAG_DOCUMENT_LOADER_ENGINE` is set to `paddleocr_vl`, the dispatch branch in `Loader._get_loader` checked only the engine name and a non-empty token, so every uploaded file was handed to the PaddleOCR-VL loader regardless of its type. Text based uploads such as `.md`, `.txt` and `.csv` were base64 encoded and posted to the `/layout-parsing` endpoint tagged as PDFs, and the API rejected them with `422 Unprocessable Entity` ("PDFium: Data format error"), so those files never indexed at all.
The loader already knows which extensions it can handle: it tags images with `fileType: 1` and treats everything else as a PDF. That list is now a module level constant, and the dispatch branch gates on `['pdf'] + images`, the same way `mistral_ocr`, `datalab_marker`, `document_intelligence` and `mineru` already limit themselves. Deriving the gate from the loader's own list keeps the two in sync, so a file can never be admitted by the gate and then mislabelled as a PDF on the wire. Everything outside that set falls through to the default loader chain, so `.md` and `.txt` load as text, `.csv` through `CSVLoader`, `.docx` through `Docx2txtLoader`, and so on.
The branch also never checked `PADDLEOCR_VL_BASE_URL`. With the URL cleared, `PaddleOCRVLLoader` raised `ValueError` from its constructor and the upload failed outright instead of falling back. Both settings are now required for the branch to be taken, matching how the other engines guard their own configuration.
Fixes#24988Fixes#26759
The Socket.IO handshake and the terminal WebSocket route each reimplement JWT authentication instead of going through the HTTP dependency chain. Both verified that the token decoded, that it had not been revoked, and that the user row existed, but neither applied the role check that `get_verified_user` enforces on every HTTP route, so any role outside `user` and `admin` was accepted.
That splits authorization across two planes. Deactivating an account by setting its role to `pending` takes effect immediately over HTTP, which returns 401, while the same JWT still opens a WebSocket. Changing a role disconnects the account's live sockets but does not revoke its token, so the client simply reconnects and gets a fresh session. Until the token expires, four weeks by default, a deactivated account keeps its channel rooms and can still read and write any note it holds an access grant on through the collaborative document handlers.
Resolve the user once, in `get_verified_user_by_token`, and route both WebSocket entry points through it. The role set moves into `VERIFIED_USER_ROLES` so the HTTP and WebSocket gates cannot drift apart, which is the underlying cause rather than either call site on its own. This also replaces five copies of the decode, revocation check and user lookup sequence.
`user-join` now resolves the user instead of reusing the identity cached in `SESSION_POOL`, which costs one extra query per handshake. Gating on the cached role would make the authorization decision depend on every future role-mutation path remembering to tear down the session pool, and that is precisely the invariant that failed here.
* i18n: complete de-DE translations
Fill in all 544 untranslated (empty) strings in the German locale and add
the two keys that were missing entirely ("Response Auto-Scroll" and
"Follow assistant responses as they are generated.").
Wording follows the conventions already used in the file: formal "Sie"
address for user-facing sentences, infinitive phrasing for labels and
buttons, third-person descriptive phrasing for setting descriptions, and
the established terminology (Kontextverdichtung, Erinnerungen,
Wissensspeicher, Werkzeuge, Chunk, Embedding, Skills, Pipelines).
Ambiguous strings were resolved against their usage in the Svelte
components, e.g. "at"/"Through" (schedule and heatmap tooltips),
"Runs"/"runs" (automation runs vs. tool invocations), "Current"
(active chat) and "Selected" (model filter).
* i18n: fix de-DE wording and two pre-existing plural bugs
Review pass over the German locale:
- "Claim" and "DN" are masculine: "Claim, der ..." instead of "Claim,
das ...", "Passwort für den Bind-DN", "Base DN, der ...".
- "hinzufügen" governs the dative, matching the existing string
"... fügen Sie sie zuerst dem Arbeitsbereich "Wissen" hinzu."
- "Beschränkt oder schließt Domains ... aus" was a zeugma; the separable
prefix only belongs to "schließt".
- Sub-agent settings render as label + input + unit suffix on one line,
so the label and suffix no longer repeat each other.
- The built-in tool descriptions are infinitive, so the notification one
is too.
- Align wording with terms already used in the file: Assistentennachrichten,
Benutzernachrichten, Vervollständigungen, Tool-Server, Wissensspeicher,
lexikalisch. Normalize the few German typographic quotes to the ASCII
quotes used everywhere else.
- The username setting claimed the chat shows "Sie", but "You" is
translated as "Du".
Also fixes bugs that predate these translations: "Starting in {{count}}
minutes" had the raw "minutes_one"/"minutes_other" suffix in its value,
and the singular and plural of "Ran {{COUNT}} analysis/analyses" were
swapped.
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
The security policy described Open WebUI as "a small volunteer team" and "a volunteer- and community-driven project", and explained response times as a shortage of capacity. Read by enterprise evaluators, security researchers and third parties trying to impose disclosure timelines, that wording makes the project look informal, under-resourced and externally steerable, which is the opposite of the position the policy is meant to hold.
Open WebUI is led and maintained by a small core team with clear ownership of the security process. This updates the wording to say that, and reframes response times as prioritisation across the project rather than a capacity shortfall. No rule, scope, commitment or timeline changes: the reporting channel, the disclosure schedule, the credit rules and the expected timeframe all stay exactly as they were.
Also removes the implicit first-come-first-served promise in the follow-up paragraph, which contradicted the severity-based prioritisation stated two paragraphs later, and bumps the last-updated date.
When navigating from a chat to a non-chat route (e.g. the admin panel),
the previously-viewed chat stayed selected in the sidebar and
deleting/archiving it wrongly redirected back to the new-chat page.
Cloning a chat also left the source chat highlighted alongside the new
clone, so two chats appeared selected at once.
Two independent sources kept the stale selection:
- The chatId store was never cleared when the Chat component unmounted,
so $chatId still pointed at the last-viewed chat (this drove the
delete/archive redirect). Clear chatId/chatTitle in Chat's onDestroy.
- The sidebar's optimistic selectedChatId highlight, set on click, was
only cleared on window blur (hence it appeared to fix itself after a
tab switch) and never followed programmatic navigation. Bind it to the
chatId store so it tracks the active chat for leave, delete and clone.
A workspace model shared publicly could be used by any user even when its
base model was private. Unregistered base models (no row in the model
table) are admin-only for direct use — get_filtered_models hides them from
non-admins and check_model_access rejects them — but has_base_model_access
treated a missing row as "no ACL" and allowed the chained request through.
has_base_model_access now takes the caller's role and only allows an
unregistered base model hop for admins, so a shared preset can no longer
reach a base model the caller could not use directly. Registered base
models keep their existing grant-based enforcement.
Claude-Session: https://claude.ai/code/session_018toPfJW1hMXAhokGaL43Ep
Co-authored-by: Claude <noreply@anthropic.com>
Since the config refactor, get_web_loader dispatched on the WEB_LOADER_ENGINE module constant, which is read from the environment once at import time. The engine selected in the Admin UI is stored under web.loader.engine in the config table but was never consulted, so UI-configured loader engines (external, playwright, firecrawl, tavily, microsoft_web_iq) were silently ignored and the built-in SafeWebBaseLoader always fetched pages directly. The same applied to the per-engine settings such as the external web loader URL and API key. This breaks egress-restricted deployments that rely on an external web loader: pages are fetched directly from the container and fail with errors like "Network is unreachable" even though an external loader is configured.
Pass the DB-backed loader settings into get_web_loader from both call sites, web search in process_web_search and web fetch via get_loader, and resolve every engine setting from them, keeping the module-level env constants as the fallback for keys that were never saved. Also initialise WebLoaderClass so an unknown engine raises the intended ValueError instead of an UnboundLocalError.
Fixes#26747
Docker compose list-form environment syntax passes quotes through verbatim, so WEB_FETCH_FILTER_LIST="" reaches the backend as two literal quote characters rather than an empty string. Config parsing turned that into the filter entry '""', which has no "!" prefix and therefore landed in the allow list. A non-empty allow list requires every host to match one of its entries, and a quotes-only pattern can never match a hostname, so every fetch_url and web loader request was rejected with "URL blocked by filter list" and surfaced to the user as "The URL you provided is invalid".
get_allow_block_lists now strips surrounding quote characters from each entry and drops entries that are empty after normalisation. Quoted but otherwise valid entries such as "example.com" or !"example.com" now behave as their unquoted forms, and garbage entries no longer convert the default blocklist into a match-nothing allowlist that blocks everything.
Fixes#26908
get_accessible_folder_files is the server-side filter that reduces a folder's attached-knowledge list (and, once #26723 lands, a direct model's) to the entries the caller may read, before that list is handed to the builtin knowledge tools as `__model_knowledge__`. It validated `file` and `collection` entries but passed `note` entries through unchecked (they fell into the `else` keep-as-is branch), even though notes are a first-class attached-knowledge type that flows through this list.
No current caller is exploitable, because every note consumer (`query_knowledge_files`, `view_note`, and the legacy retrieval path) independently re-checks note access before returning content. But relying on each consumer to remember that check is exactly the fragility this helper exists to remove, and the same `_has_read_access_to_file` membership short-circuit that makes an unvalidated `file` entry dangerous would turn any future note path that trusts list membership into an IDOR. Validate notes here so the filter enforces its own contract instead of leaning on downstream re-checks.
A note entry is now kept only when the caller owns it or holds a read grant. Notes are private by default and carry no self-grant, so ownership is checked explicitly alongside the grant lookup. Admins still bypass all checks and genuinely unknown types are still kept as-is.
Related: #26723
Since v0.10.0 a global fetch interceptor redirects to /auth and clears
localStorage.token whenever an authenticated backend request returns 401.
The OAuth callback cookie ("token", set with httponly=False so the
frontend can read it) is left behind. The auth page's oauthCallbackHandler
then immediately signs the user back in from that cookie and navigates to
"/", where the next 401 triggers the redirect again. The result is an
endless /auth and / ping-pong that renders as uncontrollable screen
flashing, and as a PWA stuck on a flashing splash screen when SvelteKit's
update check turns each navigation into a full page reload. Affected
users could only recover by clearing cookies, which matches the reports.
Clear the token cookie together with localStorage when redirecting, so
/auth stays on the login form and the user can sign in again normally.
Fixes#26731
seed_defaults inserted a row for every key in DEFAULT_CONFIG regardless
of whether the DB is authoritative for it. With ENABLE_OAUTH_PERSISTENT_CONFIG
off, the oauth.* keys were seeded from the then-current (often empty) env
values. Enabling the flag later made those stale rows override live env vars
(e.g. ENABLE_OAUTH_SIGNUP=true stopped taking effect) and further env changes
were never picked up.
Skip keys where persistent_enabled_for() is false, matching the masking the
read paths (get/get_many/get_namespace/get_all) already apply.
Claude-Session: https://claude.ai/code/session_01Vr2RCYUTXCtgtV4WMUCK86
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
The VoiceRecording component stays mounted (hidden) between recordings, and the web speech engine accumulated every session's transcript into the never-reset transcription variable. Each new recording therefore confirmed all previous utterances again, so the inserted text repeated once per session and previously deleted text reappeared in the input.
Additionally, cancelling a recording (X button, Escape or a recognition error) called stopRecording(), which stops the SpeechRecognition instance and fires its onend handler, which unconditionally confirms the transcription. Cancelled recordings therefore still inserted the accumulated transcript.
Reset the transcription at the start of each web speech session and detach the onend handler on cancel so cancelled recordings no longer confirm.
Fixes#26784
get_chat_by_id sent admins down a branch that returned the chat only when
ENABLE_ADMIN_CHAT_ACCESS was on, or the chat was internal, and never fell
through to the access-grant and shared-folder checks. With the setting off,
an admin was therefore denied a chat that had been deliberately shared with
them, either directly or through a shared folder, while any non-admin holding
the same grant could open it. The admin role removed access the user had been
given rather than only closing the admin-only path.
Try the admin path first, then let everyone fall through to the grant and
folder checks. ENABLE_ADMIN_CHAT_ACCESS=false still closes the admin-only
route to other users' chats, and internal chats stay reachable.
* fix: let only the attendee set their own calendar RSVP status
set_attendees took each attendee's status from the caller-supplied value, so an
event organiser could set another user's RSVP (for example to 'accepted') on
create or update. RSVP is meant to be self-service: the /events/{id}/rsvp
endpoint already scopes status changes to the calling user.
Derive attendee status server-side instead of from the request. An existing
attendee keeps the status they set via RSVP and a newly added attendee starts
'pending'; any caller-supplied status is ignored. Event edits no longer reset
attendees' existing responses.
Co-authored-by: legobattman <302282032+legobattman@users.noreply.github.com>
* fix: hide declined calendar invites from the attendee view
`get_events_by_range` surfaced every event where the user is an attendee regardless of their RSVP status, so declining an invite left it in the calendar with no way to remove it. Exclude `declined` attendee rows from the attendee branch, so a decline now removes the event from the user's own view while pending, accepted and tentative invitations still surface.
Co-Authored-By: legobattman <302282032+legobattman@users.noreply.github.com>
---------
Co-authored-by: legobattman <302282032+legobattman@users.noreply.github.com>
The Function and Tool database columns declare user_id as a nullable
String column, but their Pydantic read-models required a non-null
string. A record with user_id NULL therefore raised a
pydantic ValidationError inside get_functions()/get_tools(), which run
during install_tool_and_function_dependencies() at app startup —
crashing the whole application and blocking all chat completions.
Make user_id Optional in the read/response models so such records
validate gracefully (user is already rendered as None downstream when
the id has no matching user) instead of taking down startup.
Claude-Session: https://claude.ai/code/session_01Y4RRUNq7ZUFkRWbWPkDw3m
Co-authored-by: Claude <noreply@anthropic.com>
On narrow viewports the admin Models list rows previously collapsed into
unusable vertical stacks; the Models page redesign on dev has since
absorbed the truncation fixes this branch carried (min-w-0 chain, real
truncate on the name, shrink-0 action group, inline access label).
The one remaining gap: the per-row edit pencil duplicates the row tap
(both open the model editor) while costing scarce horizontal space on
mobile. Hide it below the sm breakpoint; it remains on sm+ screens.
AccessGrantsPermissions only declared allow_users, so allow_groups was
missing from the model backing the default user permissions endpoints.
Pydantic ignores undeclared fields, so POST /users/default/permissions
dropped allow_groups before model_dump(), it never reached the persisted
user.permissions config, and fill_missing_permissions restored it to the
default on the next read. Turning "Allow Sharing With Groups" off in
Admin Settings silently reverted to on, while the same toggle worked when
set per group, since group permissions are stored as a plain dict.
GET /users/default/permissions and /users/default/permissions/defaults
dropped it from their responses for the same reason.
Declare allow_groups on the model so it round-trips, matching the
access_grants block in DEFAULT_USER_PERMISSIONS. It defaults to True, so
existing payloads that omit it are unaffected.
This PR fixes two small errors in the German translation file:
- "kaann" -> "kann"
- "AAlle" -> "Alle"
No functional changes, only UI strings.
Co-authored-by: Tim Baek <tim@openwebui.com>
The channel message update and delete handlers enforced authorship only on group and dm channels. On standard channels the else branch accepted any caller holding write access on the channel, so a member who could post could also edit or delete messages authored by other members. Because the update form binds content, data and meta, and the model layer never touches message.user_id, an edited message kept the original author's attribution, so another member's message could be rewritten under their name.
Write access on a channel is the capability to post, not a moderation capability, and the frontend gates the edit and delete controls on authorship (message.user_id === user.id, or admin) for every channel type. The group and dm branch already encodes this with an explicit authorship check. Apply the same rule to the standard branch: the caller must hold write access on the channel and be the message author, unless they are an admin. Pinning is unchanged, since it is exposed to every member by design.
Since v0.10.0 chat messages are virtualized with content-visibility: auto to skip rendering of off-screen messages. Safari's implementation of content-visibility has known paint bugs (WebKit bugs 277573, 281570 and 283846) that can leave the contents of a message unpainted even when it is on screen. On iOS this makes assistant responses render as empty, both in Safari and as a PWA, while the same chats render fine in Chromium and Firefox. This matches the regression window reported in #26712, which appeared when upgrading from 0.9.6 to 0.10.2.
Detect Safari (including all iOS browsers, which use WebKit) with the same user agent check already used in MessageInput and ShareChatModal, and skip the virtualization class there. Safari falls back to rendering all messages like before v0.10.0, while other engines keep the optimization. Verified with a spoofed Safari user agent that messages render without the virtualization class and with content-visibility resolving to visible, while Chromium keeps content-visibility: auto.
Fixes#26712
The sandboxed Pyodide host (used when ENABLE_PYODIDE_FILE_PERSISTENCE is
disabled, the default) embeds its script in a String.raw template. The
matplotlib show() override was written with '\\t' escapes as if in a normal
string context, but String.raw preserves them verbatim, so the iframe's
script parser turns them into literal backslash-t characters in the
generated Python source. Pyodide then fails to compile any code that
triggers the matplotlib patch with "SyntaxError: unexpected character
after line continuation character", which is why matplotlib only worked
with the file persistence worker path enabled.
Use single '\t' escapes instead: String.raw keeps them as-is in the
script text and the sandbox's JS parser produces real tab indentation,
matching the working implementation in pyodide.worker.ts.
Fixes#26660
Assistant responses are now stored as structured output items on message.output, with message.content left empty. The action payload built in chatActionHandler still sent message.content only, so action functions received assistant messages with an empty content property. Derive the content from the structured output via getOutputText, falling back to message.content, matching how the rest of Chat.svelte resolves assistant text.
Fixes#26672
_ssrf_safe_new_conn unpacks each entry of self.socket_options straight into socket.setsockopt(), which accepts exactly 3 positional arguments. urllib3-future, a drop-in fork that shadows the urllib3 package whenever it is installed (for example as a dependency of niquests pulled in through a tool or function's requirements), declares its default socket options with a per-protocol 4th element: [(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1, "tcp")]. Its own _set_socket_options() strips that element before calling setsockopt(), but our override does not, so with urllib3-future present every synchronous web fetch (fetch_url, web search loading) fails on connect with "TypeError: setsockopt() takes exactly 3 arguments (4 given)" and returns empty content.
Mirror urllib3-future's handling in the override: for 4-element options whose last element is a protocol string, apply "tcp" options truncated to the first 3 elements and skip "udp" options (all sockets created here are SOCK_STREAM). Plain 3-element options, and any other shapes stock urllib3 would accept, are passed through unchanged, so behavior with stock urllib3 (which only ever uses 3-element tuples) is identical.
Verified locally: with urllib3-future installed the loader previously raised the TypeError on every URL and now fetches successfully; with stock urllib3 2.3.0 and 2.7.0 fetches behave the same before and after.
Note: #26015 reported this same crash but attributed it to stock urllib3 2.x, which only uses 3-element tuples; the 4-element form comes from urllib3-future shadowing urllib3.
Fixes#26791
is_regex_pattern only recognized the BRE-escaped form \| and not a bare |,
so a pattern like "Jornak|Silverlake|Orissa" was treated as one literal
string (including the pipe characters) and silently returned no matches.
This contradicted the tool docstring, which explicitly advertises
"error|warn" as an auto-detected regex example, and misled models into
concluding the searched terms were absent from the file.
Checking for a bare | also covers the escaped form, since \| contains |,
and normalize_regex already converts escaped pipes before compilation.
Literal patterns without regex metacharacters are unaffected.
Fixes#26781
SafeWebBaseLoader._unpack_fetch_results assigned the resolved parser to the parser parameter itself, so the None check only ran for the first URL. In a mixed batch every later document was parsed with whatever the first URL happened to select: an .xml feed first meant all following HTML pages went through the xml parser (broken text extraction), and an HTML page first meant .xml URLs were parsed as HTML. Web search regularly fetches mixed batches, so this silently degraded extraction quality depending on result order.
The parser is now resolved per URL; an explicitly passed parser still applies to the whole batch as before. Verified with mixed xml/html batches in both orders and with an explicit parser override.
When WEBSOCKET_MANAGER=redis, app.state.MODELS and the socket session/
usage pools are Redis-backed dicts, so every membership test and
getitem is a network round trip:
- generate_chat_completion checked `model_id not in models` (HEXISTS)
and then read `models[model_id]` (HGET) on every chat completion.
A single .get() now serves both, with the same not-found error.
- The direct-connection branch spread the pool with `{**MODELS, ...}`,
which iterates keys() then fetches each value — HKEYS plus one HGET
per model. dict(MODELS.items()) issues a single HGETALL instead.
- get_user_ids_from_room called SESSION_POOL.get(sid) twice per
session (once to filter, once for the value); the usage handler
checked membership then fetched the same key. Both now do one
lookup.
In non-Redis mode these are plain dicts and behavior is identical.
Claude-Session: https://claude.ai/code/session_01MHg5zs1VBjvRWQ54qHpfYD
Co-authored-by: Claude <noreply@anthropic.com>
The streaming handler's content accumulator is a closure cell (declared nonlocal in stream_body_handler), and CPython's in-place string append optimization only applies to plain local variables (STORE_FAST), never to cell variables (STORE_DEREF). The content += value form introduced in #27231 therefore still allocates and copies the full accumulated string on every delta, exactly like the f-string it replaced; whether that copy is cheap or expensive is up to the allocator, and measurements swing accordingly (6 to 83 ms of pure copying for a 400 KB response on Python 3.12, against 0.3 ms for this patch).
Accumulate the deltas in a list instead and join once at the single read site (publish_chat_finished_event at stream end). List append is amortized O(1) with no dependence on reference counts, bytecode specialization or allocator behaviour, so accumulation is O(n) by construction. The non-str fallback keeps the previous f-string coercion semantics.
Verified end to end against a mock SSE upstream: a streamed chat with a think-tag block plus 40 content deltas produces output items, message text, reasoning text and usage identical to current dev, with no errors in the server log.
CompressMiddleware was registered before AuditLoggingMiddleware. Starlette prepends on add_middleware, so the audit layer ended up outside compression and, at the REQUEST_RESPONSE level, recorded the zstd/brotli/gzip bytes of every response, decoded with errors='replace'. Any client that sent Accept-Encoding (i.e. every browser) therefore produced audit entries whose response_object was unreadable mojibake.
Registering the audit middleware before the compression middleware places it inside compression, so it observes the response body exactly as the route produced it while the client still receives the compressed stream.
Verified with a stacked ASGI harness: in the old order the captured body is not parseable; in the new order the captured body round-trips as the original JSON and the client response stays compressed.
The header value was truncated with int(), so every request faster than one second reported X-Process-Time: 0 and the header carried no information for exactly the requests it is meant to describe. Emit fractional seconds with microsecond precision instead, matching the pre-ASGI-refactor behavior where the raw float was sent.
With audit logging enabled, every audited request authenticated twice. The route dependency resolved the user once, and then _log_audit_entry called get_current_user again in the request's finally block: a second JWT decode, two more Redis revocation lookups, a second user row fetch with pydantic validation and, crucially, a second fire-and-forget last-active write transaction per request.
get_current_user now stashes the resolved user on the scope-backed request state (the same mechanism the auth middleware already uses for request.state.token), and the audit middleware reuses it, falling back to the old resolution only when no user was stashed (e.g. routes without an auth dependency). While in the file, the audit path patterns are compiled once in the constructor instead of per request, and the always-log endpoint set is a class attribute instead of a per-call literal; both are fixed for the process lifetime.
Benchmark:
| metric | before | after |
| --- | --- | --- |
| audit auth resolution, CPU floor (JWT decode + user validate only) | 16.7 us | 0.24 us |
| extra work per audited request | 2 Redis GETs + 1 user SELECT + 1 last-active write | none |
The before column understates the saving: it excludes the Redis and DB round trips listed in the second row, which dominate in real deployments.
Functionally verified with a stacked ASGI harness: when the route resolves a user the audit entry carries that user and the auth pipeline is not invoked again; without a stashed user the fallback path still resolves and logs correctly; the skip matrix (exclusions, whitelist mode, always-log auth endpoints, unauthenticated and non-audited methods) is unchanged.
Both session factories run with expire_on_commit=False, so ORM objects keep their attribute values after commit. Every session.refresh issued right after a commit therefore re-SELECTed a row whose values the session already held, including full chat JSON blobs and user settings, purely to overwrite identical data. Fifty such calls existed across the model layer, covering nearly every write path in the app (chat inserts, title updates, pin/archive toggles, user role and settings updates, tool, prompt, function, model, file, tag, feedback, memory, automation and grant writes).
All fifty are removed. The only refreshes with an actual job were the two update-then-reload paths in tools and skills, where a Core UPDATE statement bypasses the identity map; those now use session.get(..., populate_existing=True), which guarantees a fresh row in one SELECT whether or not the row was already present in the session (the previous code issued get plus refresh, two SELECTs, on the default configuration).
Benchmark (real SQLite DB, per write):
| write path | before | after |
| --- | --- | --- |
| chat title update, ~600 KB chat blob | 2.08 ms | 1.24 ms |
| user role update, small row | 1.21 ms | 0.68 ms |
On Postgres each removed refresh is additionally a network round trip. The chat-blob case also skips re-parsing the entire JSON document per write.
Functionally verified against a fresh database: user insert, role and settings updates, chat insert (including the server-default meta column, which is always provided client-side), title update and pin toggle, tool insert and the Core-update reload path, tag insert and the prompt insert flow that pins version_id after history creation all return correct values and persist correctly.
get_all_models runs on every models refresh and, without the base-models cache (off by default), on every /api/models request. Several of its costs multiplied by the model count for no reason:
- The active action and filter id sets were derived from get_functions_by_type, which loads full function rows including plugin source and validates them, only for the ids and is_global flags. A generalized column-only query now returns (id, is_global) tuples; the existing filter-specific helper delegates to it.
- Action priorities were computed inside the per-model sort key, constructing a pydantic Valves object per action per model; with global actions in every model's list that was models x actions constructions per refresh. Priorities are now memoized per action.
- Global action and filter item dicts were rebuilt per model from the same modules. The item lists are now built once per function and shallow-copied per model, keeping per-model dicts independent exactly as before (nested values were already shared).
- Deactivated base-model overrides were dropped with models.remove, a linear scan and shift per removal; removals are now collected and filtered out in one identity-based pass, preserving list.remove's exact object semantics.
- RedisDict.set fingerprinted the payload by serializing the already-serialized mapping a second time plus a sha256; a direct dict comparison against the last written mapping has the same skip semantics without re-serializing anything.
- /api/models did tag normalization and profile-image stripping for every model before access filtering discarded the invisible ones, and always evaluated a json.dumps debug f-string; the work now runs only on visible models and the debug line is gated on the log level. The duplicate-id dedup keeps its position before filtering so the effective-model semantics are unchanged.
Benchmark:
| metric | before | after |
| --- | --- | --- |
| model-cache fingerprint, 200 models | 45 us | 1.4 us |
| action priority Valves builds, 200 models x 4 global actions | 0.37 ms (800 builds) | 0.002 ms (4 builds) |
| function-table payload for id sets | full rows incl. source | (id, is_global) tuples |
Functionally verified: the column-only id query matches the full-row query for actions and filters including inactive exclusion, and the fingerprint skip logic writes on first set, skips identical payloads, updates plus deletes stale keys on change and clears on empty, against a scripted fake Redis.
Tools.get_tools(defer_content=True) contained the literal dead statement "stmt = stmt": the deferral was a no-op, so every tools listing loaded the full Python source of every tool (five caller sites pass defer_content=True expecting the optimization: the tools list endpoints and the user and group permission overviews). The listing now selects every column except content, and ToolModel.content becomes optional to represent deferred rows; router projections are content-less response models, so nothing downstream reads the source on these paths.
On top of that, get_tools_by_user_id issued one grant query per non-owned tool and Knowledges.get_knowledge_bases_by_user_id did the same per knowledge base (the latter also sits inside per-file access checks). Both now resolve grants for all non-owned rows in a single get_accessible_resource_ids call, the same batch helper the model listing already uses.
Benchmark (real SQLite DB):
| metric | before | after |
| --- | --- | --- |
| tools listing, 33 tools x ~200 KB source | 5.13 ms | 3.18 ms |
| grant queries per accessible-tools call, N non-owned tools | N | 1 |
| grant queries per accessible-KBs call, N non-owned KBs | N | 1 |
The listing row scales with source size; on Postgres the deferral additionally avoids shipping every tool's source over the wire per listing, and each removed grant query was a real round trip.
Functionally verified: deferred listings match full listings field for field with content None, grants included and router projections working; access filtering returns exactly owned plus granted tools and knowledge bases and nothing for strangers; full (non-deferred) reads still carry the source.
Knowledges.get_file_metadatas_by_id fetched full File rows, whose data column carries the entire extracted text of each document, validated each into a FileModel and then threw everything except id, hash, meta and the timestamps away. The function backs every knowledge base detail view and runs again after every file add or remove (eight call sites in the knowledge router), so rendering a filename list for a 50-file knowledge base parsed tens of megabytes of JSON per request.
The listing now selects exactly the five columns the response needs, joined through KnowledgeFile, mirroring the column-only helper that already existed in the files model for id-based lookups.
Benchmark (real SQLite DB, 50 files with ~200 KB extracted text each):
| metric | before | after |
| --- | --- | --- |
| KB file metadata listing | 14.2 ms | 1.20 ms |
The gap widens linearly with file size and count since extracted contents no longer get read, parsed or validated at all.
Functionally verified: output matches the old implementation field for field on all 50 files and an unknown knowledge base still returns an empty list.
has_access_to_file runs for every non-owner file GET, per RAG file check and per shared-chat or model-attached file. Its final step called Models.get_models_by_user_id, which issued one grant query per non-owned workspace model, so a single file check on an instance with M workspace models cost M grant queries plus a group query, with the deny path always paying full price. Its collection_name step listed every knowledge base the user can access (itself one grant query per knowledge base) just to scan the list for one id. And get_accessible_folder_files repeated the whole pipeline per folder entry, refetching the caller's group memberships every time.
Three changes, all using parameters and helpers that already exist:
- Models.get_models_by_user_id resolves grants for all non-owned models in one get_accessible_resource_ids call and accepts prefetched user_group_ids.
- The collection_name check fetches the one referenced knowledge base and performs a single owner-or-grant check with the already-resolved group ids, preserving the write-requires-owner guard exactly (including its short-circuit before any grant query).
- get_accessible_folder_files resolves group ids once and threads them through every per-entry check.
Benchmark:
| metric | before | after |
| --- | --- | --- |
| filter loop CPU, 300 workspace models (queries stubbed) | 47 us | 19 us |
| grant queries per file-access check, M workspace models | M | 1 |
| group membership queries per folder listing, F files | F | 1 |
The stubbed CPU row understates the win: each removed query in the other two rows was a real database round trip.
Functionally verified with stubbed accessors: owned plus granted models are returned with owned ids excluded from the batch query; model-attached file access resolves through the batched path; the collection_name path does one KB fetch and one grant check with no full listing; a missing KB falls through; write access via a KB still requires the KB owner to own the file and short-circuits before the grant query; folder listings fetch groups exactly once.
Two independent sources of fixed per-request cost:
The async SQLite engine was created with pool_pre_ping=True. A pre-ping guards against server connections dropped by timeouts or restarts, which cannot happen to a local SQLite file; each ping still costs a hop into the aiosqlite worker thread plus a SELECT 1 on every connection checkout, and with session sharing off a single request checks out a connection for every model-layer call it makes. The Postgres engines keep their pre-ping, where it is actually protective.
CommitSessionMiddleware unconditionally ran ScopedSession.commit() plus remove() after every HTTP request. The scoped registry instantiates a session on first access, so on the vast majority of requests (which never touch the sync session, per the middleware's own docstring) this built a Session, opened and committed an empty transaction and tore everything down for nothing. The middleware now checks ScopedSession.registry.has() first: requests that used the sync session are committed and removed exactly as before, on success and on the rollback path alike, and idle requests skip the machinery entirely.
Benchmark (real SQLite database):
| metric | before | after |
| --- | --- | --- |
| user row fetch incl. session + connection checkout | 681 us | 514 us |
| idle-request sync session work (create + empty commit + teardown) | 12.3 us | 0.26 us |
The first row saves per model-layer call, not per request: a request making five DB calls saves the checkout ping five times.
Functionally verified: normal reads and writes work with pre-ping off; an idle request through the middleware leaves no sync session behind; a request that uses the sync session still gets committed and removed.
The branch that normalizes plain JSON error lines from streaming upstreams (lines without the SSE data: prefix) called Chats.upsert_message_to_chat_by_id_and_message_id without await. The coroutine was never executed, so the error was never written to the chat and Python emitted a "coroutine was never awaited" RuntimeWarning instead. The frontend still received the error event, but after a reload the message showed no trace of the failure.
The parallel error-persist branch further down the same handler already awaits the call; this aligns the two.
Several spots in the chat pipeline issued sequential single-key config
SELECTs, or fetched the same key twice back-to-back, on every request:
- chat_completion_tools_handler: task model default/external and the
tools prompt template were four sequential Config round trips (the
template was fetched twice). One batched Config.get_many now serves
all of them.
- chat_completion_files_handler: the six RAG settings (top_k,
top_k_reranker, relevance_threshold, hybrid_bm25_weight,
enable_hybrid_search, full_context) were six sequential round trips
inside the retrieval call. Batched into one get_many.
- Voice and code-interpreter prompt templates were each fetched twice
within one conditional; fetch once and reuse. The code-interpreter
engine was likewise fetched twice per execution.
- Skill resolution fetched the accessible-skills list, kept only the
ids, then re-fetched each mentioned skill by id (N+1). Reuse the
rows from the access query.
Value semantics are identical: get_many applies the same defaults as
the individual gets, and the pre-existing truthiness/empty-string
checks on templates are preserved exactly.
Claude-Session: https://claude.ai/code/session_01MHg5zs1VBjvRWQ54qHpfYD
Co-authored-by: Claude <noreply@anthropic.com>
* feat: expose LDAP group sync settings in admin config
LDAP group synchronization was already wired into the login flow but its
settings (group management, auto-creation, and the group attribute) could
only be set via environment variables. OAuth, by contrast, exposes its
group-mapping settings through the admin config API and UI.
Bring LDAP to parity:
- Add enable_group_management, enable_group_creation and
attribute_for_groups to LdapServerConfig and LDAP_SERVER_CONFIG_KEYS so
the /admin/config/ldap/server endpoint reads and persists them.
- Add a "Group Mapping / Auto-Create Groups / Group Attribute" section to
the LDAP admin settings UI, mirroring the OAuth group-mapping controls.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TtCvvQ7dcadoufbRpCKcpe
* fix: harden LDAP group sync config and login flow
Address review findings on the LDAP group-sync settings:
- ldap_auth: move the auto-create-groups call inside the try/except that
wraps group sync, so a group-creation error is logged instead of
bubbling to the broad handler and failing the whole login.
- update_ldap_server: reject saving with group management enabled but an
empty group attribute, which would otherwise make sync silently no-op
(mirrors the existing required-field validation).
- Authentication.svelte: merge the LDAP server config response into the
client defaults instead of replacing the object, so any key an older
backend omits keeps its default value.
Note: the empty-directory-groups behavior was reviewed and already
matches OAuth (both skip removal when no groups are returned), so it was
left unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TtCvvQ7dcadoufbRpCKcpe
* fix: default blank LDAP group attribute to memberOf before save
The Group Attribute field advertises "Default to memberOf", but the
backend now rejects an empty group attribute when group management is
enabled. Fall back to the memberOf default client-side when the field is
left blank, so the advertised default holds and the save isn't rejected.
The backend validation remains as defense-in-depth for direct API calls.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TtCvvQ7dcadoufbRpCKcpe
* fix: initialize LDAP port default as null instead of empty string
The backend LdapServerConfig types port as `int | None`, but the frontend
initialized it to an empty string. If a save carried that default (e.g.
when the backend response omits port under version skew), Pydantic would
reject the empty string. `null` matches the model and is also what the
type="number" input yields when the field is empty.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TtCvvQ7dcadoufbRpCKcpe
* fix: parse LDAP group DNs correctly instead of splitting on commas
Group CN extraction split the DN on raw commas and sliced off "CN=",
which mangles any group whose name contains an escaped separator (e.g.
"CN=Sales\, EMEA,OU=...") into a truncated, wrong name that then fails to
match the intended Open WebUI group. Use ldap3's parse_dn to split the DN
respecting RFC 4514 escaping, and unescape the resulting value so the CN
matches what an administrator sees.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TtCvvQ7dcadoufbRpCKcpe
* chore: address review feedback on _unescape_ldap_dn_value
Trim the docstring and rename the loop index to a more descriptive name
(i -> pos) per review feedback on the group DN unescaping helper. No
behavior change.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TtCvvQ7dcadoufbRpCKcpe
---------
Co-authored-by: Claude <noreply@anthropic.com>
The middleware code-interpreter path defines `restricted_import` as `async def` and assigns it to `builtins.__import__`, which Python's import machinery calls synchronously. Calling an async function returns a coroutine without running its body, so the blocklist check never executes and `_real_import` is never called. When `CODE_INTERPRETER_BLOCKED_MODULES` is set, blocked modules are therefore not blocked, and every subsequent import inside the interpreter binds a dangling coroutine instead of the module, breaking legitimate imports as well.
Define the hook as a regular `def`, matching the working implementation in `tools/builtin.py`. A blocked top-level import now raises `ImportError`, and all other imports pass through to the real importer.
Mirrors the ENABLE_FORWARD_USER_INFO_HEADERS pattern already used by
the audio/TTS and external document loader integrations, so the
Mistral OCR backend can identify the requesting user the same way.
Co-authored-by: andrep <vpham@aut.ac.nz>
The per-request Ollama handlers (chat, generate, embed, embeddings,
and the OpenAI-compat completions/chat-completions/messages/responses
endpoints) fetched 'ollama.api_configs' up to three times and
'ollama.base_urls' separately within a single request — the .get()
default-argument pattern made the second api_configs fetch
unconditional, and get_api_key() triggered a third. Up to four
sequential SELECTs per request collapse to one.
A new get_ollama_connection_config() helper fetches base_urls and
api_configs together in one batched Config.get_many where both are
needed; handlers that only need api_configs fetch it once into a
local. Admin operations (pull/push/copy/delete) and the TTL-cached
model-list path are deliberately left untouched.
Resolution semantics (str(idx) key first, url-key legacy fallback,
same defaults) are unchanged.
Claude-Session: https://claude.ai/code/session_01MHg5zs1VBjvRWQ54qHpfYD
Co-authored-by: Claude <noreply@anthropic.com>
SecurityHeadersMiddleware called set_security_headers() on every
response — 14 os.environ.get lookups plus a regex validation per
configured header, for values that are static for the process
lifetime. Compute the header list once at construction; when no
security env vars are set, skip wrapping send entirely.
RedirectMiddleware decoded and parse_qs'd the query string of every
GET, though it only acts on /watch?v= and ?shared= URLs. Add a cheap
path/substring precheck first; a false positive just falls through to
the previous full parse, so no redirect behavior changes.
Verified byte-identical responses (status, Location, header values)
against the previous implementations across redirect, passthrough,
and no-env cases.
Claude-Session: https://claude.ai/code/session_01MHg5zs1VBjvRWQ54qHpfYD
Co-authored-by: Claude <noreply@anthropic.com>
The streaming handler rebuilt the full accumulated response with
`content = f'{content}{value}'` on every content delta — a complete
string copy per chunk, making accumulation O(n^2) over the response
length. Use in-place `content += value` for the (universal) str case,
which CPython extends in place, keeping accumulation O(n); the
f-string fallback is preserved for non-str values so coercion
behavior is unchanged.
In the ENABLE_REALTIME_CHAT_SAVE branch, full_output() — which
concatenates the entire accumulated output — was called twice per
chunk (once for the DB upsert, once for the emitted delta). Compute
it once and reuse.
Claude-Session: https://claude.ai/code/session_01MHg5zs1VBjvRWQ54qHpfYD
Co-authored-by: Claude <noreply@anthropic.com>
The Automatic1111 branch of the image generation route called set_image_model whenever the request carried a model field. On this backend set_image_model is not request-scoped: it persists image_generation.model to the global configuration and posts the new sd_model_checkpoint to the shared server, because Automatic1111 holds a single checkpoint instance-wide. A non-admin holding features.image_generation could therefore change the instance-wide image model and the shared backend checkpoint for every user by sending a model on an ordinary generation request, even though the setting is otherwise managed only through the admin-only image configuration route and the frontend never sends this field.
Gate the switch on an admin caller. A non-admin now generates on the currently configured checkpoint and the model field no longer mutates global state; admins retain per-request model switching here and through the image configuration route. Image editing is unaffected, as it selects its model per request without writing global configuration.
The chat action route loaded a Function by its raw action_id and executed its action callable after only checking that the id and the requested model existed. The model list that the client renders actions from resolves each model's actions to the active action-type Functions that are global or assigned to that model, and the action route did not mirror that resolution, so a disabled, unassigned, or wrong-type Function, or an action on a model the caller cannot access, could be reached by calling the route directly.
Gate the route on the same rules the model resolution applies: the Function must be an active action, and for server-resolved models the caller must have model access and the action must be one the model actually surfaces (matched by function id, the prefix of each model actions entry, so single and sub-actions both resolve). Direct connections carry a client-supplied model the caller already owns, so the model-bound checks are scoped to non-direct calls; the active-action check always applies. Executing admin-authored Function code remains intended behaviour — this only keeps the route consistent with which actions each model exposes.
Co-authored-by: komyunghan <komyunghan@users.noreply.github.com>
get_all_models ran four function-table queries: global actions, active
actions, global filters, active filters. Global functions are by
definition (type, is_active=True, is_global=True) — a subset of the
active set — so the global id sets can be derived from the active-rows
queries' is_global flag. Four queries become two, and each dropped
query returned full rows including every plugin's source code.
Also folds the mid-function 'models.default_metadata' read into the
Config.get_many already issued at the top of the function (one fewer
round trip; the existing `or {}` default handling is preserved).
Claude-Session: https://claude.ai/code/session_01MHg5zs1VBjvRWQ54qHpfYD
Co-authored-by: Claude <noreply@anthropic.com>
* perf: cut per-instance CPU cost of shared socket.io Redis pub/sub channel
Profiling a multi-instance deployment (py-spy --gil) showed ~44% of worker
CPU in the socket.io pub/sub listener. Two causes, two fixes:
- Add hiredis so redis-py parses the RESP protocol in C instead of pure
Python (redis/_parsers/resp3.py alone accounted for ~28% of GIL samples;
redis-py auto-selects the hiredis parser when importable).
- Subclass AsyncRedisManager to drop emits whose target room has no local
participants before upstream _handle_emit re-encodes the full packet.
Every instance receives every emit published on the shared channel, so
with N instances all but the hosting one were paying full packet
re-serialization per message just to deliver it to nobody. Broadcasts
(room=None) are unaffected.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Q9CQ9qnp3sZGYQQwztsCJT
* fix: restrict pub/sub emit early-out to string rooms
Adversarial review against python-socketio 5.16.2 found one divergence
from upstream: for a degenerate empty-sequence room (emit to room=[]) on
an instance whose namespace has no local clients, the filter's
get_participants probe raises IndexError from room[0] where upstream
returns silently at the namespace guard and still publishes to Redis.
Open WebUI only ever emits to scalar string rooms or room=None, so the
case is unreachable today; guard on isinstance(room, str) anyway so any
non-string room shape passes through to upstream behavior unchanged.
Every open-webui emit uses a string room, so the fast path still covers
all real traffic.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Q9CQ9qnp3sZGYQQwztsCJT
* Update requirements.txt
* Update pyproject.toml
* Update requirements-min.txt
---------
Co-authored-by: Claude <noreply@anthropic.com>
When an upstream provider rejects a request (e.g. a 400 for a
max_tokens value above the model's ceiling), the actionable error
message was only published to event sinks, which are invisible unless
an event function or webhook is configured. Admins had to query the
provider's API directly to diagnose failures (open-webui#27237).
Add a single log line in publish_model_provider_request_failed — the
chokepoint every upstream failure path (OpenAI-compatible chat,
embeddings, responses, token counting, and Ollama) already routes
through — recording status, provider, url, model, error code, and the
upstream message truncated to 1000 chars. 4xx logs at WARNING, 5xx at
ERROR. Client-facing responses are unchanged, so no additional error
detail is exposed in the chat.
Claude-Session: https://claude.ai/code/session_018VecyiPejru1EVF5yfe2sU
Co-authored-by: Claude <noreply@anthropic.com>
`sharing.folders` is present in `DEFAULT_USER_PERMISSIONS` but absent from the `SharingPermissions` response/update schema, so the admin default and group permission API silently drops it on every round-trip and the setting is never saved. Add `folders: bool = False`, matching the config default (`USER_PERMISSIONS_FOLDERS_ALLOW_SHARING`), restoring parity with `DEFAULT_USER_PERMISSIONS`.
Fixes#27120.
* Change spelling to British English in translation.json
Updated translations to use British English spelling for various terms.
* Update translation.json
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
generate_function_chat_completion rewrote form_data['model'] to the base
model id in place. Because that dict is the same object process_chat_response
later re-submits for the post-tool-call continuation, the continuation was
access-checked against the base model instead of the user-facing preset.
For a public preset over a private/unregistered pipe base, a non-admin's
first message succeeded but the continuation after a native tool call was
denied and silently swallowed, aborting the response (admins skip the
user-only check, so they were unaffected).
Operate on a shallow copy of form_data, mirroring the openai/ollama routers
which already substitute the base model on a copy. The continuation now
re-checks the preset the user actually has access to.
Claude-Session: https://claude.ai/code/session_018toPfJW1hMXAhokGaL43Ep
Co-authored-by: Claude <noreply@anthropic.com>
When a chat is started from the home page, the URL is switched to /c/{id} with
history.replaceState, so the Chat component is never remounted and chatIdProp
stays empty for the lifetime of that view. Both websocket recovery paths,
handleSocketConnect and the chat:active fallback, are gated on chatIdProp and
therefore never run for these chats. After any websocket drop during a response
(mobile backgrounding, VPN or IP change, wake from sleep) the completed response
is never fetched and the chat stays stuck in a loading state until a manual page
refresh. Chats opened directly via /c/{id} recover fine, which is why the bug
only reproduces reliably on freshly started chats.
Gate both recovery paths on the chatId store instead, which is set for every
persisted chat, and let loadChat fall back to it so the recovery reload also
works when chatIdProp is empty. Chats opened via /c/{id} behave exactly as
before and temporary chats stay excluded.
The same gate likely explains the remaining reports in #26315.
Fixes#26844
The system-terminal HTML file preview (FilePreview.svelte, serveUrl branch) rendered served HTML in an iframe that hardcoded allow-same-origin. The terminal proxy serves the file root-relative (same origin as the app) and injects no CSP, and there is no default global CSP, so script in a previewed HTML file executed in the application's own origin and could read localStorage (the session token), enabling account takeover and, for admin or workspace.functions victims, server-side RCE via Functions. The sibling srcdoc branch already gates allow-same-origin behind the iframeSandboxAllowSameOrigin setting (off by default) and injects a CSP; the serveUrl branch never received that defense. Gate allow-same-origin on the serveUrl branch identically, so by default the preview runs at an opaque origin and its scripts cannot reach the parent context. Legitimate HTML preview rendering is unaffected.
Co-authored-by: manus-use <manus-use@users.noreply.github.com>
Enabling ENABLE_MILVUS_MULTITENANCY_MODE with the default MILVUS_URI (embedded Milvus Lite at DATA_DIR/vector_db/milvus.db) fails on the first embedding write: _create_shared_collection calls collection.create_index(RESOURCE_ID_FIELD) with no index params. A Milvus server auto-selects a scalar index type in that case, but Milvus Lite rejects the call with "create_index missing required 'index_type' parameter", so shared collection creation raises and every embedding write 500s (memory add, file upload, knowledge writes).
Keep the parameterless call as the first attempt so behavior on Milvus servers is unchanged, fall back to an explicit INVERTED scalar index, and if that also fails log a warning and continue. The scalar index only accelerates resource_id filters; inserts and filtered queries work without it, so a missing index must not break collection creation.
Verified against embedded Milvus Lite: shared collections now create (with the warning), and memory add, file upload and memory query succeed end to end. Against a Milvus server the first attempt is identical to the current code, so nothing changes where it works today.
SecurityHeadersMiddleware was the last middleware in the stack still
subclassing BaseHTTPMiddleware, after CommitSession, AuthToken,
WebsocketUpgradeGuard and Redirect were all moved to pure ASGI in
utils/asgi_middleware.py. BaseHTTPMiddleware re-buffers the response
body through an anyio task group, which has known issues with
streaming and Content-Length-bearing responses (e.g. the FileResponse
returned by /api/v1/audio/speech).
Reimplement it as a pure-ASGI middleware that stamps the configured
security headers onto the http.response.start message via
MutableHeaders and forwards all body chunks untouched, matching the
pattern already used by its four siblings. set_security_headers() and
all its helpers are unchanged.
Co-authored-by: classic298 <classic298@users.noreply.github.com>
The sidebar time-ago indicator rendered `created_at`, so the relative time stayed pinned to the chat's creation age and never reflected new activity. After sending a message the chat would jump to the top of the list (which sorts by `updated_at`) while still showing a stale label such as "3w", which is confusing.
The indicator was originally added using `updated_at` and was inadvertently switched to `created_at` during a later refactor. Restore `updated_at` (falling back to `created_at` when absent) so the timestamp matches the list ordering and updates whenever a chat is modified.
Fixes#26451
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
- Move setModels() and setToolServers() out of the loaded gate in
+layout.svelte so the page renders immediately instead of blocking
behind slow/unreachable connection timeouts
- Track failed Ollama backend URLs from /api/tags and skip them in the
subsequent /api/ps fan-out, eliminating a redundant second timeout
- Silence the autocomplete toast that fires during the transient empty
model state
Translate all remaining untranslated strings in the German locale: 228 empty
values plus the genuinely translatable members of the value-equals-key set (240
entries total). Covers webhooks/events, external knowledge sources, context
compaction, terminal servers, OAuth/OIDC claims, skills, and directory
management.
Translations follow the existing locale conventions: formal "Sie" address,
established German terms (Websuche, Wissen, Werkzeuge, Wissensspeicher), and
brand/proper nouns, format names, and identical-in-German words left untouched.
Documents (especially AI/LLM documentation) legitimately contain special token strings like <|endoftext|> as literal text. The tiktoken encoder raises a ValueError when encountering these during chunk size measurement in merge_docs_to_target_size(), preventing the entire file from being indexed. Pass disallowed_special=() to encoding.encode() to treat all text as normal content.
Audit of asyncio.sleep vs time.sleep and event-loop-blocking calls:
- utils/plugin.py: run pip `install_frontmatter_requirements`
(subprocess.check_call) via asyncio.to_thread in load_tool_module_by_id,
load_function_module_by_id, and install_tool_and_function_dependencies.
- retrieval/utils.py: move the synchronous SSRF-guarded requests probe and
loader.load() in get_content_from_url into a sync helper run via
asyncio.to_thread.
- routers/audio.py: write uploaded audio to disk off the event loop in
transcription().
- routers/pipelines.py: write uploaded pipeline file off the event loop in
upload_pipeline().
The existing time.sleep call sites are all in genuinely synchronous
functions (sync requests/DB drivers/daemon threads) with async
counterparts that already use asyncio.sleep, so no time.sleep -> asyncio.sleep
changes were needed.
Claude-Session: https://claude.ai/code/session_01LXR5bYfsfSS42RGHQZu2Ta
Co-authored-by: Claude <noreply@anthropic.com>
Adds a {{USER_AGENT}} custom-header placeholder that relays the inbound
client's User-Agent to upstream model backends, so providers see the real
client instead of Open WebUI's internal aiohttp UA. This makes upstream
usage/cost attribution and backend telemetry possible, and is opt-in
per-connection (no global flag): admins add {{USER_AGENT}} to a connection's
custom headers in Admin > Settings > Connections.
The placeholder is sourced from the live inbound request (with a metadata
fallback for detached RAG/tool calls), so it resolves on every prompt-sending
path, not just chat completions:
- OpenAI completions, Responses API, and proxy — all route through
get_headers_and_cookies, which now passes the request into get_custom_headers.
- Anthropic Messages API (/api/v1/messages) — already covered, it delegates
to the chat completion handler.
- Ollama (/api/chat, /v1/completions, /v1/chat/completions, /v1/messages,
/v1/responses) — previously had no custom-header support at all; send_request
now applies per-connection custom headers (with templating) for every
Ollama prompt endpoint.
Custom headers are applied after the built-in user-info headers so explicit
admin-configured headers take precedence. The other existing placeholders
({{CHAT_ID}}, {{USER_ID}}, ...) now also work on the newly covered paths.
Frontend: the connection editor's Headers field is now shown for Ollama
connections too (previously gated to non-Ollama), so the placeholder can be
configured there.
Ref: open-webui/open-webui#26159
Add a custom MIME type (application/x-open-webui-drag) to intentional
chat and folder drag sources. The onDragOver handler in MessageInput
now checks for this type instead of the generic text/plain, which
SortableJS also sets during reorder operations for pinned menu items
(Notes, Workspace) and pinned Models.
* feat(ui): add drag-to-reorder for pinned notes in sidebar
* feat(chat): support folder, note, and model drag-and-drop from sidebar as context references
---------
Co-authored-by: Tim Baek <tim@openwebui.com>
get_protected_resource_metadata() only attempted RFC 9728 discovery when
the anonymous `initialize` probe returned 401 with a WWW-Authenticate
header. Some remote MCP servers — notably Google's Gmail/Drive/Calendar
MCPs (gmailmcp.googleapis.com, etc.) — answer 200 to an anonymous
initialize, so OAuth scope and authorization-server discovery silently
failed and connections to them could not be established.
Run the discovery regardless of the probe's HTTP status: prefer the
resource_metadata URL from WWW-Authenticate when present, and otherwise
fall back to the RFC 9728 §4.2 well-known URIs. The trade-off is a couple
of extra well-known GETs during MCP connection setup for servers that
expose no PRM document; behavior for 401-responding servers is unchanged.
The base model selector in the model creation workspace listed every
model from the global store, ignoring the per-model visibility (hidden)
flag. Models marked as not visible were correctly hidden from the chat
model selector but still appeared in the workspace 'Base Model (From)'
dropdown.
Filter out hidden models (info.meta.hidden) to match the chat selector
behavior, while still keeping a model visible if it is the currently
selected base, so editing a preset built on a now-hidden base does not
show a blank dropdown.
Fixes#25665https://claude.ai/code/session_01CUvzYN9DjvwTpNCT5QL33B
Co-authored-by: Claude <noreply@anthropic.com>
routers/openai.py and routers/ollama.py decorated get_all_models with
`@cached(key=lambda _, user: ...)`. In aiocache 0.12, `key=` is a STATIC
cache key: get_cache_key returns `self.key` verbatim (the lambda object)
without calling it, so every caller collides to ONE shared entry within
the TTL. The intended per-user namespacing never happened — one user's
permission-filtered model list could be served to another user (or an
anonymous caller) during the cache window.
The per-call hook is `key_builder=` (called as key_builder(func, *args,
**kwargs)). Switch both sites to key_builder with a (func, request,
user=None) signature so the key is built per call. user=None mirrors
ollama's optional-user signature and stays correct whether user is passed
positionally, as a kwarg, or omitted.
Verified offline: old form returns the same key object for distinct users
(collision); new form yields distinct openai_all_models_<id> /
ollama_all_models_<id> keys, and the unauthenticated base key when no user.
These two were the only @cached(key=lambda ...) sites in the backend.
The Ollama model upload and download handlers perform three stages
of sync file I/O on multi-GB model files inside async handlers:
1. Persist upload — file.file.read() + write() in a loop
2. SHA-256 hash — calculate_sha256() reads the file sequentially
3. Read for blob push — open().read() loads entire model into memory
All three stages block the event loop for the duration of the I/O.
For a 4GB model, each stage freezes the event loop for 10+ seconds
while every other user's request stalls.
Wrap each blocking stage in asyncio.to_thread() in both handlers:
- upload_model(): persist upload, calculate_sha256, blob read
- download_file_stream(): calculate_sha256, blob read
Benchmark (full pipeline: write + SHA-256 + read back, 3 trials):
- 200MB: max jitter 145ms → 1ms (145x improvement)
- 500MB: max jitter 1,026ms → 1ms (1,026x improvement)
File I/O is pure I/O — no GIL contention. asyncio.to_thread()
completely eliminates event loop blocking.
GET /api/v1/channels/{id}/messages/{message_id}/thread authorized only the URL channel, but get_messages_by_parent_id() appended the thread parent (loaded by id) without checking it belonged to that channel, so a caller could read a message from a channel they cannot access by passing its id as the thread root. Require the parent to be in the requested channel before returning it, and reject a posted parent_id/reply_to_id that does not belong to the channel.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Wrap get_content_from_url() in asyncio.to_thread() inside
get_sources_from_items() to prevent sync requests.get() from blocking
the event loop when users attach URL sources to chat messages.
The same function was already properly wrapped at retrieval.py:1839.
- Wrap hashlib.sha256(contents).hexdigest() in asyncio.to_thread()
inside upload_file_handler() to prevent CPU-bound hashing from
blocking the event loop during file uploads (44ms/100MB, scales
linearly with file size).
A trailing space in the Open Terminal bearer token broke only the
interactive terminal: HTTP calls put the token in the Authorization
header, where the spec strips trailing whitespace, so the connection
test, file listing and tool calls all worked. The terminal WebSocket
can't set headers from the browser, so it sends the token inside a JSON
auth message that preserves the space verbatim, failing auth with
[Connection closed]. Normalize the key on save so whitespace never
enters storage, and trim it in the WebSocket auth message so existing
saved configs work without re-saving.
Fixes#25613
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Web search results stored as a vector collection were silently dropped
before retrieval when BYPASS_RETRIEVAL_ACCESS_CONTROL is False (the
default). The server-generated web_search file item carries a
'collection_name' but its 'web_search' type is not matched by any
explicit dispatch branch in get_sources_from_items, so it fell through
to the untrusted client-supplied collection_name branch and was ignored.
Add an explicit branch for type == 'web_search' items so the collection
is queried again. Access control is preserved: the collection still
passes through filter_accessible_collections, which already allowlists
web-search-* and bypasses only for admins.
Regression introduced when the retrieval access-control hardening gated
the bare collection_name fallback behind BYPASS_RETRIEVAL_ACCESS_CONTROL.
get_embedding_function runs at import time (main.py), so the empty-engine
+ no-model guard added in 55ca719b raised ValueError during startup and
bricked the instance: a blank embedding model set via the UI made the app
unbootable, with no way back into settings to undo it. Move the check into
the returned coroutine so construction always succeeds and the error only
fires when something actually embeds, surfacing as a clear request error
instead of the old cryptic 'NoneType has no attribute encode'.
Fixes#25634Fixes#25165
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Sign Out button on the Account Pending overlay only cleared the
local token and redirected to /auth, skipping the backend signout. This
left the OAuth/OIDC session alive at the IdP, so re-login bounced the
user straight back into pending and never re-fetched updated role/group
claims. Call userSignOut() and honor the returned redirect_url, matching
the sidebar logout flow.
Fixes#25644
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The file picker and Files modal only render filenames and metadata, but
searchFiles()/getFiles() never passed content=false, so the backend
serialized data.content (full extracted text) of every file into each
response. On instances with many large documents this turns a metadata
list into tens of MB per request, re-issued on every picker keystroke.
The backend already supports content=false on GET /files/ and
GET /files/search; this just makes the list/search callers opt out of the
payload by default. The param stays overridable for callers that need it.
Fixes#25741
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A chat created in place (new chat -> first message) keeps chatIdProp
undefined because the URL is swapped via history.replaceState without a
remount, so none of the chatIdProp-gated updateLastReadAt paths fire and
the chat is never persisted with a last_read_at. Starting another new
chat via initNewChat reset $chatId without marking the outgoing chat
read, so on refresh updated_at > last_read_at (NULL) and the chat shows
as unread in the sidebar.
Mark the outgoing chat read in initNewChat, mirroring navigateHandler.
Fixes#25108
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
validate_url() calls socket.getaddrinfo() for SSRF protection, which
blocks the event loop for 100-700ms per DNS lookup. This affects:
- Image generation (get_image_data) — every external image URL
- Image editing (load_url_image) — every external image URL
- OAuth profile pictures (_process_picture_url) — every login
- Webhook delivery (post_webhook) — every notification
- Image base64 conversion (get_image_base64_from_url) — chat images
Wrap all 5 async call sites in asyncio.to_thread() so DNS resolution
runs in the thread pool. The event loop remains free to serve other
requests during the lookup.
Benchmark (3 domains, 3 trials averaged):
- BEFORE: max jitter 479ms, 1 blocked ping per trial
- AFTER: max jitter 1ms, 0 blocked pings (324x improvement)
Co-authored-by: Tim Baek <tim@openwebui.com>
get_event_call() routed execute:python / execute:tool events to a client-supplied session_id after only checking the session was connected, not that it belonged to the requester. Verify the target session is owned by the requesting user (metadata user_id) before delivering, so a client cannot route code/tool execution into another user's session.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Azure TTS handler (_tts_azure) interpolated the user-supplied voice,
and the locale derived from it, into the SSML xml:lang and <voice name>
attributes without XML-escaping, while the text body was already escaped
(2e75c6dbd). Escape both attributes too, so every user-derived value in
the SSML document is consistently encoded.
Co-authored-by: alanturing881 <alanturing881@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(calendar): add repeat/recurrence dropdown to event modal
* fix(calendar): anchor recurring event expansion to event start date
The expand_recurring_event() utility used the view range start as dtstart,
causing FREQ=WEEKLY events to land on the wrong day of the week. Use the
event's original start date instead so recurrence patterns stay correct.
The web-ingest probe in get_content_from_url validated the URL at resolve
time (validate_url) but then fetched with a bare requests.get, which
re-resolves the hostname at connect time. An attacker-controlled name
server can answer with a public IP during validation and an internal IP
at connect, reaching cloud metadata / loopback / internal services (blind
always; binary content-types are read back to the caller). The
connection-layer guard (#24759) that closes this for the SafeWebBaseLoader
path was never mounted on this probe.
Route the probe through the same _SSRFSafeAdapter the loader uses, so the
resolution that feeds the TCP connect is re-validated against the
global-IP check.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
has_collection did `collection_name in self.client.list_collections()`,
but chromadb's list_collections() returns Collection objects (1.x), not
name strings — so the membership test is always False, even when the
collection exists. Compare against the collection names instead (with a
hasattr guard tolerating versions that yield plain names).
Found via the dependency-contract test suite (unit/deps/test_chromadb.py).
* Update SECURITY.md
* Update SECURITY.md
* Update SECURITY.md
* Update SECURITY.md
* Update SECURITY.md
* Update SECURITY.md
* Update SECURITY.md
* Update SECURITY.md
* Extend the already-fixed/monitoring rule to public PRs and credit
Broaden the rule from "already fixed" to also cover issues already being fixed in
the open (e.g. an open pull request), extend the commit-monitoring pattern to PRs,
and fold in the credit consequence on provable grounds rather than an unprovable
bad-faith claim: a report of an already-public, already-fixed-or-being-fixed issue
filed strictly last is a duplicate we cannot distinguish from scraping, so it earns
no advisory. Credit belongs to whoever found or fixed it, who forfeits it by
disclosing publicly instead of reporting confidentially first — so a public fix
earns no advisory and no credit for anyone.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Remove Rule 14 (One Vulnerability Per Report)
The one-CVE-per-vulnerability constraint it restated is a CVE Program counting
rule, already binding through the "Alignment with the CVE Program" section.
Dropping the standalone rule removes the duplication; bundled reports are still
split on that basis when they arise.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* security policy: surface "What a Valid Report Gets You" near the top, refresh date
Move the "What a Valid Report Gets You" section up to directly under the good-faith
reporting section (it leads with what reporters receive, rather than burying it
below the rules), and update the last-updated date.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Update SECURITY.md
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
_upload_file_async built its multipart body with
`aiohttp.streams.FilePayload(...)`, which no longer exists in aiohttp
(payload classes live in aiohttp.payload, and there is no FilePayload).
The reference sits inside a lazy closure, so import succeeds and only the
async Mistral OCR file-upload path blows up at runtime with
AttributeError on every call.
Mirror the working sync path: open the file in a context manager and let
MultipartWriter.append(file_obj, {...}) build a streaming
BufferedReaderPayload, with the POST issued inside the open() block so
the handle stays valid for the whole upload. Preserves the streaming /
memory-efficiency intent.
Found via the dependency-contract test suite (unit/deps/test_aiohttp.py),
which pins aiohttp.streams.FilePayload as absent.
constants.py subclassed opentelemetry.semconv.trace.SpanAttributes, which
is deprecated since semconv 1.25.0 (emits a DeprecationWarning; the pinned
0.63b1 has it live). Source the legacy span-attribute keys from the
non-deprecated _incubating attribute modules instead:
http.url / http.method / http.status_code <- _incubating http_attributes
db.name / db.statement / db.operation <- _incubating db_attributes
The stable http module renamed these (http.request.method, ...), so only
the incubating module preserves the original values. Custom keys
(db.instance/type/ip/port, error.*, result.*) stay literals.
Verified: emitted attribute keys are byte-identical before/after for every
key instrumentors.py reads, and importing constants no longer emits a
DeprecationWarning.
Replace per-message database queries with batch IN-clause queries in
channel message handlers. This eliminates the N+1 query pattern that
caused ~102 queries per channel page load (50 messages × 2 queries each).
Changes:
- Add get_reactions_by_message_ids() to MessageTable: single query
fetches all reactions for multiple messages using IN clause with
User JOIN, returns dict[message_id, list[Reactions]]
- Add get_thread_reply_counts_by_message_ids() to MessageTable: single
GROUP BY aggregate query returns (count, max_created_at) per parent,
replacing full object loads just to call len()
- Refactor get_channel_messages(): 102 → 4 queries per page
- Refactor get_pinned_channel_messages(): 22 → 3 queries per page
- Refactor get_channel_thread_messages(): 53 → 4 queries per page
- Refactor send_notification(): N+1 membership check → batch set lookup
pydub imports the stdlib `audioop`, removed in Python 3.13, so audio
preprocessing would break there. requires-python is already capped at
< 3.13; this one-line pointer flags what to handle (audioop-lts, or drop
pydub) before raising that cap.
* chore: bump Python backend dependencies, drop unused peewee
Minor/patch + reviewed major bumps across requirements.txt,
requirements-min.txt, pyproject.toml and uv.lock; playwright image bumped in
docker-compose.playwright.yaml. peewee/peewee-migrate removed (zero imports).
Security-relevant: cryptography 46->48, authlib 1.6.10->1.7.2, PyJWT 2.11->2.13,
requests 2.33.1->2.34.2, RestrictedPython 8.1->8.2, pillow 12.1.1->12.2.0.
Reviewed majors: redis 7->8, pymilvus ->2.6.14, azure-search-documents 11->12,
chardet 5->7, unstructured 0.18->0.22, pycrdt 0.12->0.13.
Testing:
- Resolution: `uv lock` resolves the full bumped set with no conflicts; uv.lock
regenerated to match (peewee dropped, every pin including
azure-search-documents==12.0.0 resolves).
- Per-dependency contract tests (external tests repo, unit/deps/): 105 files,
2205 passed / 6 skipped, ruff-clean. One file per dependency pins the symbols,
signatures and behaviour the backend actually uses, so an API removal/rename in
a bumped version fails loudly instead of at runtime. Offline/deterministic.
- End-to-end embed->retrieve test driving transformers + sentence-transformers +
chromadb together through Open WebUI's real RAG path (cached model, in-memory
chroma, semantic retrieval asserted).
- Install/startup/health resolution gate added to the dep-bump workflow and the
integration suite (uv/pip resolve + uvicorn /health + Playwright dev visibility).
- Bugs surfaced while testing each got an isolated fix branch + regression test:
Mistral OCR aiohttp FilePayload (#25779), chroma has_collection (#25780),
aiocache per-user model-cache key (security), otel semconv deprecation,
pydub/audioop <3.13 note.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Bump python-multipart 0.0.22 -> 0.0.27 (CVE-2026-42561, CVE-2026-40347)
0.0.22 is affected by two DoS CVEs in the multipart parser that
Starlette/FastAPI run for every multipart/form-data request, so any
authenticated user hitting an upload endpoint can trigger them:
- CVE-2026-42561: unbounded part-header count/size -> CPU exhaustion (fixed 0.0.27)
- CVE-2026-40347: large multipart preamble/epilogue DoS (fixed 0.0.26)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Add a security-reporting link to the issue chooser
With blank issues disabled, the New Issue chooser only offered Bug Report and
Feature Request, leaving security reporters no obvious route and nudging them
toward filing vulnerabilities as public issues. Add a contact_links entry that
points to the Security Policy (/security/policy), where the "Report a
vulnerability" button opens a private GitHub advisory — keeping security
reports out of public issues.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Update config.yml
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the recursive parseJSONString with an equivalent iterative unwrap.
The recursion re-parsed its own already-parsed result until JSON.parse threw;
on scalar JSON values (e.g. "5" -> 5) that recursed until a stack overflow
which was then silently caught — wasted work on every complete tool-call
payload. The loop returns the identical value in all cases (verified
byte-identical across 29 inputs incl. double/triple-encoded and partial JSON)
without the stack churn.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
- Append Unicode text variation selector U+FE0E to the ↩ character in
ShortcutItem.svelte to prevent iOS/iPadOS rendering it as a colorful
emoji instead of a plain text glyph.
- Fix 'Keyboard shortcuts' → 'Keyboard Shortcuts' (capital S) in the
user menu to match the modal title and standard title case.
Pressing Enter to confirm an IME composition (e.g. Japanese, Chinese, or Korean)
in the chat search modal and the sidebar search input was treated as a normal
Enter and triggered a search action such as starting a new chat, instead of just
confirming the composed text.
Guard the search keydown handlers so a keydown fired while composing is ignored:
- SearchInput.svelte on:keydown
- SearchModal.svelte document-level onKeyDown
using e.isComposing (with a keyCode === 229 fallback), mirroring the IME guard
already used by MessageInput. A second Enter, after the composition is confirmed,
still triggers the action as before.
Fixes#26172
Co-authored-by: Yogi <251807336+ybnjm1234-source@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
has_access_to_file() derives file access from the objects a file is attached to
(knowledge bases, workspace models). Those branches returned True for any access_type
whenever the user held that permission on the object, write/delete included. Since a
user can create their own KB or model and attach any file they can merely READ (KB
attach and the model meta.knowledge validator both gate on read access only), a user
with read access to a victim file could launder it into write/delete: attach it to an
object they own, then rename, overwrite or delete it via the write-gated file routes
(POST /files/{id}/rename, /data/content/update, DELETE /files/{id}). This is the
residual of GHSA-vjqm-6gcc-62cr (CVE-2026-54012) left open by the read-only attach
validator (CWE-863).
An object now confers write/delete on a file only when the object's owner owns that
file, so delegation originates from the file's own owner. Read is unchanged (RAG and
shared-object reads still work), and legitimate delegation is preserved: a write grant
on an object whose owner owns the attached file still confers write. Applied to all
three object branches: knowledge base, file home collection, and workspace model.
Co-authored-by: rexpository <30176934+rexpository@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The direct image-edit route was the only image-edit surface with no authorization: it ran
on get_verified_user alone, while POST /generations enforces ENABLE_IMAGE_GENERATION +
features.image_generation and the built-in edit_image tool enforces ENABLE_IMAGE_EDIT +
features.image_generation. A verified non-admin user could therefore reach the configured
image-edit provider (spending IMAGES_EDIT_OPENAI_API_KEY) even when the administrator had
globally disabled image editing (ENABLE_IMAGE_EDIT=False) or denied the user image
generation.
Split the route from the shared impl (mirroring generate_images/image_generations): the new
/edit wrapper enforces ENABLE_IMAGE_EDIT and the per-user image-generation permission, then
delegates to image_edits(). The internal callers (the edit_image tool and the chat
middleware) already gate themselves and call image_edits() directly, so they are unaffected.
Co-authored-by: jagstack <52110932+jagstack@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds a 'Reset to Defaults' button to both the Edit Default Permissions
modal and the Edit User Group permissions modal.
- Default permissions modal: resets to stock/env-var configuration
- User group modal: resets to current global default permissions
Backend: new GET /api/v1/users/default/permissions/defaults endpoint
that returns DEFAULT_USER_PERMISSIONS (env-var-based initial defaults).
ws_terminal() interpolated the path parameter session_id directly into the upstream
terminal WebSocket URL and then appended ?user_id=<caller>, with no encoding or
validation (the HTTP sibling proxy_terminal runs _sanitize_proxy_path; this path ran
nothing). An encoded '?'/'&' smuggled through session_id survives Open WebUI's single
decode and is re-decoded by the upstream, injecting an attacker-chosen user_id ahead
of the appended one. Query parsing binds the first occurrence, so the orchestrator
resolves the spoofed user's terminal scope, letting a normal authenticated user
present another user's identity to the upstream (CWE-116/863).
Encode session_id as an opaque path segment with urllib.parse.quote(session_id,
safe=''). This neutralises '?'/'#'/'&' at any decode depth (the upstream's single
decode reverses only the quote, leaving the original delimiters inert as path
content), while legitimate UUID session ids pass through unchanged. The appended
user_id is then the only query parameter the upstream binds.
The separate concern that the forwarded identity is a bearer claim with no integrity
binding spans Open WebUI and the upstream terminal server and is not addressed here.
Co-authored-by: rexpository <30176934+rexpository@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
These were the only two ydoc Socket.IO handlers missing the SESSION_POOL guard that
every sibling (join/state/update) enforces. With always_connect=True admitting
unauthenticated sockets, a client that knew a note's document_id could broadcast
spoofed awareness (cursors/presence) and fake ydoc:user:left events into a live
editing room, attributed to any client-supplied user_id.
Both handlers now require an authenticated SESSION_POOL session; the awareness handler
additionally requires prior ydoc:document:join (room membership); and the broadcast
user_id is fixed to the authenticated identity instead of the client-supplied value,
removing the impersonation. Document content was never reachable (ydoc:document:update
already enforces write permission).
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The SCIM update_user (PUT) and patch_user (PATCH) handlers mapped the SCIM active field
unconditionally onto the role column (role = 'user' if active else 'pending'), so a routine
IdP sync or a misconfigured IdP that marked a locally-provisioned admin inactive would
silently strip that admin's role and could lock an instance out of its own administration.
Gate both active->role assignments on user.role != 'admin' so SCIM provisioning can
activate/deactivate ordinary users but never demotes an existing admin; admin role changes
continue to go through the dedicated admin endpoints. SCIM already cannot promote to admin
(active only maps to user/pending), so this is symmetric.
Credit to @HOHK0923 for surfacing the admin-demotion footgun.
Co-authored-by: HOHK0923 <118590749+HOHK0923@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(oauth): use Protected Resource Metadata scopes in MCP DCR flow
The Dynamic Client Registration flow seeded the registration request
`scope` from the Authorization Server's `scopes_supported` (RFC 8414),
which is a full catalog of every scope the AS can grant across all
resources. Per RFC 9728 and the MCP Scope Selection Strategy, the
resource-specific Protected Resource Metadata `scopes_supported` is the
correct, least-privilege source.
The PRM is already fetched in this function; this change prefers its
`scopes_supported` and keeps the AS `scopes_supported` only as a
fallback when the PRM advertises none. Mirrors the static-credentials
fix in #24690.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(oauth): trim DCR scope comment per review feedback
Shortens the inline comment per review feedback on #25958.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
is_string_allowed does endswith() matching and was called with the full URL
(retrieval/web/utils.py) against WEB_FETCH_FILTER_LIST, so a blocklisted host with any
path (https://blocked.example/x) ended with /x, not the host, and slipped through; the
allowlist direction false-rejected legitimate URLs and admitted attacker URLs ending in
an allowed string. The same endswith caused label confusion at the hostname call site
(retrieval/web/main.py): corp.com matched evilcorp.com, 10.0.0.1 matched 110.0.0.1.
Add is_host_allowed(host, ...) matching on DNS label boundaries (host == pattern or
host.endswith('.' + pattern)), called with the parsed hostname at both web-fetch call
sites. is_string_allowed is left unchanged for the unrelated function-name filters
(utils/middleware.py, utils/tools.py).
The separate is_global guard (validate_url / _ssrf_safe_new_conn, active when
ENABLE_RAG_LOCAL_WEB_FETCH is off) already blocks RFC1918/loopback/link-local, so this
restores the admin's intended blocking of specific public hosts.
Co-authored-by: addcontent <59762500+addcontent@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The connection-layer DNS-rebinding guard (_SSRFSafeResolver / _SSRFSafeAdapter, PR #24759) was
mounted only on SafeWebBaseLoader. Two user-reachable image fetches validate the URL then fetch
it through the shared get_session() pool with the default resolver, so a TTL-0 rebinding answer
that passed validate_url reaches an internal address at connect:
- get_image_base64_from_url (utils/files.py): user image_url on every chat completion.
- load_url_image (routers/images.py, POST /api/v1/images/edit): user-supplied image field.
Add get_ssrf_safe_session() (a one-off aiohttp session mounting _SSRFSafeResolver) and use it
for both fetches, so the connect-time IP is re-validated and a rebound loopback / RFC1918 /
metadata address is rejected. The shared pool is left untouched for the admin-configured
image-generation callers, which legitimately reach internal hosts.
Co-authored-by: dhyabi2 <32069256+dhyabi2@users.noreply.github.com>
Co-authored-by: geo-chen <2404584+geo-chen@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
process_uploaded_file auto-links an uploaded file to the knowledge base named in
client-supplied metadata.knowledge_id, but it called Knowledges.add_file_to_knowledge_by_id
directly with no authorization, while the dedicated POST /knowledge/{id}/file/add endpoint
gates the same operation on owner / admin / write grant. A read-only collaborator (or any
verified user who knows the KB id) could therefore attach arbitrary files to a knowledge base
they cannot write: the KnowledgeFile membership row was committed, and the later vector-index
write check failed closed but left the row in place.
Gate the auto-link on the same owner/admin/write check before inserting the membership row;
otherwise skip the link and log. The file still uploads as the user's own file, it just is
not linked to a KB they cannot write.
Co-authored-by: jagstack <52110932+jagstack@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
generate_chat_completion() checks model access on line 200 only when bypass_filter is
False. When an arena model reaches this function without a pre-resolved selected_model_id,
which is the task and background path (the /api/v1/tasks/* endpoints call
generate_chat_completion directly rather than through process_chat_payload), the fallback
resolves the arena to an underlying model and recurses with bypass_filter=True, so the
resolved model's access check is skipped. An authenticated user with access to an arena
could therefore reach a model they are denied directly, and for the default or exclude
arena, whose candidate pool is every non-arena model, any model on the instance (CWE-862).
The normal chat path resolves the arena in process_chat_payload before this function, so its
resolved model is checked on line 200; the task path was not, which is the inconsistency.
Enforce check_model_access() on the resolved model in the fallback, before the
bypass_filter=True recursion, mirroring the normal-path check. Admins and already-bypassed
recursive calls are unaffected, and legitimate arena use of accessible models is unchanged.
Co-authored-by: rexpository <30176934+rexpository@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two lifecycle/authorization gaps let a deactivated (pending) account keep acting:
1. The background automation scheduler (execute_automation) rehydrated the owner by ID and
dispatched the chat pipeline without re-checking the owner. A user later set to pending,
or one whose features.automations permission was revoked, kept running scheduled
automations on the operator's provider credentials, even though the HTTP create/update/run
routes already gate on get_verified_user + features.automations. Re-gate the rehydrated
owner before dispatch: require role user/admin and, for non-admins, the features.automations
permission; otherwise record an error and skip the run.
2. check_model_access enforced per-model ACLs only for exactly role == 'user', so any other
non-admin role (a pending principal) fell through and was granted access. Enforce for every
non-admin role (admins still bypass), so the check fails closed (CWE-862, CWE-863).
Co-authored-by: rexpository <30176934+rexpository@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The decode-until-stable loops in _sanitize_proxy_path (terminals.py, cap 8) and
_safe_static_redirect_path (models.py, cap 2) proceed with whatever remains after the
cap instead of rejecting it. A path encoded more times than the cap therefore exits the
loop still percent-encoded, passes the literal '..' / prefix checks (the dots are still
%2E, not '..'), and is forwarded to the upstream terminal server, or emitted as a
redirect Location, which then decodes it once more and resolves the traversal. This is
the residual of the decode-until-stable hardening added for CVE-2026-54017: the cap is a
fixed depth, not a true stability guarantee.
Reject when the value is still not stable after the cap (unquote(x) != x), so anything
encoded more deeply than the cap fails closed rather than being forwarded. Legitimate
paths stabilize within a pass or two and are unaffected.
Co-authored-by: DavidCarliez <271374756+DavidCarliez@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix: block private-IP webhook URLs to close SSRF on caller-controlled URL
post_webhook(url, ...) in utils/webhook.py forwards the URL straight to
aiohttp.ClientSession.post with no SSRF gate. The URL is caller-controlled
on two surfaces:
- User notification settings under ENABLE_USER_WEBHOOKS=true — any
authenticated user can set the URL their notifications POST to.
- Automation notification triggers (calendar alerts, etc.).
Without a gate, the URL can target cloud metadata (169.254.169.254 /
fd00:ec2::254), localhost-bound services, RFC1918 internal hosts, or any
other private address reachable from the server process. Blind SSRF — no
response body returned to the caller — but enough to enumerate internal
services via response timing / status codes, and on cloud deployments
enough to issue requests against IMDSv1 if available.
Call validate_url() at the top of post_webhook. The function blocks
private/reserved IPs when ENABLE_RAG_LOCAL_WEB_FETCH is False (the
default), is the project's chosen SSRF gate, and is already applied to
the equivalent fetch surfaces (retrieval, image-load, OAuth profile
picture). Operators who legitimately need to webhook to private IPs
(internal monitoring, self-hosted Slack alternatives, etc.) can set
ENABLE_RAG_LOCAL_WEB_FETCH=True — same opt-out as the other gated
surfaces.
Scope intentionally limited to webhooks. The OAuth discovery and
external reranker paths cwanglab also flagged are admin-configured with
intentional private-IP defaults (reranker defaults to
http://localhost:8080/v1/rerank) and are out of scope per Rule 9 — the
admin owns the URL choice and the operator opt-out exists for them too.
Reported by cwanglab in GHSA-5x9f-85cg-w3hf (cluster canonical with six
closed siblings: g36v-23gj-j69x, 6j8f-h58v-xgmw, xpwv-52pm-p8hj,
v9gp-hv2c-9qv8, fw7w-jrw7-p3v9, x7xq-74rg-m8mf).
Co-authored-by: cwanglab <cwanglab@users.noreply.github.com>
* fix: also pass allow_redirects=False on webhook post_webhook session.post
Companion to the previous commit. validate_url() only validates the
initial URL; aiohttp's default allow_redirects=True would still follow
a 302 to a private-IP target. Same redirect-bypass class as the rh5x
cluster's five call sites, sixth call site to receive the same gate.
Co-authored-by: cwanglab <cwanglab@users.noreply.github.com>
---------
Co-authored-by: cwanglab <cwanglab@users.noreply.github.com>
The Qdrant backends implemented delete(ids=...) as a payload filter on
metadata.id, but points are stored with the item id as the Qdrant point id
(see _create_points), and not every point carries an id in its payload.
Memory points store only {created_at} in metadata (KB metadata embeddings
likewise), so deleting a single memory matched nothing and left an orphaned
vector that kept being injected into RAG context.
Delete by point id instead: PointIdsList for the standard backend, and a
tenant-scoped HasIdCondition for multitenancy (point ids are unique, so tenant
isolation is preserved). Filter-based deletion is unchanged.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The static credentials OAuth flow currently sets scope=None, relying on
the OAuth provider's default scopes. This breaks providers like GitHub
that default to minimal/public-only access when no scope is requested.
This change reads scopes_supported from the Protected Resource Metadata
document (RFC 9728) and uses them in the authorization request. Unlike
the Authorization Server's scopes_supported (a full catalog of every
scope the AS can grant), the PRM scopes_supported represents what the
specific resource requires — making it safe to request without breaking
providers like Entra ID that reject broad scope requests.
Fixes the regression introduced in 349ea4ea where all scope handling was
removed from the static flow.
* fix: gate chat_completion channel: branch on channel access + message scoping
When chat_id starts with 'channel:' the chat-completion handler skips
the chat ownership / storage block below it. Nothing replaced that
gate. The downstream channel emitter in socket/main.py:_make_channel_
emitter writes to Messages.update_message_by_id using a caller-supplied
message_id pulled from form_data['id'], with no membership check, no
write-access check, and no validation that the message_id belongs to
the channel.
Net effect: any authenticated user could submit
chat_id='channel:<any-channel-uuid>' + id='<any-message-uuid>' and
overwrite that message with attacker-controlled LLM output. Cross-
channel writes worked too — private channels, DMs, channels the
caller has no access to. Original author attribution stayed intact on
the overwritten row.
Add the missing checks at the channel: branch:
1. Channel must exist (404 otherwise).
2. Non-admin caller must have write access to the channel — membership
for group/dm channels, AccessGrants permission='write' for others.
3. The message_id (if supplied) must belong to the same channel — a
caller with write access to channel A cannot use this path to
overwrite a message in channel B.
Behaviour change is limited to callers who were exploiting the gap:
legitimate flows that supply a message_id under their own channel
membership continue to work unchanged.
Co-authored-by: sfwani <sfwani@users.noreply.github.com>
* chore: trim verbose comment on channel: branch gate
---------
Co-authored-by: sfwani <sfwani@users.noreply.github.com>
The query_memory endpoint embeds the search query without the configured
RAG_EMBEDDING_QUERY_PREFIX, while every RAG retrieval path in
retrieval/utils.py correctly passes it. Instruction-tuned embedding
models (e.g. Qwen3-Embedding) produce poor results without the prefix,
causing memory search to return semantically unrelated results.
- Introduced skills functionality in Chat.svelte, MessageInput.svelte, and related components.
- Added SkillsModal for displaying and managing available skills.
- Updated state management to include selectedSkillIds and integrate skills API.
- Enhanced UI to show available skills and their descriptions.
- Updated translations to support skills-related text.
* feat: add support for Valkey vector database
Signed-off-by: Riley Des <riley.desserre@improving.com>
* feat: add CLIENT SETNAME to Valkey vector store connections
Set client_name on GlideClientConfiguration for both the main client
and batch client so connections are identifiable in CLIENT LIST,
monitoring dashboards, and CloudWatch metrics.
Signed-off-by: Riley Des <riley.desserre@improving.com>
---------
Signed-off-by: Riley Des <riley.desserre@improving.com>
The history diff, delete, and version-restore routes authorize the URL
prompt_id but then act on a caller-supplied history/version id without
checking it belongs to that prompt (IDOR). Filter by prompt_id in
compute_diff and delete_history_entry, and reject a cross-prompt version_id
in update_prompt_version.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Firecrawl /search returns either `{"data": [...]}` (flat list — v1, and
what frost19k reported on #23966) or `{"data": {"web": [...]}}` (v2,
current production). The parser only handled the dict shape:
data = response.get('data') or {}
results = data.get('web') or []
On a list-shape response, `data.get('web')` raised AttributeError,
caught by the function's outer try/except, and `search_firecrawl`
silently returned []. Web search worked against v2 endpoints but is one
upstream-format-change away from failing closed again. Accept either.
The native function-calling tool resolver in utils/tools.py applies five
gates before exposing execute_code as a builtin tool: builtin-category
enable, ENABLE_CODE_INTERPRETER global config, model capability,
features.code_interpreter request flag, and the per-user
features.code_interpreter permission.
The legacy XML-tag detection path in streaming_chat_response_handler
applied only the request-flag gate. Brings the legacy path to parity by
running the same five-gate check before activating tag detection.
Behaviour change is limited to deployments that previously relied on
the asymmetry — admins who set ENABLE_CODE_INTERPRETER=False or revoked
the per-user permission, on the legacy tool-calling mode, with the
client supplying features.code_interpreter=true. Any of those three
conditions met now correctly disables tag detection.
Co-authored-by: sfwani <sfwani@users.noreply.github.com>
Open WebUI's collection ACL accepted any unknown name as a
legacy/ephemeral collection. In Milvus multi-tenancy mode that name
becomes the `resource_id` and is interpolated unescaped into a SQL-like
Milvus expression — `resource_id == '<name>'` — so a name like
x' or resource_id != '' or resource_id == 'x
turns the filter into a tautology and returns every tenant's chunks
from the shared collection.
All collection names Open WebUI generates are UUIDs, SHA-256 hex
digests, or fixed-prefix variants of those — they all fit
[A-Za-z0-9_-]. Add a strict format check in
filter_accessible_collections (utils.py) that drops any name outside
that set before any ACL or vector-store lookup, applied even on the
admin bypass path. _validate_collection_access then surfaces the dropped
name as a 403.
As defense in depth, MilvusClient now validates resource_id at every
expression-construction site and escapes single quotes / backslashes in
any other string interpolated into a filter (delete ids, metadata
filter values). Non-string filter values are typed-checked instead of
str()-formatted.
Co-authored-by: Claude <noreply@anthropic.com>
update_chat_by_id re-derived assistant content from `output` on every
save (serialize_output) so frontend edits to output items reflect in
content. But it ran unconditionally, so content set independently of
output — an __event_emitter__ {"type":"replace"} from an Action, or an
outlet filter footer — was reverted to the original output-derived text
on the next save. The reload reads chat.chat directly, so the change
vanished after navigating away (regression vs 0.9.2, which predates the
output mechanism).
Re-derive only when the message's `output` actually changed versus what's
stored, which still reflects genuine output edits but leaves
independently-set content intact.
Fixes#24585
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Symptom
On a fresh install (zero users) the frontend shows the mandatory
"Create Admin Account" onboarding screen, but POST
/api/v1/auths/signup returns 403 ACCESS_PROHIBITED ("You do not have
permission to access this resource."). Wiping the database does not
help when the config layer is backed by Redis (the value survives in
the Redis/valkey volume), or when only the user table is cleared (the
config row survives in Postgres). The instance is then unrecoverable
through the UI.
Root cause
signup_handler() auto-sets request.app.state.config.ENABLE_SIGNUP =
False immediately after the first admin is created. That value is
persisted by the config layer (the Postgres config table, and Redis
when REDIS_URL is set). On a later zero-user database the persisted
False is read back, so ENABLE_SIGNUP resolves False even though no
users exist. The old gate was:
if WEBUI_AUTH:
if not ENABLE_SIGNUP or not ENABLE_LOGIN_FORM:
if has_users or not ENABLE_INITIAL_ADMIN_SIGNUP:
403
ENABLE_INITIAL_ADMIN_SIGNUP defaults to False, so with zero users the
inner test (has_users or not ENABLE_INITIAL_ADMIN_SIGNUP) is True, and
a stale ENABLE_SIGNUP=False trips the outer test, producing a 403 on
the only UI path that can create the first admin. The frontend decides
to show onboarding purely from user_count == 0, so frontend and
backend disagree and the instance bricks.
Change
Split the gate by has_users. Subsequent signups (has_users True) are
unchanged: still gated by ENABLE_SIGNUP and ENABLE_LOGIN_FORM. The
first user (has_users False, the bootstrap admin the onboarding screen
invites) is gated only by the admin-chosen ENABLE_LOGIN_FORM (the
documented SSO-only hard-disable) unless ENABLE_INITIAL_ADMIN_SIGNUP is
set. It is no longer gated by ENABLE_SIGNUP, which in the zero-user
state is never an admin decision but the post-first-admin auto-disable
leaking across a database reset.
Why this is safe (full case analysis)
For WEBUI_AUTH the gate has 16 input combinations over (has_users,
ENABLE_SIGNUP, ENABLE_LOGIN_FORM, ENABLE_INITIAL_ADMIN_SIGNUP). Old and
new are identical in 15 of them:
* All 8 has_users=True cases: both reduce to "403 iff not
ENABLE_SIGNUP or not ENABLE_LOGIN_FORM". Unchanged.
* 7 of the 8 has_users=False cases: identical.
The only changed case is has_users=False, ENABLE_SIGNUP=False,
ENABLE_LOGIN_FORM=True, ENABLE_INITIAL_ADMIN_SIGNUP=False: old
behaviour 403, new behaviour allow. The new condition is a strict
subset of the old (new-403 implies old-403), so the change never newly
blocks any request that previously succeeded; it only stops blocking
that one bootstrap state.
That state has no legitimate deployment. With the login form enabled
and zero users the onboarding form is already served, and the only
operator-configurable way to keep the first signup closed (SSO-only:
ENABLE_LOGIN_FORM=False, optionally with ENABLE_INITIAL_ADMIN_SIGNUP)
is preserved byte for byte. ENABLE_SIGNUP=False with zero users is not
an operator choice, it is the automatic post-first-admin disable, so
the old behaviour there was purely a brick with no recovery path. No
security control is weakened: ENABLE_LOGIN_FORM and
ENABLE_INITIAL_ADMIN_SIGNUP keep their exact meaning, and the
WEBUI_AUTH=False path is untouched.
This is not Redis-specific: it reproduces with Redis disabled through
the Postgres config table alone (clear the user table, keep the config
row).
Verification
Drove the real signup endpoint across a 10-case matrix on freshly
migrated databases, including the full end-to-end first-admin creation
(returns role=admin, row persisted as admin) and the preserved
SSO-only, subsequent-signup and no-auth behaviours. All pass.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The startup banner uses Unicode box-drawing characters. On a stdout that
can't encode them (Windows cp1252, or redirected/headless/pythonw output)
print() raises UnicodeEncodeError and aborts startup. This blocks running
open-webui serve headless on Windows.
Guard the banner print and fall back to a plain ASCII line so startup
always proceeds regardless of the console encoding.
Fixes#24965
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
aiohttp's ClientResponse.json() validates the Content-Type header against
'application/json' by default and raises ContentTypeError for any other
value — including 'application/x-ndjson', which Ollama returns for its
OpenAI-compatible /v1/images/generations endpoint.
Pass content_type=None to skip this check while keeping all other parsing
behaviour unchanged. The fix covers image generation (openai, gemini,
automatic1111 engines) and image editing (openai, gemini engines).
serve_cache_file gated the resolved path with file_path.startswith(os.path.abspath(CACHE_DIR)) without a trailing os.sep, so any path resolving to a sibling whose name starts with the cache-dir basename (e.g. cache_backup, cached_models) passed the prefix check. Authenticated users could read files from such siblings via /cache/../<sibling>/<file>. Appending os.sep to the prefix closes the bypass; deep traversal and absolute paths were already correctly blocked.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The proxy gathered _client_to_upstream and _upstream_to_client with
return_exceptions=True. When upstream sends a graceful CLOSE,
_upstream_to_client returns but gather keeps waiting on
_client_to_upstream, which is blocked in ws.receive() until the browser
disconnects. The handler stays pending and the finally: session.close()
cleanup is deferred, leaking a ClientSession and an open browser socket.
Use asyncio.wait(return_when=FIRST_COMPLETED) and cancel the pending
sibling, so the proxy unwinds as soon as either direction finishes. The
pumps' bare except Exception already lets CancelledError (a BaseException)
propagate, so cancellation is clean and they need no change.
Fixes#25464
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat: cap profile image data URI size to bound model/avatar bloat
validate_profile_image_url() validated data-URI format (MIME allowlist,
SVG rejection, scheme checks) but never its length, so a valid
data:image/...;base64,<huge> passed for both custom-model icons and user
avatars. Large inline images bloat Postgres and the Redis MODELS hash and
degrade model-list latency.
Add PROFILE_IMAGE_MAX_DATA_URI_SIZE (default 256 KiB, 0 disables) and
reject oversized data URIs in the shared validator, so both model meta
(ModelMeta.profile_image_url) and user avatars (UpdateProfileForm) are
bounded at one chokepoint. ModelMeta already clears invalid values to
None on read, so existing oversized icons stop propagating into the
MODELS hash on the next refresh.
Fixes#25468
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix: default PROFILE_IMAGE_MAX_DATA_URI_SIZE to None (no cap)
Per review: opt-in rather than a 256 KiB default. Unset leaves data URIs
uncapped; the validator already skips the check on a falsy value.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
getChatEventEmitter starts a setInterval emitting a `usage` socket event
every second; clearInterval ran only after sendMessageSocket resolved, so
any throw/reject left the interval firing for the page lifetime. Each
failed send added another orphaned interval, inflating server-side usage
accounting and growing CPU/memory over a session.
Wrap the send in try/finally so the interval is always cleared, on both
the happy path and any thrown/rejected path. The exception still
propagates unchanged.
Fixes#25465
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
_sanitize_proxy_path decoded the proxy path once before the '..' check, so a double-encoded payload (%252e%252e) survived the check as %2e%2e and was then re-decoded into '..' by the upstream terminal server, defeating the traversal guard. Decode until stable so no encoded traversal sequence can reach the upstream. Single-encoded payloads were already rejected; this closes the double (and deeper) encoding bypass.
Co-authored-by: sermikr0 <230672901+sermikr0@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add missing custom={true} prop to AdvancedParams in General.svelte and
ModelSettingsModal.svelte so the 'Add Custom Parameter' option appears
consistently across all advanced parameter surfaces.
Also forward custom_params in the General.svelte save handler so custom
parameters are persisted instead of silently dropped on save.
Both presence_penalty and repeat_penalty in the saveHandler read from
params.frequency_penalty instead of their own values due to a
copy-paste error. This causes users adjusting either parameter to
silently save the frequency_penalty value instead.
Thread.svelte: Add null check for messagesContainerElement in scrollToBottom()
to match the existing pattern in Channel.svelte. Prevents potential TypeError
when the DOM element is not yet bound during rapid thread switches.
PinnedMessagesModal.svelte: Move res.length check inside the if (res) block.
Previously, res.length was accessed unconditionally after a guarded block,
causing TypeError when the API call fails and the .catch() returns null.
The SVG-XSS hardening introduced in f5f4b5895 correctly rejects
data:image/svg+xml URIs on new input, but also caused a
pydantic_core.ValidationError when reading pre-existing models from
the database that had SVG data URIs stored as their profile images.
This ValidationError propagated unhandled through _to_model_model and
get_all_models, crashing the entire /api/models endpoint with HTTP 500
and leaving users with no models available in the UI.
Fix:
- Wrap validate_profile_image_url() in a try/except ValueError inside
ModelMeta.check_profile_image_url. Legacy entries are cleared to None
with a warning log instead of raising — the /model/profile/image API
endpoint already falls back to /static/favicon.png when the value is
empty.
- Default ModelMeta.profile_image_url to None instead of hardcoding
/static/favicon.png, since the serving endpoint handles the fallback.
- Add a per-model try/except in ModelsTable.get_all_models so that any
future unexpected validation failure on a single record skips that
model with an error log rather than aborting the entire list.
bypass_system_prompt is an internal flag used by utils/middleware.py and utils/chat.py to skip applying the model system prompt on recursive base-model calls, but it was still declared as a positional argument on the openai/ollama chat-completion route handlers, so FastAPI bound it from the query string. Move it to request.state so external clients cannot set it, matching how bypass_filter is handled.
Drop the argument from both route signatures and read getattr(request.state, 'bypass_system_prompt', False); utils/chat.py sets request.state.bypass_system_prompt alongside bypass_filter and drops the kwarg from the two route-handler calls (the recursive self-calls keep it). Mirrors c0385f60b.
Co-authored-by: anishgirianish <161533316+anishgirianish@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The 't0p-s3cr3t' default was dead code on every supported startup path:
start.sh, start_windows.bat and `open-webui serve` all set or
auto-generate WEBUI_SECRET_KEY before the backend imports env.py. It was
only ever reachable by invoking uvicorn directly, which is unsupported
and unsafe (the app would then sign tokens/cookies with a public,
hardcoded key). It also keeps getting reported as a vulnerability because
it looks dangerous, even though it is unreachable in practice.
Drop the fallback (default to '') so an unset key is caught by the
existing WEBUI_AUTH guard, and replace the vague error with a clear,
actionable message explaining that the key is a hard requirement and how
the supported start methods provide it. Exit cleanly via SystemExit
instead of raising a ValueError traceback.
WEBUI_AUTH=False keeps working unchanged (key defaults to '').
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Long usernames overflow the Edit User modal, User Preview modal header,
and the sidebar user area because the flex containers lack width
constraints.
- EditUserModal: add min-w-0 to the flex-1 container so the existing
truncate class takes effect
- UserPreviewModal: add min-w-0 and truncate to the title container,
flex-shrink-0 to the close button so it stays visible
- Sidebar: add truncate to the username display and flex-shrink-0 to
the avatar container to prevent it from being squeezed
renderMermaidDiagram returned raw mermaid SVG, which FilePreview.svelte injects
via wrapper.innerHTML = svg. Mermaid runs with securityLevel: 'loose', so it
neither sanitizes click hrefs (formatUrl skips sanitizeUrl) nor DOMPurifies its
output; a .md file with a click X href "javascript:..." directive (or an
HTML-label payload) therefore executes script in the app origin when previewed.
The chat path was already safe because SVGPanZoom DOMPurifies before rendering;
file preview was not.
Sanitize at the source: renderMermaidDiagram now returns DOMPurify-cleaned SVG
via a shared sanitizeSvg helper (same policy as SVGPanZoom), so every consumer
including the FilePreview innerHTML sink receives safe output.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
insert_chat_files() stored any caller-supplied file_id with no ownership
check, so a user could attach another user's file to their own chat and
then read it through the shared-chat access path in has_access_to_file().
Filter file_ids to those the caller owns, is admin for, or can read.
Also repairs an UnboundLocalError introduced in 260ead64d: the existing
duplicate-check referenced `session` before it was assigned (db=session),
so the function threw on every call and no chat_file rows were persisted.
ModelMeta.profile_image_url now runs validate_profile_image_url, rejecting SVG/script data URIs (matching UserUpdateForm and ChannelWebhookForm). The /model/profile/image endpoint enforces the PROFILE_IMAGE_ALLOWED_MIME_TYPES allowlist and sets X-Content-Type-Options: nosniff, so an SVG data URI can no longer be served inline on-origin. Closes the fourth profile-image XSS sink missed by the user and webhook fixes.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Starlette reconstructs request.url.path from the HTTP Host header without
validation. An attacker can inject a path into the Host header to make
request.url.path return a different value than the path Starlette routes on.
The API key endpoint restriction check was using request.url.path to decide
whether to allow or deny access — making it bypassable via a crafted Host
header on any Starlette version prior to 1.0.1.
Fix: replace request.url.path with request.scope["path"], which reads the
raw ASGI scope path that Starlette uses for routing. This value is set by
the ASGI server from the actual request path and cannot be injected via
HTTP headers, making it safe regardless of Starlette version.
Affected code path:
get_current_user_by_api_key() in backend/open_webui/utils/auth.py
(only triggered when ENABLE_API_KEYS_ENDPOINT_RESTRICTIONS is enabled)
References:
CVE-2026-48710 / BadHost
https://arstechnica.com/information-technology/2026/05/millions-of-ai-agents-imperiled-by-critical-vulnerability-in-open-source-package/
When called without attached model knowledge and given a caller-supplied knowledge_id, search_knowledge_files passed it straight to Knowledges.search_files_by_id, which does not enforce ownership on knowledge_id. An authenticated user who happened to know a target UUID could enumerate file metadata (filename, file id, KB id, KB name, updated_at) from any knowledge base, bypassing the AccessGrants permission model.
Mirror the same admin/owner/AccessGrants check the attached-KB branch already uses, matching the sibling query_knowledge_files function.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Both _get_access_grants and _to_prompt_model referenced an undefined
local variable 'session' instead of the 'db' parameter passed to each
method. Because these helpers are called outside of any
'async with get_async_db_context()' block, 'session' did not exist in
their scope, causing a NameError on every prompt fetch.
The NameError was silently swallowed by the broad 'except Exception'
clause in get_prompt_by_id, which returned None — causing the frontend
[id]/+page.svelte to immediately redirect back to /workspace/prompts
rather than rendering the prompt editor.
Also adds the missing 'logging' import and module-level 'log' logger,
which was referenced (but never imported) in insert_new_prompt,
update_prompt_version, and delete_prompt_by_id.
The lookup-prompt-by-command endpoint's only frontend wrapper,
getPromptByCommand, is dead: nothing imports or calls it, the path is
referenced nowhere else, and the route handler has no internal caller
(slash-command resolution happens client-side from the loaded prompt
list). Removes the route handler, its section header, and the dead
wrapper.
The shared Prompts.get_prompt_by_command data-layer method is kept: it
is still used by create/update prompt validation (prompts.py:175, 275,
340). PromptAccessResponse and AccessGrants are untouched.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
_process_picture_url validated the initial picture URL with validate_url()
but then aiohttp followed 3xx redirects without re-validating the target,
so a validate_url-passing public URL could 302 to an internal address and
the body was base64-stored in the user's profile_image_url. This is the
sixth call site of the CVE-2026-45401 redirect-bypass cohort; the other
five already pass allow_redirects=AIOHTTP_CLIENT_ALLOW_REDIRECTS. Apply
the same.
validate_url() resolves DNS to check IPs but discards the result; the
HTTP client resolves again independently. Between those two lookups an
attacker can swap the DNS record from a public IP to an internal one
(DNS rebinding).
Push the IP-is-global check into the actual connection layer so the
validated resolution is the one used for the TCP connect:
- aiohttp (_fetch): _SSRFSafeResolver wraps DefaultResolver and rejects
non-global IPs at resolve time (zero TOCTOU window).
- requests (_scrape): _SSRFSafeAdapter mounts custom urllib3 connection
classes whose _new_conn resolves, validates, and connects to the
validated IP in one shot (zero TOCTOU window).
Both paths respect ENABLE_RAG_LOCAL_WEB_FETCH (skip validation when on).
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
has_access_to_file granted access whenever the file was attached to a
shared chat the user could read, ignoring the requested access_type. A
read-only shared-chat recipient therefore satisfied write and delete
checks and could delete or mutate the chat owner's attached file. Gate
the shared-chat branch on read access, matching the channels branch
directly above it.
Co-authored-by: oxsignal <oxsignal@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
SafePlaywrightURLLoader validated only the initially submitted URL and
then let the browser follow HTTP redirects and client-side navigations
without re-checking them, so a public URL could redirect into the
internal network (cloud metadata, RFC1918, loopback). Intercept
document-type requests, re-run validate_url on each, and apply the same
redirect policy as the requests loader (blocked unless
AIOHTTP_CLIENT_ALLOW_REDIRECTS). Sub-resource requests pass through
unchanged so page rendering performance is unaffected.
Co-authored-by: POV9en <POV9en@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
generateFollowUps in src/lib/apis/index.ts is dead: it appears only at
its own definition, nothing imports or calls it, and it targets a
non-existent path (/api/v1/tasks/follow_ups/completions, plural) while
the real route is /tasks/follow_up/completions (singular). Follow-up
suggestions are generated server-side in the chat-completion middleware
and delivered over the chat:message:follow_ups websocket event, so this
wrapper was never on the live path.
Removes only the dead wrapper. The backend POST /tasks/follow_up/completions
endpoint is intentionally kept: it is a member of the actively-used
/tasks/*/completions family (title, tags, emoji, queries, moa) and its
handler delegates to the core generate_follow_ups function.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
update_event only verified write access on the event's source calendar.
CalendarEventUpdateForm accepts a new calendar_id which the model layer
applies unconditionally, so a user with write access to their own calendar
could move (inject) an event into any other user's calendar. Mirror the
destination check create_event already performs.
The __main__ block called search_bing() with 4 positional arguments, but
the function requires 5 (subscription_key, endpoint, locale, query,
count). Running `python -m open_webui.retrieval.web.bing` raised a
TypeError and, before failing, silently misrouted every argument. Read
the key/endpoint from environment variables, matching config.py defaults.
Closes#24765
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The bulk-clear-chat-tags endpoint's only frontend wrapper,
deleteTagsById in src/lib/apis/chats/index.ts, is dead: nothing imports
or calls it, the path is referenced nowhere else, and the route handler
has no internal caller. Removes the route handler, the dead wrapper, and
the now-orphaned Chats.delete_all_tags_by_id_and_user_id model method
(its sole caller was this route). The shared
Chats.delete_orphan_tags_for_user method is untouched.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
POST /utils/markdown rendered a markdown string to HTML server-side. Its
only frontend wrapper, getHTMLFromMarkdown in src/lib/apis/utils/index.ts,
is dead: nothing imports or calls it, the route is hit by no other code
path, and the path string appears nowhere else in the repo (no direct
fetch, no test, no docs). Markdown is rendered client-side in the UI, so
this endpoint was redundant.
Fully self-contained removal: the endpoint, its MarkdownForm model, the
now-orphaned 'import markdown' in the utils router (used only here), and
the dead getHTMLFromMarkdown wrapper. Nothing else depends on any of them.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
GET /evaluations/feedbacks/all returned the entire feedback table in a
single response (flagged as a Medium OOM risk for admins in
open-webui#22206). Its only frontend wrapper, getAllFeedbacks in
src/lib/apis/evaluations/index.ts, is dead: nothing imports or calls it
anywhere in the codebase. The endpoint is a redundant view-only twin of
GET /evaluations/feedbacks/all/export, which is what the admin Feedbacks
UI actually uses.
Removes the endpoint, the now-unused FeedbackResponse import in the
evaluations router, and the dead getAllFeedbacks frontend wrapper. The
shared Feedbacks.get_all_feedbacks data-layer method is kept, since the
live /feedbacks/all/export endpoint still uses it.
Ref: open-webui#22206
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
These 19 exported wrappers are dead: each appears exactly once in the
codebase (its own definition), nothing imports or calls any of them, and
none has a corresponding backend route. They are leftovers from settings
that were consolidated server-side into /auths/admin/config,
/openai/config, /ollama/config and /api/config:
- index.ts: getModelFilterConfig, updateModelFilterConfig,
getCommunitySharingEnabledStatus, toggleCommunitySharingEnabledStatus,
getModelConfig, updateModelConfig (+ orphaned GlobalModelConfig type)
- auths: getSignUpEnabledStatus, toggleSignUpEnabledStatus,
getDefaultUserRole, updateDefaultUserRole, getJWTExpiresDuration,
updateJWTExpiresDuration
- openai: getOpenAIUrls, updateOpenAIUrls, getOpenAIKeys, updateOpenAIKeys
- ollama: updateOllamaUrls
- prompts: restorePromptFromHistory
- folders: updateFolderItemsById (+ orphaned FolderItems type)
Shared types (ModelConfig/ModelMeta/ModelParams) and all live wrappers
are untouched. Removal is import-safe: nothing referenced these.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reported by bwgabrielsusai on #24719: granting a user only
`workspace.skills` doesn't show the Workspace menu, and visiting
`/workspace` directly bounces them to `/`.
The per-route guard in `/workspace/+layout.svelte` already covered
skills, but two earlier gates in the chain didn't:
* `Sidebar.svelte` case 'workspace' OR'd models/knowledge/prompts/tools
to decide menu visibility — skills was missing, so the entry never
rendered for skills-only users.
* `/workspace/+page.svelte` redirect chain picked the first available
section — skills was missing, so the fallback `goto('/')` fired.
Adding skills to both.
start.sh runs with `set -euo pipefail`, but three call sites added in
070ab2650 (refac: reorganize scripts and ci workflows) reference
optional env vars via bash's `,,` lowercase expansion without any
default. Containers that don't set these vars — the default for every
deployment that isn't explicitly opting into Playwright / bundled
Ollama / CUDA — crash on startup with:
start.sh: line 15: WEB_LOADER_ENGINE: unbound variable
(and the same for USE_OLLAMA_DOCKER, USE_CUDA_DOCKER once the first
were set in turn.) Reported in open-webui#24560 by urbenlegend.
The same refactor correctly defaulted every other optional env var
with `${VAR:-…}`. The three `,,` references slipped through because
bash can't combine `:-default` with `,,` in a single substitution —
`${VAR:-default,,}` makes the default literal `,,`, not what's wanted.
Fix: normalise the three vars in a one-line preamble with `${VAR:=}`,
which assigns an empty default if unset. The downstream `${VAR,,}`
expressions stay exactly as Tim wrote them, preserving the file's
visual style and matching the existing `${VAR:-…}` idiom for "this
variable is optional".
* fix: enforce features.direct_tool_servers on chat-completion tool_servers
The features.direct_tool_servers per-user permission was correctly
enforced on the storage path (routers/users.py user/settings/update,
which strips toolServers from saved settings when the caller lacks the
permission), but the inference path (/api/chat/completions) popped
tool_servers straight from the request body into metadata with no
permission check. The middleware (utils/middleware.py:2799) then
consumed direct_tool_servers to inject system_prompt into the message
array and register external tool specs that get invoked during the
completion. End result: any authenticated user could bypass the
admin-set per-user feature toggle and use inline tool_servers in their
chat-completion requests, even when admin had explicitly denied the
permission.
Default for USER_PERMISSIONS_FEATURES_DIRECT_TOOL_SERVERS is False
(config.py:2750), so under default config no regular user is supposed
to be able to use direct tool servers — making this a real boundary
bypass on out-of-the-box deployments rather than a corner case.
Mirror the storage-side behaviour at the inference entry point: pop
tool_servers from the request body, then silently drop the value if
the caller is non-admin and lacks features.direct_tool_servers. Admins
always pass; users with the explicit grant always pass; everyone else
gets None propagated into metadata, which the middleware already
handles as the no-tool-servers case.
Reported by berkant-koc in GHSA-f582-c373-jjf6.
Co-authored-by: berkant-koc <berkant-koc@users.noreply.github.com>
* chore: trim verbose comment on tool_servers permission check
---------
Co-authored-by: berkant-koc <berkant-koc@users.noreply.github.com>
POST /api/v1/chats/new and POST /api/v1/chats/{id}/folder accepted a
caller-supplied folder_id with no validation — neither ownership, nor
existence, nor UUID format. The row was persisted with the supplied
value verbatim, so the DB ended up with chat rows whose folder_id
referenced another user's folder, a non-existent UUID, or even a
non-UUID string.
No read path surfaces this across users — every chat-folder read is
user_id-filtered on both sides — so this is referential-integrity
hardening rather than a security boundary fix. But there's no reason
to accept dangling references either, and the downstream consumers
shouldn't have to assume the column is clean.
Add a Folders.get_folder_by_id_and_user_id() lookup at both writers:
if a folder_id is supplied, it must match a folder owned by the
caller. None remains allowed (chat-without-folder is the default).
Non-existent and non-UUID values fall through to 404.
Reported by ShigekiTsuchiyama in GHSA-4vrg-2vcq-q7jc.
Co-authored-by: ShigekiTsuchiyama <ShigekiTsuchiyama@users.noreply.github.com>
Adds an IFRAME_CSP environment variable that injects a Content-Security-Policy
<meta> tag into all srcdoc iframes rendering untrusted content:
- Artifacts (LLM-generated HTML previews)
- FullHeightIframe (tool/embed output)
- FilePreview (user-uploaded HTML files)
- CitationModal (RAG document HTML)
Shared utility in src/lib/utils/csp.ts handles injection with HTML-safe
attribute escaping. URL-based iframes (src=) are correctly excluded.
Env-var only — no PersistentConfig, no admin UI, no DB. Set once at deploy
time, requires restart. Empty string (default) means no CSP restriction.
* fix: prevent mass-assignment user_id spoofing in POST /api/v1/evaluations/feedback
Two independent gaps in backend/open_webui/models/feedbacks.py let an
authenticated caller forge the `user_id` (and `id`, `version`) on a new
feedback record submitted to POST /api/v1/evaluations/feedback:
1. `FeedbackForm` declared `model_config = ConfigDict(extra='allow')`,
so Pydantic preserved any extra fields supplied in the request body —
including `user_id`, `id`, `version`. The form is the public input
boundary for the endpoint and should not accept unknown fields.
2. In `insert_new_feedback`, the dict literal placed
`**form_data.model_dump()` AFTER `'id': id`, `'user_id': user_id`,
`'version': 0`. Python dict-literal duplicate-key resolution is
last-wins, so any of those fields present in `form_data` overwrote
the server-derived values.
Combined effect: a regular user could POST a feedback record with an
arbitrary `user_id`, attributing the rating to any other user. The Elo
leaderboard at backend/open_webui/routers/evaluations.py computes model
rankings from these records, and the admin export
(GET /api/v1/evaluations/feedbacks/export) and admin list
(GET /api/v1/evaluations/feedbacks/all) display the spoofed attribution.
Two fixes, defense-in-depth:
- FeedbackForm: switch `extra='allow'` to `extra='ignore'` so Pydantic
drops unknown fields at parse time. Sub-models (RatingData / MetaData /
SnapshotData) intentionally keep `extra='allow'` because their contents
are deliberately schema-flexible — the spoofing surface was the form,
not the sub-payloads.
- insert_new_feedback: spread `form_data.model_dump()` first, then
overlay server-controlled fields (`id`, `user_id`, `version`,
`created_at`, `updated_at`) so the explicit keys win on duplicate-key
resolution regardless of what reaches the function. Matches the secure
pattern already used in backend/open_webui/models/functions.py:120.
Reported by yantongggg in GHSA-rjmp-vjf2-qf4g. Same root-cause class as
the prior published GHSA-hr43-rjmr-7wmm (folder mass-assignment, fixed
in v0.9.0); that fix did not generalize across the codebase, this fix
closes the feedback variant.
Co-authored-by: yantongggg <yantongggg@users.noreply.github.com>
* chore: trim comments
---------
Co-authored-by: yantongggg <yantongggg@users.noreply.github.com>
* fix: prevent redirect-based SSRF in get_image_base64_from_url
Cohort follow-up to PR #24491. That PR patched three call sites
(SafeWebBaseLoader._scrape, get_content_from_url, load_url_image) to
pass allow_redirects=False on the underlying HTTP client; this fourth
call site in utils/files.py was missed.
get_image_base64_from_url() is invoked from convert_url_images_to_base64
in utils/middleware.py on every /api/chat/completions request whose
message content includes an image_url part. validate_url() is called on
the originally-submitted URL only; the aiohttp session.get() call had
no allow_redirects argument and the shared session pool does not
override the aiohttp default (allow_redirects=True). An authenticated
user sending a chat message with image_url pointing at an attacker host
that 302-redirects to 169.254.169.254 / 127.0.0.1 / RFC1918 reached the
internal target. This is the most reachable variant in the redirect
cluster: no special endpoint, no admin permission, no feature flag.
Apply the same one-line fix as the other three call sites: pass
allow_redirects=AIOHTTP_CLIENT_ALLOW_REDIRECTS (defaults to False).
Reported by nayakchinmohan in GHSA-88jq-grjp-jx6f; consolidated under
GHSA-rh5x-h6pp-cjj6.
Co-authored-by: nayakchinmohan <nayakchinmohan@users.noreply.github.com>
* fix: enforce collection write access on process_file endpoint
Cohort follow-up to ba83613ff. That commit added _validate_collection_access
to process_text and process_web (the user-supplied collection_name path)
but missed process_file in the same router.
process_file accepts a user-supplied collection_name and writes the file's
embedded content into that collection via save_docs_to_vector_db. The
file_id is gated by file ownership (line 1562) but collection_name was
unchecked, so an authenticated user could append content from a file they
own into another user's knowledge-base collection by passing the victim's
KB UUID as collection_name. Identical pattern to the process_text and
process_web gaps that ba83613ff closed.
Apply the same one-line gate as the sibling endpoints: when
collection_name is user-supplied (not the default file-{file.id} fallback),
require write access via _validate_collection_access. The shared validator
delegates to filter_accessible_collections, which already correctly
handles file-* prefixes (via has_access_to_file) and KB UUIDs
(via Knowledges.check_access_by_user_id) — admins bypass.
Reported by tenbbughunters (Tenable) in GHSA-4g37-7p2c-38r9 (the
comprehensive write-path filing covering process_text / process_file /
process_web / process_youtube and the _validate_collection_access UUID
root cause), and independently re-identified for the missed process_file
call site by kodareef5 in GHSA-4m74-3cmc-293g.
Co-authored-by: tenbbughunters <tenbbughunters@users.noreply.github.com>
Co-authored-by: kodareef5 <kodareef5@users.noreply.github.com>
* fix: enforce collection write access on process_files_batch endpoint
Cohort follow-up to ba83613ff and the prior process_file fix on this
branch. process_files_batch (line 2604) is the third write endpoint in
the same router that accepts a user-supplied collection_name; it was
covered in the same Tenable filing as process_file and was missed by
the same cohort fix. The endpoint validates per-file ownership at line
2642 but does not check whether the caller has write access to the
target collection_name before save_docs_to_vector_db writes into it
at line 2683-2690 with add=True.
Apply the same one-line gate as the sibling endpoints. Validate only
when collection_name is user-supplied (truthy) so the existing fall
through behavior for the None case is unchanged.
Same Tenable / kodareef5 cohort as the previous commit.
Co-authored-by: tenbbughunters <tenbbughunters@users.noreply.github.com>
Co-authored-by: kodareef5 <kodareef5@users.noreply.github.com>
---------
Co-authored-by: nayakchinmohan <nayakchinmohan@users.noreply.github.com>
Co-authored-by: tenbbughunters <tenbbughunters@users.noreply.github.com>
Co-authored-by: kodareef5 <kodareef5@users.noreply.github.com>
* fix: gate tool content updates behind workspace.tools to match create endpoint
`update_tools_by_id` (routers/tools.py:452) authorizes a caller as long as
they are the tool's owner, hold a `write` access grant on the tool, or are
an admin. This means a verified user who has been given a write grant on
a tool — typically as part of a metadata-collaboration workflow (edit
description, adjust valves, manage access grants) — can also overwrite
the tool's Python source. Because `load_tool_module_by_id` further down
calls `exec(content, module.__dict__)` at module-import time, anything
the new content puts outside the `class Tools:` body executes immediately
on the server with the worker's privileges (root in the default Docker
deployment).
The `create_new_tools` endpoint already requires
`workspace.tools` (or `workspace.tools_import`) precisely because creating
a tool means submitting executable code. The update endpoint did not
mirror that check, producing an asymmetric authorization surface in which
a write-grantee with no workspace permission can still reach the same
exec sink as a workspace.tools-trusted creator. SECURITY.md frames
`workspace.tools` as the trust signal an admin uses to delegate
code-execution capability; the previous behavior let that signal be
bypassed by a per-resource share.
Fix: after the existing ownership / write-grant / admin gate, add a
content-change check. If `form_data.content != tools.content`, require
`workspace.tools` or `workspace.tools_import` (or admin role). Metadata
edits — `name`, `description`, valves config, access grants — continue
to flow through the existing gate, so the legitimate share-for-
collaboration workflow is unaffected.
Reported by KadirArslan in GHSA-p4fx-23fq-jfg6 with a working three-user
PoC (Alice trusted with workspace.tools creates a tool and shares write
to Bob; Bob updates content and the new code runs as root inside the
container, with Burp Collaborator confirming outbound exfiltration).
Co-authored-by: KadirArslan <KadirArslan@users.noreply.github.com>
* chore: trim comment
---------
Co-authored-by: KadirArslan <KadirArslan@users.noreply.github.com>
* fix: enforce message ownership in group/DM channel update + delete endpoints
`update_message_by_id` (channels.py:1348) and `delete_message_by_id`
(channels.py:1550) branch on `channel.type`. The `else` branch (standard
channels) correctly enforces `message.user_id != user.id` ownership before
mutating, but the `if channel.type in ['group', 'dm']` branch only checked
`is_user_channel_member` — channel membership alone, with no message
ownership verification.
Effect on group/DM channels: any verified member of the conversation could:
- overwrite another member's message content while the server preserved
`user_id=victim`, producing tampered content that renders to other
members as the original author's authentic post (integrity + authenticity);
- silently delete another member's messages, removing them from
conversation history without trace (integrity).
Reproduced end-to-end against v0.9.4 with three users (attacker, victim,
viewer) sharing a group channel: attacker overwrites victim's message and
deletes another, viewer reads the tampered content as victim-authored.
Two patches, identical shape, mirror the `else` branch's existing
ownership semantics:
- `update_message_by_id` group/DM branch: add
`if user.role != 'admin' and message.user_id != user.id: raise 403`
immediately after the `is_user_channel_member` check.
- `delete_message_by_id` group/DM branch: same.
The standard-channel branch is unchanged (it already enforced ownership).
Admins remain able to moderate any message, matching the existing semantic
in the standard-channel branch.
Reports consolidated under GHSA-wwhq-cx22-f7vv (earliest live filing of the
group/DM-specific variant). Same gap previously surfaced and partially
fixed under GHSA-jxwr-g6r6-j3fx (which addressed the standard-channel
branch only) — this completes the cohort.
* chore: trim comments
`pin_channel_message` (channels.py:1242) checked `permission='read'` on
the standard-channel branch before mutating `is_pinned` / `pinned_by` /
`pinned_at` via `Messages.update_is_pinned_by_id`. Pin/unpin is a write
operation; gating it on read access let any user with read-only channel
access pin or unpin any message in the channel, including admin posts.
One-character fix: change `permission='read'` to `permission='write'`.
Reported by kikayli in GHSA-5gc6-xhv4-2wg6.
Co-authored-by: kikayli <kikayli@users.noreply.github.com>
* chore: add validate_url() to get_image_data() for cohort consistency hardening
`get_image_data()` in `backend/open_webui/routers/images.py` fetches the
URL returned by the configured image generation API directly via
`session.get(data)` without first calling `validate_url()`. The sibling
`load_url_image()` in the same file (called from /images/edit) calls
`validate_url(data)` first — that gate was added under
GHSA-jgx9-jr5x-mvpv. The two functions handle structurally identical
input (an attacker-or-server-supplied URL string) and should enforce the
same SSRF gate as a matter of code hygiene.
In the current call graph, the URL passed to `get_image_data()` comes
from the admin-configured image generation API's response, so an
exploitable SSRF chain additionally requires admin-side trust delegation
(misconfigured/untrusted upstream image API, or a custom
OpenAI-compatible server that reflects user input into response URLs).
That makes the missing call a defense-in-depth gap rather than a
vulnerability per SECURITY.md Rule 9 — a position the GHSA-h7cc-wwjp-5xqh
advisory is being closed under. This change is hardening: it brings the
two image-fetch helpers into alignment so any future caller that begins
passing user-influenced URLs into `get_image_data()` is gated by the
same private-IP / loopback / metadata-IP filter the rest of the
codebase enforces.
Surface raised by brodmart in GHSA-h7cc-wwjp-5xqh.
Co-authored-by: brodmart <brodmart@users.noreply.github.com>
* chore: trim comment
---------
Co-authored-by: brodmart <brodmart@users.noreply.github.com>
GET /api/v1/models/model?id=<model_id> at routers/models.py:412
returned the full model.model_dump() to any caller with read access,
including the params dict that holds the admin-curated system prompt
and other behavior config. The user-facing /api/models endpoint
already strips this via utils/models.py:170,210 with the comment
"Remove params to avoid exposing sensitive info", and /api/v1/models/list
gates by write permission so non-curators don't see the model in their
workspace listing at all. The per-id endpoint missed the same gate, so
a user with read-only access (e.g. granted access to use the model in
chat) could open /workspace/models/edit?id=<not-mine> in the browser
and read the system prompt verbatim from the network response, even
though saving was correctly blocked.
Compute write_access once at the top of the handler so it can serve
both the response-shape decision and the response field. When the
caller lacks write access, replace params with an empty dict in the
serialised response. Owners, admins under BYPASS_ADMIN_ACCESS_CONTROL,
and explicit write-grant holders still get the full payload so the
workspace edit UI keeps working for users who legitimately curate the
model.
Read-permission users continue to receive everything else they need to
chat with the model — the chat path resolves prompt/params server-side
from the stored ModelModel and never echoes them back through this
endpoint.
Reported by destination-one in GHSA-h2cw-7qw9-56xr.
Co-authored-by: destination-one <destination-one@users.noreply.github.com>
urllib.parse.urlparse and requests/aiohttp disagree on how to split URLs
containing backslash, tab, CR, or LF in or around the netloc. urlparse
treats backslash as part of userinfo and uses what follows '@' as the
host; requests treats backslash as the start of the path and connects
to whatever precedes it. The same URL therefore passes the private-IP
filter (urlparse sees a public host) but reaches an internal target
(requests connects to e.g. 127.0.0.1). End result is an SSRF that the
existing IP block list cannot catch because it's evaluating the wrong
host.
PoC: http://127.0.0.1:6666\@1.1.1.1 — urlparse hostname is 1.1.1.1
(global, passes), requests reaches 127.0.0.1 (loopback).
Reject up front any URL containing one of the four documented parser-
confusing characters before either parser gets a chance to interpret
it. None of these characters is valid in an unencoded URL (\ should
always be %5C, whitespace should be %09 / %0A / %0D), so this is a
pure defensive rejection with no legitimate-input false positives.
Reported by Fushuling and RacerZ-fighting in GHSA-8w7q-q5jp-jvgx.
Co-authored-by: Fushuling <Fushuling@users.noreply.github.com>
Co-authored-by: RacerZ-fighting <RacerZ-fighting@users.noreply.github.com>
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
The `get_status()` handler at retrieval.py:263 (`@router.get('/')`) returned
the live RAG pipeline configuration (CHUNK_SIZE, CHUNK_OVERLAP, RAG_TEMPLATE,
RAG_EMBEDDING_ENGINE, RAG_EMBEDDING_MODEL, RAG_RERANKING_MODEL, etc.) without
any authentication dependency, while every adjacent endpoint on the same
router (/embedding, /embedding/update, /config, /config/update) requires
get_admin_user.
Exhaustive search of the repository confirms the endpoint has no callers:
- Frontend (src/): no `RETRIEVAL_API_BASE_URL}/'`-style fetch; the existing
`getRAGConfig()` in src/lib/apis/retrieval/index.ts targets `/config`,
not the root, and is the only consumer of admin-level retrieval state.
- Backend self-references: none.
- Cypress e2e (chat, documents, registration, settings): none.
- Backend tests (backend/open_webui/test/): none.
- Build/CI scripts (scripts/): none.
- Direct symbol import of `get_status` from this router: none.
The endpoint is dead code, almost certainly a relic from before the
/config GET split. Removing it has zero UX impact and eliminates the
unauthenticated-config-disclosure surface raised in advisory triage on
GHSA-65pg-qhhw-mxwg. External monitoring scripts that may have hit the
bare root will receive a 404 and can switch to the existing /config
endpoint, which returns the same fields plus the rest of the RAG config
under admin auth.
Surface raised by 0xRyuzak1 in GHSA-65pg-qhhw-mxwg. The advisory was closed
as not-a-vulnerability per SECURITY.md Rule 1 (no security boundary
crossed in default config — RAG_TEMPLATE default is a citation-format
instruction, not a system prompt; no integrity/availability impact); this
removal is independent code-hygiene that aligns the router cohort.
Reported-by: 0xRyuzak1 <https://github.com/0xRyuzak1>
The /create (L155-193) and /id/{id}/update (L248-297) endpoints in
routers/skills.py persisted form_data.access_grants directly to
AccessGrants.set_access_grants without filter_allowed_access_grants,
while every other shareable resource in the codebase (channels, knowledge,
models, notes, prompts, tools, calendars) and the dedicated
/id/{id}/access/update endpoint on this same router (L309-348) all do call
the filter. A user with workspace.skills permission (default False, but
admins can grant it to skill-creating users) could therefore attach
{"principal_type":"user","principal_id":"*","permission":"read"|"write"}
to the create or update payload and have it persisted unfiltered, bypassing
the sharing.public_skills gate that the rest of the cohort enforces.
Two changes:
- create_new_skill: call filter_allowed_access_grants with
'sharing.public_skills' immediately before insert, after the existing
permission check and ID-taken check.
- update_skill_by_id: call filter_allowed_access_grants with the same key
after the access check, before form_data.model_dump() flows into
Skills.update_skill_by_id. The pre-existing access check at L263-277 only
restricts WHO may modify the skill; the new filter restricts WHICH grants
they may set.
All supporting plumbing was already in place from prior PRs:
filter_allowed_access_grants is already imported at L22, the
USER_PERMISSIONS_WORKSPACE_SKILLS_ALLOW_PUBLIC_SHARING constant exists,
DEFAULT_USER_PERMISSIONS['sharing']['public_skills'] is wired up,
SharingPermissions.public_skills is in the Pydantic, and the admin UI
already renders the toggle. This is a pure 2-line router fix that closes
the cohort-consistency gap.
Same shape as the calendar fix in #24493, reported by Matteo Panzeri while
auditing the resource-cohort cohort during follow-up on #24493.
Co-authored-by: Matteo Panzeri <28739806+matte1782@users.noreply.github.com>
* fix: gate public sharing of calendars behind sharing.public_calendars permission
The calendar router did not call filter_allowed_access_grants on either the
create or update endpoint, while every other shareable resource in the
codebase (channels, knowledge, models, notes, prompts, skills, tools) does.
A verified non-admin owner could therefore attach
`{"principal_type":"user","principal_id":"*","permission":"read"|"write"}`
to their own calendar in the create or update payload and have it persisted
unfiltered. Any other verified user with the (default-on) features.calendar
permission could then read or, for write grants, write events on it via the
existing /events* endpoints, bypassing the per-user sharing.public_<X>
permission gate the rest of the resource cohort enforces.
Three changes:
- config.py: add USER_PERMISSIONS_CALENDAR_ALLOW_PUBLIC_SHARING (default
False, env-overridable) and surface it in DEFAULT_USER_PERMISSIONS
['sharing']['public_calendars'] so admins can grant it per group via the
same UI used for notes/models/etc.
- routers/calendar.py: import filter_allowed_access_grants and call it in
create_calendar with the new sharing.public_calendars key, identical to
the channel router's pattern.
- routers/calendar.py: call filter_allowed_access_grants in update_calendar
too. The pre-existing owner-only gate at L350 only restricts WHO may
change grants; the new filter restricts WHICH grants they may set, so a
non-admin owner cannot make their own calendar publicly readable or
writable without the corresponding sharing permission.
Same shape as GHSA-7rjh-px4v-5w55 (channels). Reported by Matteo Panzeri.
Co-authored-by: Matteo Panzeri <28739806+matte1782@users.noreply.github.com>
* fix: expose public_calendars + features.calendar through admin permissions surface
The earlier commit added DEFAULT_USER_PERMISSIONS['sharing']['public_calendars']
and the runtime filter call, but the new key was not yet plumbed through the
admin /users/default/permissions endpoint. Without these changes the toggle
would round-trip as silently dropped:
- routers/users.py SharingPermissions: any payload POSTed to
/default/permissions ran through `form_data.model_dump()`, and Pydantic
drops fields not declared on the model. The new public_calendars key
would have been stripped on every save, leaving admins unable to grant
the permission via the UI even though the runtime filter would honor it.
- src/lib/constants/permissions.ts: the frontend's DEFAULT_PERMISSIONS dict
is the seed shape used by the admin Groups Permissions panel; without
the new key it could not bind a Switch component to it.
- Permissions.svelte: add a Calendars Public Sharing toggle alongside the
Notes/Chats Public Sharing toggles, gated on the existing
features.calendar flag (matches the pattern used for notes/chats).
Also closes a pre-existing parity gap on features.calendar: DEFAULT_USER_
PERMISSIONS['features']['calendar'] has existed since the calendar feature
shipped, and Permissions.svelte already renders a Calendar feature toggle,
but FeaturesPermissions Pydantic and the frontend defaults never knew
about it. Adding it everywhere completes the round-trip so admin saves no
longer silently drop the calendar feature flag either.
---------
Co-authored-by: Matteo Panzeri <28739806+matte1782@users.noreply.github.com>
validate_url() in retrieval/web/utils.py only validates the initial URL.
The HTTP clients used downstream (sync requests, sync requests via the
parent WebBaseLoader._scrape, aiohttp via load_url_image) followed 3xx
redirects by default and did not re-validate the redirect target against
the private-IP / metadata-IP block list. An authenticated user could
submit a public URL that 302-redirected to an internal address (RFC1918,
127.0.0.1, 169.254.169.254, etc.) and the redirected response was returned
to them, enabling SSRF reads of internal services and cloud metadata.
Three call sites needed allow_redirects=False to match the policy already
enforced on the async _fetch() path:
- SafeWebBaseLoader: override requests_kwargs in __init__ so that the
inherited synchronous _scrape() path passes allow_redirects=False to
self.session.get() (the parent WebBaseLoader uses requests' default
allow_redirects=True).
- get_content_from_url (retrieval/utils.py): pass allow_redirects=False
on the streamed requests.get(...) call.
- load_url_image (routers/images.py, image-edits endpoint): pass
allow_redirects=False on the aiohttp session.get(...) call.
Reports consolidated under GHSA-rh5x-h6pp-cjj6:
- GHSA-rh5x-h6pp-cjj6 (tenbbughunters / Tenable) - sync _scrape
- GHSA-5vxg-6gmv-m2qr (YLChen-007) - load_url_image
- GHSA-hf76-c83f-63w2 (tempcollab) - aiohttp _fetch (already fixed)
- GHSA-h55f-h5fh-mvm4 (sneaXOR) - get_content_from_url
* refac(routers): reject external URLs in profile/model image handlers
* refac(ui): centralize image URL validation in safeImageUrl helper
* refac(auths): make signout POST-only
* refac: gate external profile image redirect behind ENABLE_PROFILE_IMAGE_URL_FORWARDING
Restore the 302 redirect for external http(s) profile image URLs in
the user and model profile-image endpoints, but gate it behind a new
ENABLE_PROFILE_IMAGE_URL_FORWARDING env flag (default: True).
Existing deployments that rely on external profile image forwarding
continue to work unchanged. Operators who want to suppress the
redirect (to prevent client-side IP/UA/Referer leaks) can set the
flag to False.
Non-admin GET /api/v1/prompts/tags went through get_prompts_by_user_id,
which loaded every active prompt with its full content/data/meta plus
owner records and all access grants, then ran one has_access query per
prompt that wasn't owned by the caller - all so the endpoint could
collapse the result to a sorted tag list. With 600 prompts this took
several seconds while the admin path (a single SELECT) returned in <1s.
Add Prompts.get_tags_by_user_id which selects only the tags column and
applies the same EXISTS-based access filter used by /list. Also tighten
the admin get_tags to project just the tags column instead of full rows.
The endpoint is now one DB query (plus one for groups), no row hydration,
no N+1.
Co-authored-by: Claude <noreply@anthropic.com>
get_prompts_by_user_id used to fetch every active prompt (with users +
all access grants), then call AccessGrants.has_access() once per prompt
that the user did not own. With 600+ prompts this issued ~600 extra
round-trips per request and explained the multi-second delay reported in
the GET /api/v1/prompts and /api/v1/prompts/tags endpoints for non-admin
users.
Push the access check into a single SQL query via the existing
AccessGrants.has_permission_filter (EXISTS subquery), so only accessible
rows come back from the DB. Users and access grants for the surviving
rows are still batch-fetched, no N+1 anywhere on this path.
Co-authored-by: Claude <noreply@anthropic.com>
Convert the /chats/all endpoint from loading all user chats into memory
at once to a streaming NDJSON response that fetches chats in batches of
100. This prevents Out-of-Memory crashes for users with large chat
histories.
Backend: Added async generator that paginates through chats with
short-lived DB sessions per batch (critical for SQLite lock release).
Frontend: Updated getAllChats to consume the NDJSON stream via
ReadableStream reader, accumulating results for the export file.
Ref: open-webui#22206
Previously, clicking the file name link did not open the file content
because the condition checked `!isPDF && item.url`, which failed for
`type === 'file'` items that use an ID-based URL path.
Update the condition to trigger on `item.type === 'file' || item.url`,
and resolve the correct URL by extracting `fileId` from `item.id` or
`item.tempId` instead of using `item.url` directly as the file
identifier.
The transcription endpoint was async but called the synchronous transcribe() function directly, blocking the single-threaded uvicorn event loop for the entire duration of inference. This caused all HTTP and WebSocket connections to stall for every user on the instance during STT processing.
- Add asyncio import
- Use async UploadFile.read() instead of synchronous file.file.read()
- Offload the blocking transcribe() call via asyncio.to_thread()
Closes#24169
The chat table has no computed columns (no DEFAULT, SERIAL/IDENTITY,
or TRIGGER that populate server-side values on UPDATE), and every
column modified by update_chat_by_id is set explicitly from Python
values earlier in the function. db.refresh therefore issues a SELECT
that replaces those just-written Python values with the round-tripped
database representation of the same values, which is a no-op for
functional purposes but pulls the entire chat.chat JSON blob back over
the network and through the driver's JSON decoder.
On large, active chats where chat.chat can reach tens of megabytes,
skipping the refresh measurably reduces latency and eliminates one
~JSON-sized transient allocation per write.
- Return 302 to /static/favicon.png instead of streaming the same PNG per
model id so browsers can cache one asset for default avatars.
- Validate stored /static/ paths with decode, normpath, and /static
prefix checks; invalid paths fall back to favicon.
Made-with: Cursor
After migration to async db operations, the throttle decorator also
needs to support async. Since the decorator is only used for async funcs
now, we can just switch it to async instead of supporting sync and async
at the same time.
Signed-off-by: Adam Tao <tcx4c70@gmail.com>
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
* perf(channels): batch user lookup in model_response_handler thread history
The thread-history builder in model_response_handler called
Users.get_user_by_id once per thread message (deduped via an intra-loop
dict), producing N individual SELECTs for a thread of N unique authors.
Replace with a single Users.get_users_by_user_ids call that returns all
authors in one WHERE id IN (...) query, matching the batch pattern
already used elsewhere in this file (lines 739, 804, 1320).
Behavior is preserved: deleted users still resolve to None and fall
through to the existing 'Unknown' fallback via .get().
* refac(channels): rename loop vars to full words per review
Address reviewer feedback to use descriptive names `message` and `user`
instead of single-letter `m` and `u` in the batch user-lookup
comprehensions.
---------
Co-authored-by: Claude <noreply@anthropic.com>
After set_access_grants, the handler was reloading the same knowledge
record via get_knowledge_by_id, which triggers an extra SELECT plus a
nested fetch of access grants. set_access_grants already returns the
newly-written grants and the local knowledge object is otherwise
unchanged, so update it in place and reuse it for the response.
https://claude.ai/code/session_01S18Lgqbih7Ry2JZUUv8TxF
Co-authored-by: Claude <noreply@anthropic.com>
Avoid loading the full Chat row (including the potentially large `chat`
JSON column) just to read tag IDs from `meta.tags`. Issue a narrow
SELECT on `Chat.meta` instead, which is much cheaper for chats with
large message histories.
Co-authored-by: Claude <noreply@anthropic.com>
Convert chat_message_file_ids from list to set so the membership test
in the comprehension is O(1) instead of O(m), turning the dedupe from
O(n*m) into O(n+m). Also replace the redundant set([...]) with a set
comprehension.
https://claude.ai/code/session_01Le3NnqNhcgaJvFrDZGqmwe
Co-authored-by: Claude <noreply@anthropic.com>
Pass the request-scoped AsyncSession into Models.get_model_by_id so the
endpoint no longer opens a fresh DB session on every call, avoiding an
extra connection acquisition per profile image request.
Co-authored-by: Claude <noreply@anthropic.com>
* perf(users): drop redundant get_user_by_id refetch in session-user endpoints
Five /user/* handlers refetched the user row via Users.get_user_by_id(user.id)
immediately after receiving an identical UserModel from Depends(get_verified_user).
Since get_verified_user already populated the user within the same request
microseconds earlier, the refetch is pure overhead. The dead else branches
(unreachable — get_verified_user raises 401 on missing user) are removed as
a natural consequence.
Affected endpoints:
- GET /user/settings
- GET /user/status
- POST /user/status/update
- GET /user/info
- POST /user/info/update
Eliminates one SELECT per request to each of these endpoints with no behavioral
change.
* fix(users): preserve USER_NOT_FOUND error on status update failure
update_user_status_by_id returns None when the target user is missing or
the update raises. The previous commit removed the pre-update existence
gate (get_user_by_id) and returned the update result directly, which
turned not-found/failure cases into 200 OK with a null body instead of
the expected 400 USER_NOT_FOUND.
Guard the update result explicitly to preserve the original API contract,
matching the equivalent pattern already applied in /user/info/update.
* docs(users): note lost-update tradeoff on /user/info/update
Make the concurrency tradeoff explicit: merging against the auth-time
snapshot slightly widens the lost-update window compared to the previous
pre-merge refetch, but the refetch only narrowed (did not eliminate) that
window. Real safety requires row locking or a version column.
---------
Co-authored-by: Claude <noreply@anthropic.com>
Add configurable reranker batch size (env var RAG_RERANKING_BATCH_SIZE,
default 32) following the same pattern as RAG_EMBEDDING_BATCH_SIZE.
- config.py: PersistentConfig for RAG_RERANKING_BATCH_SIZE
- main.py: import, state init, pass to get_reranking_function
- colbert.py: accept batch_size param in predict() (was hardcoded 32)
- utils.py: get_reranking_function passes batch_size at call time
- retrieval.py: expose in config GET/POST endpoints and ConfigForm
- Documents.svelte: add Reranking Batch Size input in admin settings
Closes#23730
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
Loader.load() dispatches to the underlying langchain document loaders
(PyMuPDF, Unstructured, python-docx, Tika, …) which are all
synchronous and CPU/IO-bound. process_file() awaited it directly on
the event loop, so parsing a non-trivial PDF/DOCX would freeze the
entire FastAPI app for the duration of the parse — which is what users
experience as "the server hangs whenever I upload a file."
Add an `aload()` async wrapper on Loader that runs the sync load on a
worker thread via asyncio.to_thread, and update process_file() to
await it. The sync API is preserved so existing callers that already
run inside run_in_threadpool (e.g. save_docs_to_vector_db) are
unaffected.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
Co-authored-by: Claude <noreply@anthropic.com>
* fix(retrieval): offload sync VECTOR_DB_CLIENT calls in async paths via AsyncVectorDBClient
The vector DB backends (Chroma, pgvector, Qdrant, Milvus, Pinecone,
Weaviate, …) are uniformly synchronous and their methods perform
blocking network or disk I/O. Multiple async route handlers and helpers
were calling them directly on the event loop — file processing,
memories, knowledge bases, hybrid search bookkeeping — so a single
upsert/delete/search would freeze every other in-flight request for the
duration of the call.
Introduce `AsyncVectorDBClient`, a thin async facade that wraps the
existing sync client and dispatches each method through
`asyncio.to_thread`. It mirrors `VectorDBBase` exactly and forwards
*args/**kwargs so backend-specific extra parameters keep working.
Update every async-context call site (routers/retrieval, routers/files,
routers/memories, routers/knowledge, retrieval/utils,
tools/builtin) to await `ASYNC_VECTOR_DB_CLIENT` instead of calling the
sync client directly. Two helpers that were sync-only also acquire
async siblings or are awaited via `asyncio.to_thread` at their async
call site (`remove_knowledge_base_metadata_embedding`,
`get_all_items_from_collections`, `query_doc`).
The original sync `VECTOR_DB_CLIENT` is unchanged, so callers that
already run inside `run_in_threadpool` (e.g. `save_docs_to_vector_db`
and the sync `query_doc`/`get_doc` helpers) are unaffected.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
* fix(retrieval): restore explicit AsyncVectorDBClient signatures matching VectorDBBase
Per PR review: the original *args/**kwargs forwarding lost type
safety and IDE/static-analysis support. Restore explicit signatures
that mirror VectorDBBase exactly, so:
* Bad kwargs fail at the facade boundary instead of inside the
worker thread (where the resulting TypeError tends to be
swallowed by surrounding `try/except`).
* IDE autocomplete and static analysis work as expected.
* The stated intent ("mirror VectorDBBase exactly") now holds at
the API contract level, not just behaviourally.
While doing this, surface a pre-existing bug in
`delete_entries_from_collection` that the stricter typing flagged:
the call passed `metadata={'hash': hash}` which is not a parameter
on `VectorDBBase.delete` nor any backend. The TypeError raised
inside the sync delete was silently swallowed by `except Exception`
so the endpoint always reported `{'status': False}` for every
request instead of actually deleting matching vectors. Replace with
`filter=...` to do what the endpoint name promises.
The thorough review's other note (no concurrency/backpressure on
the shared default threadpool) is intentionally not addressed here:
asyncio.to_thread on the shared executor is the right primitive for
this use case; per-domain bounded executors would add lifecycle
complexity disproportionate to the problem and the loop is no
longer blocked, which was the actual bug.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
* fix(retrieval): parallelize hybrid-search collection prefetch; document async facade contracts
Address PR review findings:
1. Hybrid-search prefetch was sequential
`query_collection_with_hybrid_search` previously awaited
`ASYNC_VECTOR_DB_CLIENT.get(name)` once per collection in a for
loop. Each call already off-loaded to a worker thread, but
awaiting them serially meant total prefetch latency scaled
linearly with the number of collections. Run them concurrently
with `asyncio.gather` so multi-collection queries actually
benefit from the threadpool. Per-collection exception handling
is preserved by wrapping each fetch in a small helper that
logs and returns `(name, None)` on failure, so a single bad
collection cannot poison the whole gather.
2. Document the thread-safety expectation explicitly
The facade now formally states what was always implicit: the
sync `VECTOR_DB_CLIENT` is shared across worker threads, so the
underlying backend driver must be thread-safe. This is not a
new exposure — `save_docs_to_vector_db` already called the sync
client from `run_in_threadpool`. Adding a global lock here
would defeat the responsiveness the facade exists to provide;
backends that cannot tolerate concurrent access should grow
their own internal serialization.
3. Document the API-surface choice and `.sync` escape hatch
The strict `VectorDBBase` mirror was a deliberate choice (the
previous `*args/**kwargs` revision let a `metadata=` typo
silently break an endpoint). Document it, and call out the
`.sync` escape hatch with an example for callers that genuinely
need a backend-specific parameter not on `VectorDBBase`.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
* fix(retrieval): guard /delete against null file.hash and let HTTPException reach the client
Address PR review finding on the `metadata=` → `filter=` change in
`delete_entries_from_collection`.
The new `filter={'hash': hash}` query was correct for files that
have a hash, but did not handle `file.hash is None` (unprocessed,
failed, or legacy records). The match semantics of a null filter
value are backend-dependent — some ignore the key entirely, some
treat it as "metadata field absent" and match every such row — so
issuing the query risked deleting unrelated entries.
* Reject `hash is None` up front with a 400 explaining the file
has no hash to target.
* Narrow the surrounding `except Exception` so it no longer
swallows `HTTPException`. Without this fix the new 400 (and the
pre-existing 404 for missing files) would be silently re-shaped
into `{'status': False}` and the caller could not distinguish a
bad-request input from a backend error.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
---------
Co-authored-by: Claude <noreply@anthropic.com>
* fix(middleware): replace BaseHTTPMiddleware HTTP middlewares with pure ASGI implementations
Starlette's BaseHTTPMiddleware (and the @app.middleware('http')
decorator that uses it) wraps the downstream app in an anyio task
group whose cancel scope tears down the inner task on every exit —
client disconnect, response complete, or any outer middleware bailing.
That CancelledError gets injected into whatever the inner task was
awaiting, so DB queries, embedding calls, and other long awaits get
killed mid-flight. Under aiosqlite the cleanup path then logs a
multi-page `terminate_force_close() not implemented` traceback at
ERROR for every cancelled DB call.
Open WebUI had four such middlewares stacked
(`commit_session_after_request`, `check_url`, `inspect_websocket`,
`RedirectMiddleware`) so a single cancellation would compound through
all four.
Move the four middlewares to a new `open_webui.utils.asgi_middleware`
module as plain ASGI classes (`__call__(scope, receive, send)`):
* `CommitSessionMiddleware` — was `commit_session_after_request`;
now also rolls back if commit fails
before releasing the connection.
* `AuthTokenMiddleware` — was `check_url`; sets request.state
token + enable_api_keys + stamps
X-Process-Time via a wrapped send.
* `WebsocketUpgradeGuardMiddleware`
— was `inspect_websocket`; rejects
/ws/socket.io HTTP requests that
claim transport=websocket without a
proper Upgrade/Connection header.
* `RedirectMiddleware` — was the BaseHTTPMiddleware subclass;
same /watch + share-target rewrites.
Pure ASGI does not introduce a cancel scope around the downstream app,
so client disconnects propagate via `receive()` (the way ASGI was
designed) instead of being injected as CancelledError. Middleware
ordering is preserved.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
* fix(middleware): CommitSessionMiddleware — rollback on downstream error, never commit failed requests
The first cut put commit() in a finally block, which meant that even
when a downstream handler raised, the middleware would still commit
whatever partial sync writes that handler had made before the
failure. That regressed the previous BaseHTTPMiddleware semantics
where commit only ran on the success path.
Restructure the failure handling:
* Downstream raised → rollback any pending sync work, release the
connection, re-raise so the outer error middleware turns it into
an error response. We never commit a request that did not complete.
* Downstream returned → commit. On commit failure, log loudly,
rollback, and re-raise. ScopedSession.remove() always runs in
finally so the connection cannot leak.
Document the inherent pure-ASGI limitation explicitly: by the time
`await self.app(...)` returns the response messages have already
been emitted, so a commit failure can no longer change what the
client sees on the wire. Buffering the response to gate it on commit
success would break streaming responses (chat completions, SSE) which
are core to Open WebUI; the trade-off is intentional. Routes that
need commit-before-send must manage the sync session explicitly.
Also drop unused `typing` imports flagged by review.
https://claude.ai/code/session_01JSr4NZSskEUQvoJnavVXh8
---------
Co-authored-by: Claude <noreply@anthropic.com>
* fix: drop extra='allow' on FolderForm and FolderUpdateForm
These request models were configured to accept arbitrary extra fields,
which were then merged into the folder row via form_data.model_dump().
In insert_new_folder the server-assigned user_id is placed before the
form spread, so a client-supplied user_id in the request body would
override it and the folder would be persisted against another account.
Strictly typed inputs are the correct shape for these endpoints — the
client has no legitimate reason to send fields beyond the declared
ones, and dropping extra='allow' closes the mass-assignment sink at
the validation layer instead of relying on every callsite to merge
fields in the right order.
* fix: reject unknown fields on FolderForm and FolderUpdateForm
Address review feedback: dropping extra='allow' fell back to Pydantic
v2's default extra='ignore', which only silently drops unknown fields
instead of rejecting them. The intent for these request models is a
strict input contract — fail fast when a client sends anything the
server does not expect — so explicitly set extra='forbid'. This also
makes the hardening visible in the form definition rather than implicit
in the default.
These four Redis calls were the only places in the backend still using
bare cache keys. Every other call already namespaces through
REDIS_KEY_PREFIX, which is how multiple Open WebUI instances sharing a
Redis database stay isolated. Two instances with different
TOOL_SERVER_CONNECTIONS or TERMINAL_SERVER_CONNECTIONS would otherwise
clobber each other's cached OpenAPI specs.
* fix: replace brittle profile_image_url allowlist with safe-scheme validation
The previous validation used a hardcoded allowlist of specific static
paths and a single Gravatar prefix. This rejected OWUI's own internal
API paths (e.g. /api/v1/users/{id}/profile/image) and external OAuth
avatar URLs, making it impossible to save user profiles from the admin
panel.
Replace with scheme-based validation that allows relative paths,
HTTP(S) URLs, and data:image URIs while blocking dangerous schemes
like javascript:, file:, and ftp:.
Fixes open-webui#23387
* fix: harden profile image URL validation per review feedback
- Restrict data URIs to safe raster formats (png/jpeg/gif/webp);
SVG is excluded because it can carry embedded scripts.
- Block scheme-relative URLs (//host/path) which browsers resolve
against the current protocol, bypassing the relative-path check.
* fix: use structural validation instead of prefix checks
- Use urlparse for HTTP(S) URLs: gives case-insensitive scheme
matching and rejects bare schemes with no host (e.g. https://).
- Use a compiled regex for data URIs: enforces the ;base64, boundary,
restricts to safe raster formats, and is case-insensitive per spec.
- Removes the startswith-based prefix tuple in favour of proper
URL and data URI parsing.
* fix: validate hostname not netloc, fix misleading comment
- Use parsed.hostname instead of parsed.netloc so URLs like
http://:80/path (non-empty netloc but no actual host) are rejected.
- Update data URI comment to accurately state we validate MIME type
and structure, not base64 payload integrity.
* fix: constrain relative paths to known-safe prefixes
Accepting any relative path starting with / allowed a user to set
their profile_image_url to an arbitrary internal GET endpoint. When
another user (e.g. an admin) views that profile, the browser fires
the GET with the viewer's session cookies — an authenticated GET
trigger surface.
Constrain to known-safe prefixes (/api/v1/users/, /static/) and
exact matches (/user.png, /favicon.png) which are the only relative
paths OWUI itself generates.
* fix: use exact matches and anchored regex, eliminate all prefix wildcarding
Replace all startswith-based path checks with:
- frozenset exact matches for static assets (/user.png, /favicon.png,
/static/favicon.png)
- Anchored regex for the OWUI profile image API route that accepts
only /api/v1/users/{id}/profile/image (no trailing components,
no path traversal across segments)
This eliminates every prefix-based attack surface:
- /api/v1/users/{id}/anything-else is rejected
- /static/../../etc/passwd is rejected
- /api/v1/users/../../admin/config is rejected
- Arbitrary internal GET triggers are no longer possible
* fix: exclude query/fragment delimiters from user-ID regex segment
Change [^/]+ to [^/?#]+ so that inputs like
/api/v1/users/alice?x=1/profile/image are rejected — the browser
would interpret ? as the query string start, making the actual
request target /api/v1/users/alice instead of the intended route.
* Add ownership checks to global task endpoints
- Restrict GET /api/tasks and POST /api/tasks/stop/{task_id} to admin-only
- Add new scoped POST /api/tasks/chat/{chat_id}/stop endpoint with ownership
check so regular users can stop their own chat tasks
- Allow admins to access the scoped chat task endpoints alongside owners
- Update frontend to use the new scoped stop endpoint when a chatId is available
https://claude.ai/code/session_01K7zPDvvjRu8AxJ4Br2HhZc
* Handle temporary (local:) chat IDs in scoped task endpoints
Temporary chats use local:<socketId> as chat_id which doesn't exist in
the DB. The scoped endpoints now skip ownership checks for local: IDs
(they aren't enumerable) and use {chat_id:path} to handle the colon in
the URL path.
https://claude.ai/code/session_01K7zPDvvjRu8AxJ4Br2HhZc
* Verify session ownership for local: chat IDs and URL-encode chat_id
- For local:<socketId> chat IDs, look up the socket's owner in
SESSION_POOL and verify it matches the requesting user (or admin)
- URL-encode chat_id in frontend fetch calls to handle special
characters (colon in local: IDs) safely
https://claude.ai/code/session_01K7zPDvvjRu8AxJ4Br2HhZc
---------
Co-authored-by: Claude <noreply@anthropic.com>
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
The validators.ipv6(ip, private=True) call always returns a falsy ValidationError because validators==0.35.0 does not support the private kwarg for IPv6. This means any hostname resolving to a private IPv6 address (::1, fd00::*, ::ffff:169.254.169.254) bypasses SSRF protection entirely, circumventing the fix for CVE-2025-65958.
Replace both the IPv4 and IPv6 validators-based private checks with Python's stdlib ipaddress module using an allowlist approach (not addr.is_global). This blocks all non-globally-routable addresses — private, loopback, link-local, reserved, multicast, and unspecified — for both IPv4 and IPv6, including IPv4-mapped IPv6 addresses.
Per RFC 4513, a Simple Bind with a non-empty DN but empty password is unauthenticated simple authentication. Many LDAP servers (OpenLDAP default, some AD configs) accept these binds, allowing account takeover without valid credentials.
Rejects empty and whitespace-only passwords before attempting the LDAP bind.
The APIKeyRestrictionMiddleware only inspected the Authorization header for sk- tokens, but get_current_user also reads API keys from cookies and x-api-key headers. This allowed complete bypass of endpoint restrictions by sending the key via an alternate transport.
Moves the restriction check into get_current_user_by_api_key so it runs regardless of how the API key was delivered. Removes the now-redundant middleware.
Unlike all other resource routers (knowledge, models, notes, prompts, tools, skills), the channel router did not call filter_allowed_access_grants. This allowed any user to set wildcard access grants on group channels, bypassing the admin's public sharing permission framework.
Adds filter_allowed_access_grants with the sharing.public_channels permission key to both create and update endpoints, matching the pattern used by all other resource routers.
The OAuth token exchange endpoint skipped the domain allowlist check that the normal OAuth callback enforces. An attacker with a valid OAuth token from a non-allowed domain (e.g. gmail.com) could bypass the admin's domain restriction policy entirely.
Adds the same domain validation check used in the OAuth callback, denying access when the email domain is not in the allowed list.
SESSION_POOL caches user.role at connection time and never refreshes it. When an admin demotes or deletes a user, their socket sessions retain the old cached role until voluntary disconnect, allowing continued use of admin-gated socket features (ydoc editing, channel access).
Adds disconnect_user_sessions() helper that disconnects all sockets for a user ID. Called from update_user_by_id (on role change) and delete_user_by_id. The client auto-reconnects and re-authenticates with fresh DB data.
The catch-all /{path:path} proxy forwards any request to the upstream OpenAI-compatible API with the admin's API key and no access control. This is an intentional proxy but should be opt-in.
Adds ENABLE_OPENAI_API_PASSTHROUGH env var (defaults to False). When disabled, the catch-all returns 403. No other routers (Ollama, responses) have catch-all proxies.
The GET /channels/{id}/members endpoint checked membership for group/dm channels but had no access gate for standard channels, allowing any authenticated user with channels permission to enumerate members of private standard channels by UUID.
The model name from user input was interpolated directly into Azure deployment URL paths without validation. A user could send a model name like '../../management/foo' to traverse the URL path and hit unintended Azure endpoints with the admin's API key.
Adds _sanitize_model_for_url that rejects path separators and traversal sequences, and percent-encodes the name. Applied at convert_to_azure_payload (covers chat completions + proxy) and the responses endpoint's direct URL construction.
These four endpoints checked model existence but never verified the user has read access via AccessGrants, allowing any authenticated user to use restricted models.
Uses the canonical check_model_access helper from utils.access_control.
Both LDAP and OAuth registration checked user count before insert to determine whether to assign admin role. With multiple workers, concurrent first-user registrations could each see zero users and both create admin accounts.
Applies the insert-first-check-after pattern already used by signup_handler: insert with DEFAULT_USER_ROLE, then atomically check get_num_users()==1 and promote only the sole user to admin.
is_user_channel_member and is_user_channel_manager did not filter on is_active, allowing deactivated members to retain read/write access to group channels via direct API calls.
The /responses proxy endpoint only required authentication via
get_verified_user but did not check per-model access grants. This
allowed any authenticated user to access any model through this
endpoint, bypassing the access control system.
Extract a shared check_model_access helper into utils/access_control
and replace all inline access control blocks across openai.py and
ollama.py (7 locations) with calls to this helper. This eliminates
code duplication and prevents future policy drift between endpoints.
CWE-862: Missing Authorization
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:H (6.5 Medium)
Introduces REDIS_HEALTH_CHECK_INTERVAL and wires it through to every
Redis client created by get_redis_connection (plain, cluster and
sentinel paths, sync and async). When set, redis-py will PING any
connection idle longer than the interval on checkout, so dead sockets
are surfaced as reconnectable errors before a real command lands on
them.
Defaults to unset (empty string) so existing deployments see no
behavioural change. Operators who want the protection should set it
shorter than the Redis server `timeout` setting and any firewall/LB
idle timeout on the path to Redis.
Co-authored-by: Claude <noreply@anthropic.com>
Introduces REDIS_SOCKET_KEEPALIVE and wires socket_keepalive=True
through to every Redis client created by get_redis_connection
(plain, cluster and sentinel paths, sync and async). When enabled,
the kernel sends TCP keepalive probes on idle connections so
half-closed sockets (e.g. after a silent firewall/LB reset or a NIC
flap) are detected before the next command lands on them and the
request never sees a "Connection reset by peer" error.
Defaults to off so existing deployments see no behavioural change.
Operators who want the protection set REDIS_SOCKET_KEEPALIVE=true
in their environment.
Co-authored-by: Claude <noreply@anthropic.com>
* fix(redis): honor REDIS_SOCKET_CONNECT_TIMEOUT on non-sentinel clients
Previously only the sentinel path passed REDIS_SOCKET_CONNECT_TIMEOUT
through to the Redis client. Plain redis:// and cluster URLs fell back
to redis-py's default (no explicit connect timeout), so a hung Redis
or a black-holed network path could stall the whole worker until the
kernel gave up. Forwarding the same env var to from_url()/RedisCluster
keeps the behavior consistent across all deployment topologies.
* fix(redis): gate socket_connect_timeout on is-not-None, not truthiness
Addresses review feedback: the truthiness check on REDIS_SOCKET_CONNECT_TIMEOUT
silently dropped an explicit 0 value and was inconsistent with the sentinel
construction path, which forwards the value directly. Switch to `is not None`
so any user-configured value (including 0) is passed through to from_url()
and RedisCluster.from_url().
---------
Co-authored-by: Claude <noreply@anthropic.com>
Differentiate between "Allow File Upload" and "Allow Web Upload"
in Chinese translations to help administrators understand the
distinction:
- "Allow File Upload" = local file, cloud storage uploads
- "Allow Web Upload" = URL, YouTube, web content uploads
Using user.id as client_id causes WebSocket deadlocks when the same
user generates images concurrently (e.g., multi-model chat). ComfyUI
routes messages by clientId, so shared IDs mean only one connection
receives the completion — others hang forever.
Generate a unique UUID per request, matching ComfyUI's own examples.
Ollama recently added Responses API support via its OpenAI-compatible
endpoint (/v1/responses). This adds a proxy endpoint to the Ollama
router that forwards requests to Ollama's /v1/responses, applying
the same model resolution, access control, and prefix_id handling
used by the existing /v1/chat/completions and /v1/messages proxies.
Uses a typed ResponsesForm Pydantic model with required model field
and extra='allow' for forward compatibility, consistent with other
endpoint schemas in the file.
This allows API consumers (Codex, Claude Code, etc.) to use the
Responses API directly with Ollama-hosted models without requiring
a separate OpenAI-compatible connection.
Azure offers two URL formats: the legacy deployment-based format
(/openai/deployments/{model}/...) and the newer v1 format
(/openai/v1/...) where the model stays in the payload body and no
api-version query parameter is needed.
Previously, the code always ran convert_to_azure_payload which
rewrites the URL to the deployment format, causing 404 errors for
users with v1-style base URLs. Now, when the base URL contains
'/openai/v1', we skip deployment URL construction and route
directly.
Applied consistently across all three Azure routing paths:
generate_chat_completion, /responses proxy, and generic proxy.
The built-in search_web tool hardcoded count=5 as the default,
ignoring the admin-configured WEB_SEARCH_RESULT_COUNT setting.
When the LLM did not specify a count, the tool always returned 5
results regardless of admin configuration.
Now the tool defaults to the admin-configured value when the LLM
omits the count parameter, while still capping LLM-requested
values at the admin maximum to prevent abuse.
Closes#23485
Adds Tamil (India) translations for all existing keys including
Automations feature strings (create, schedule, run, logs, etc.).
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
The add_file_context function used a positional zip() to pair API
payload messages with DB-stored messages. After
process_messages_with_output() expands assistant messages containing
tool calls into multiple OpenAI-format messages (assistant + tool
results), the payload list becomes longer than the stored list. This
caused the zip to misalign, so subsequent user messages never received
their attached_files tags -- the model could see uploaded images via
vision but had no file URL to pass to edit_image.
Fix: filter both lists to user-role messages only before zipping.
User messages maintain the same order in both lists regardless of
assistant message expansion, restoring correct file context injection.
Fixes#21878
New **pt-BR** translations for items introduced in the latest releases, plus a consistency/quality pass across existing strings (grammar, tone, capitalization, pluralization). Placeholders and hotkeys preserved. No logic changes.
The previous regular expression used manual surrogate-pair ranges to
match emojis and missed a large category of commonly used symbols:
/[\uD800-\uDBFF][\uDC00-\uDFFF]|\uD83C[\uDC00-\uDFFF]|\uD83D[\uDC00-\uDE4F]/g
This approach only covers emojis encoded as surrogate pairs (U+1F000 –
U+1F4FF range), but silently skips BMP emojis that use a text-
presentation code point followed by the variation selector U+FE0F,
such as ❤️ (U+2764 U+FE0F), ☀️, ✅, ⚡, ⭐, and keycap sequences
like 1️⃣, as well as ZWJ family sequences (👨👩👧👦) and flag sequences.
Replace with the Unicode property escape \p{RGI_Emoji} using the 'v'
(unicodeSets) flag introduced in ES2024. This single pattern covers
every standardised emoji sequence defined by Unicode, including all
the cases above.
Browser support: Chrome 112+, Firefox 116+, Safari 17+, Node.js 20+.
All browsers targeted by open-webui already support this syntax.
Co-authored-by: Tim Baek <tim@openwebui.com>
Co-authored-by: joaoback <156559121+joaoback@users.noreply.github.com>
Co-authored-by: yoloni <yoloni@tencent.com>
Replace bare except clauses with except Exception to follow Python best practices and avoid catching unexpected system exceptions like KeyboardInterrupt and SystemExit.
Add hamburger menu with bulk actions (Enable All, Disable All, Show All,
Hide All) to the workspace Models page filter bar, matching the
functionality from the admin Settings Models page.
Add flex-wrap, max-h-20, and overflow-y-auto to the groups container
in UserStatus.svelte to prevent horizontal overflow when users have
many groups. Groups now wrap to multiple rows with a scrollbar when
exceeding the max height.
Co-authored-by: QinQin <qinqin@geotab.com>
Add AUDIT_INCLUDED_PATHS env var for whitelist-based audit filtering.
When set, only matching paths are audited and AUDIT_EXCLUDED_PATHS is
ignored. Auth endpoints (signin/signout/signup) are always logged
regardless of filtering mode.
* Prefer model-provided web search result count over admin default
Update `search_web` to prioritize the model-provided `count` parameter before falling back to the admin-configured `WEB_SEARCH_RESULT_COUNT`, and finally defaulting to 5.
Changes:
- Set `count` default to `None` instead of `5`.
- Adjust fallback order to: model-provided `count` → admin-configured value → `5`.
- Update comment to reflect the new precedence logic.
This ensures explicit model requests for result count are respected while preserving sensible defaults.
* Enforce maximum web search result count from config
Update `search_web` to cap the model-provided `count` parameter at the admin-configured `WEB_SEARCH_RESULT_COUNT` to prevent excessive result requests.
Changes:
- Set default `count` parameter to `5`.
- Replace fallback logic with enforcement logic that limits `count` to the configured maximum.
- Update comment to reflect that the result count is now capped to prevent abuse.
This ensures web search requests cannot exceed the configured limit while maintaining a sensible default.
Some databases contain chat records where 'history' or 'messages' are
stored as lists instead of dicts. This causes an AttributeError
('list' object has no attribute 'items') during the
8452d01d26d7_add_chat_message_table migration.
Add isinstance checks to skip chat records with unexpected data shapes
gracefully, matching the existing pattern used for individual message
validation.
Rework shimmer animation timing and gradient for a smoother sweep.
Slower start with acceleration through the text using a custom
cubic-bezier curve. Softer highlight contrast in both light and
dark modes.
description:Create a detailed bug report to help us improve Open WebUI.
description:Tell us what broke in Open WebUI.
title:'issue:'
labels: ['bug', 'triage']
assignees:[]
body:
- type:markdown
attributes:
value:|
# Bug Report
## Important Notes
Use this for real, reproducible bugs. A clear issue is the most useful contribution: include the affected workflow, the expected result, the actual result, and the details needed for someone else to reproduce it.
- **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.**
Before submitting, search open and closed [Issues](https://github.com/open-webui/open-webui/issues) and [Discussions](https://github.com/open-webui/open-webui/discussions). The issue may already be reported or fixed on `dev`.
- 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!**
Please do not open a code pull request for this report unless a maintainer asks for one, or the change is only i18n/localization. If you want to share code as reference, include it here as a local diff or patch. Actionable reproduction details are the most useful next step.
- **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.
- **Contributing**: If you encounter an issue, consider submitting a pull request or forking the project. We prioritize preventing contributor burnout to maintain Open WebUI's quality.
- **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).
Security vulnerabilities must not be reported publicly. Use the [GitHub security page](https://github.com/open-webui/open-webui/security) instead.
- type:checkboxes
id:issue-check
attributes:
label:Check Existing Issues
description:Confirm that you’ve checked for existing reports before submitting a new one.
label:Before Submitting
options:
- label:I have searched for any existing and/or related issues.
- label:I searched open and closed issues and discussions for an existing report.
required:true
- label:I have searched for any existing and/or related discussions.
- label:I checked whether this is already fixed on the `dev` branch or latest source.
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!).
- label:I understand that maintainers want a well-written issue before any code pull request.
required:true
- label:I am using the latest version of Open WebUI.
- label:I am using the latest available version of Open WebUI for my install method.
required:true
- label:This is not a security vulnerability.
required:true
- type:dropdown
@@ -44,9 +38,9 @@ body:
label:Installation Method
description:How did you install Open WebUI?
options:
- Git Clone
- Pip Install
- Docker
- Pip Install
- Git Clone
- Other
validations:
required:true
@@ -55,67 +49,51 @@ body:
id:open-webui-version
attributes:
label:Open WebUI Version
description:Specify the version (e.g., v0.6.26)
description:Specify the version, commit, or image tag.
placeholder:v0.11.0, dev commit SHA, or Docker tag
validations:
required:true
- type:input
id:ollama-version
attributes:
label:Ollama Version (if applicable)
description:Specify the version (e.g., v0.2.0, or v0.1.32-rc1)
validations:
required:false
- type:input
id:operating-system
attributes:
label:Operating System
description:Specify the OS (e.g., Windows 10, macOS Sonoma, Ubuntu 22.04, Debian 12)
- List URLs visited, user input (incl. example values/emails/passwords if needed),
- Describe all options and toggles enabled or changed,
- Include any files or environmental changes,
- Identify the expected and actual result at each stage,
- Ensure any reasonably skilled user can follow and hit the same issue.
required:true
label:Ollama Version
description:Include this if Ollama is involved.
placeholder:v0.32.5
validations:
required:false
- type:textarea
id:summary
attributes:
label:Summary
description:What is wrong, in a few sentences?
validations:
required:true
- type:textarea
id:expected-behavior
attributes:
label:Expected Behavior
description:Describe what should have happened.
description:What should have happened?
validations:
required:true
@@ -123,7 +101,7 @@ body:
id:actual-behavior
attributes:
label:Actual Behavior
description:Describe what actually happened.
description:What actually happened?
validations:
required:true
@@ -131,32 +109,21 @@ body:
id:reproduction-steps
attributes:
label:Steps to Reproduce
description:|
Please provide a **very detailed, step-by-step guide** to reproduce the issue. Your instructions should be so clear and precise that anyone can follow them without guesswork. Include every relevant detail—settings, configuration options, exact commands used, values entered, and any prerequisites or environment variables.
**If full reproduction steps and all relevant settings are not provided, your issue may not be addressed.**
**If your steps to reproduction are incomplete, lacking detail or not reproducible, your issue can not be addressed.**
description:Include the exact commands, settings, URLs, model/provider setup, and user actions needed to hit the bug.
placeholder:|
Example (include every detail):
1. Start with a clean Ubuntu 22.04 install.
2. Install Docker v24.0.5 and start the service.
3. Clone the Open WebUI repo (git clone ...).
4. Use the Docker Compose file without modifications.
5. Open browser Chrome 115.0 in incognito mode.
6. Go to http://localhost:8080 and log in with user "test@example.com".
7. Set the language to "English" and theme to "Dark".
8. Attempt to connect to Ollama at "http://localhost:11434".
9. Observe that the error message "Connection refused" appears at the top right.
Please list each step carefully and include all relevant configuration, settings, and options.
1. Start Open WebUI with ...
2. Configure ...
3. Open ...
4. Click ...
5. See ...
validations:
required:true
- type:textarea
id:logs-screenshots
attributes:
label:Logs & Screenshots
description:Include relevant logs, errors, or screenshots to help diagnose the issue.
placeholder:'Attach logs from the browser console, Docker logs, or error messages.'
label:Logs, Screenshots, and Config
description:Include relevant browser console logs, server/container logs, screenshots, and configuration. If something does not apply, say so.
validations:
required:true
@@ -164,13 +131,6 @@ body:
id:additional-info
attributes:
label:Additional Information
description:Provide any extra details that may assist in understanding the issue.
description:Anything else that might help us understand the report.
validations:
required:false
- type:markdown
attributes:
value:|
## Note
**If the bug report is incomplete, does not follow instructions or is lacking details it may not be addressed.** Ensure that you've followed all the **README.md** and **troubleshooting.md** guidelines, and provide all necessary information for us to reproduce the issue.
about:Do NOT open a public issue for security vulnerabilities, suspected vulnerabilities, or any security-related concern. Please review our Security Policy and report privately via the "Report a vulnerability" button so it can be handled as a private advisory.
description:Describe what you would like Open WebUI to support.
title:'feat:'
labels: ['triage']
body:
- type:markdown
attributes:
value:|
## Important Notes
### Before submitting
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.
# Feature Request
#### Scope
Describe the requested behavior, the problem it solves, and any examples, mockups, screenshots, or workflows that clarify the request. A clear issue or discussion is the most useful contribution.
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**.
Search open and closed [Issues](https://github.com/open-webui/open-webui/issues) and [Discussions](https://github.com/open-webui/open-webui/discussions) before submitting. If the request needs broad product, UX, architecture, compatibility, or maintenance discussion, please start in [Discussions](https://github.com/open-webui/open-webui/discussions) so the community can weigh in.
Remember:
- Open WebUI is a **volunteer-driven project**
- It's managed by a **single maintainer**
- It's supported by contributors who also have **full-time jobs**
Please do not open a code pull request for this request unless a maintainer asks for one, or the change is only i18n/localization. If you want to share code as reference, include it here as a local diff or patch. Clear product context is the most useful next step.
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.
### Bug reproducibility
If a bug cannot be reproduced with a `:main` or `:dev` Docker setup, or a `pip install` with Python 3.11, it may require additional help from the community. In such cases, we will move it to the "[issues](https://github.com/open-webui/open-webui/discussions/categories/issues)" Discussions section due to our limited resources. We encourage the community to assist with these issues. Remember, it’s not that the issue doesn’t exist; we need your help!
Security vulnerabilities must not be reported publicly. Use the [GitHub security page](https://github.com/open-webui/open-webui/security) instead.
- type:checkboxes
id:existing-issue
id:existing-request
attributes:
label:Check Existing Issues
description:Please confirm that you've checked for existing similar requests
label:Before Submitting
options:
- 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.
- label:I searched open and closed issues and discussions for an existing 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.
- label:I checked whether this already exists on the `dev` branch or latest source.
required:true
- label:I understand that maintainers want a well-written issue or discussion before any code pull request.
required:true
- label:This request is not a security vulnerability.
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... / Not related to a problem"
label:Problem
description:What is missing, frustrating, confusing, or unnecessarily hard today?
placeholder:"I'm trying to..., but..."
validations:
required:true
- type:textarea
id:solution-description
id:desired-behavior
attributes:
label:Desired Solution you'd like
description:Clearly describe what you want to happen.
label:Desired Behavior
description:What would you like to happen instead?
placeholder:"I would like Open WebUI to..."
validations:
required:true
- type:textarea
id:why-it-matters
attributes:
label:Why This Matters
description:Who benefits, and what workflow does this unlock or improve?
validations:
required:true
- type:textarea
id:examples
attributes:
label:Examples or References
description:Add mockups, screenshots, links, prompts, workflows, or examples from other tools.
validations:
required:false
- type:textarea
id:alternatives-considered
attributes:
label:Alternatives Considered
description:A clear and concise description of any alternative solutions or features you've considered.
- type:textarea
id:additional-context
attributes:
label:Additional Context
description:Add any other context or screenshots about the feature request here.
label:Alternatives or Workarounds
description:What have you tried instead, if anything?
⚠️ 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.
Important checks for contributors:
1. Target the `dev` branch. PRs targeting `main` will be closed.
2. Code pull requests are not the default contribution path.
3. Do not open a code PR as the first step. Start with a well-written Issue or Discussion unless a maintainer asked for the PR or the change is only i18n/localization.
4. Do not delete the Contributor License Agreement section at the bottom. The CLA bot requires it.
-->
# Pull Request Checklist
# 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.
Thanks for wanting to improve Open WebUI. The most useful contribution isusually a clear, well-written Issue, not an unsolicited code 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.
Open a code pull request only when a maintainer asks for one, or when the change is only i18n/localization. For real, reproducible bugs, start with a well-described [Issue](https://github.com/open-webui/open-webui/issues). For feature requests, UI/UX changes, behavior changes, architecture changes, suspected fixes, or unconfirmed approaches, start with an active [Discussion](https://github.com/open-webui/open-webui/discussions).
**Before submitting, make sure you've checked the following:**
Before continuing, make sure the linked Issue or Discussion explains the user-facing problem, the expected outcome, the affected workflow, and any examples, logs, screenshots, constraints, or reproduction details needed for maintainers to evaluate it.
- [ ]**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:** 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
- **ci**: Changes to our continuous integration processes or workflows
- **chore**: Refactor, cleanup, or other non-functional code changes
- **docs**: Documentation update or addition
- **feat**: Introduces a new feature or enhancement to the codebase
- **fix**: Bug fix or error correction
- **i18n**: Internationalization or localization changes
- **perf**: Performance improvement
- **refactor**: Code restructuring for better maintainability, readability, or scalability
- **style**: Changes that do not affect the meaning of the code (white space, formatting, missing semi-colons, etc.)
- **test**: Adding missing tests or correcting existing tests
- **WIP**: Work in progress, a temporary label for incomplete or ongoing work
If you have implementation notes, include them as reference in the Issue or Discussion. If you want to share code as reference, include it there as a local diff, patch, or branch note. Do not open a pull request for reference code.
# Changelog Entry
Unsolicited PRs may be closed without review, especially when they introduce product, architecture, compatibility, dependency, or maintenance decisions that have not been discussed.
### Description
## Checklist
- [Concisely describe the changes made in this pull request, including any relevant motivation and impact (e.g., fixing a bug, adding a feature, or improving performance)]
- [ ] This PR targets the `dev` branch.
- [ ] This PR links to a well-described, confirmed Issue or active Discussion: `Closes #___` / `Relates to #___`.
- [ ] A maintainer explicitly asked me to open this PR, or this PR only updates i18n/localization.
- [ ] The change is one logical unit with no unrelated commits.
- [ ] I matched nearby code patterns and avoided unnecessary new settings, abstractions, or dependencies.
- [ ] I manually tested the changed workflow and any nearby behavior that could be affected.
- [ ] I updated relevant docs, including the [Open WebUI Docs Repository](https://github.com/open-webui/docs), if needed.
- [ ] I added screenshots for UI changes, and a recording when motion or interaction matters.
- [ ] I reviewed any AI-generated code before submitting it.
- [ ] The PR title uses one of the prefixes listed below.
- **chore**: Refactoring, cleanup, or non-functional changes
- **docs**: Documentation additions or updates
- **feat**: New features or enhancements
- **fix**: Bug fixes or corrections
- **i18n**: Internationalization or localization changes
- **perf**: Performance improvements
- **refactor**: Code restructuring
## Summary
Describe the change, the problem it solves, and the impact on users.
## Testing
List the exact manual checks you ran. Include commands, setup details, screenshots, or recordings where helpful.
## Changelog Entry
### Added
- [List any new features, functionalities, or additions]
-
### Changed
- [List any changes, updates, refactorings, or optimizations]
### Deprecated
- [List any deprecated functionality or features that have been removed]
### Removed
- [List any removed features, files, or functionalities]
-
### Fixed
- [List any fixes, corrections, or bug fixes]
-
### Removed
-
### Security
- [List any new or updated security-related changes, including vulnerability fixes]
-
### Breaking Changes
- **BREAKING CHANGE**: [List any breaking changes affecting compatibility or functionality]
-
---
## Additional Context
### Additional Information
Add anything maintainers should know before review.
- [Insert any additional context, notes, or explanations for the changes]
- [Reference any related issues, commits, or other relevant information]
### Screenshots or Videos
- [Attach any relevant screenshots or videos demonstrating the changes]
### Contributor License Agreement
## 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.
Your PR will NOT be reviewed or merged until you check the box below confirming that you have read and agree to the terms of the CLA.
DO NOT DELETE THIS SECTION.
Your PR will not be reviewed or merged until you check the box below confirming that you have read and agree to the CLA.
-->
- [ ] 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.
@@ -20,7 +20,7 @@ Examples of behavior that contribute to a positive and professional community in
- **Respecting others.** Be considerate, listen actively, and engage with empathy toward others' viewpoints and experiences.
- **Constructive feedback.** Provide actionable, thoughtful, and respectful feedback that helps improve the project and encourages collaboration. Avoid unproductive negativity or hypercriticism.
- **Recognizing volunteer contributions.** Appreciate that contributors dedicate their free time and resources selflessly. Approach them with gratitude and patience.
- **Recognizing volunteer contributions.** Appreciate that **contributors dedicate their free time and resources selflessly**. Approach them with gratitude and patience.
- **Focusing on shared goals.** Collaborate in ways that prioritize the health, success, and sustainability of the community over individual agendas.
Examples of unacceptable behavior include:
@@ -32,6 +32,7 @@ Examples of unacceptable behavior include:
- **Entitlement, demand, or aggression toward contributors.** Volunteers are under no obligation to provide immediate or personalized support. Rude or dismissive behavior will not be tolerated.
- **Unproductive or destructive behavior.** This includes venting frustration as hostility ("tantrums"), hypercriticism, attention-seeking negativity, or anything that distracts from the project's goals.
- **Spamming and promotional exploitation.** Sharing irrelevant product promotions or self-promotion in the community is not allowed unless it directly contributes value to the discussion.
- Posting low-effort, hard to read, essay-length AI generated comments or other forms of low-quality, hard to parse content that puts the burden of understanding on the reader.
Copyright (c) 2023- Open WebUI Inc. [Created by Timothy Jaeryang Baek]
All rights reserved.
@@ -15,11 +17,27 @@ modification, are permitted provided that the following conditions are met:
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
4. Notwithstanding any other provision of this License, and as a material condition of the rights granted herein, licensees are strictly prohibited from altering, removing, obscuring, or replacing any "Open WebUI" branding, including but not limited to the name, logo, or any visual, textual, or symbolic identifiers that distinguish the software and its interfaces, in any deployment or distribution, regardless of the number of users, except as explicitly set forth in Clauses 5 and 6 below.
4. Notwithstanding any other provision of this License, and as a material
condition of the rights granted herein, licensees are strictly prohibited
from altering, removing, obscuring, or replacing any "Open WebUI"
branding, including but not limited to the name, logo, or any visual,
textual, or symbolic identifiers that distinguish the software and its
interfaces, in any deployment or distribution, except in the following
circumstances: (i) deployments or distributions where the total number
of end users (defined as individual natural persons with direct access
to the application) does not exceed fifty (50) within any rolling
thirty (30) day period; (ii) the licensee has obtained specific prior
written permission from the copyright holder; or (iii) where the
licensee has obtained a duly executed enterprise license expressly
permitting such modification. For all other cases, any removal or
alteration of the "Open WebUI" branding shall constitute a material
breach of license.
5. The branding restriction enumerated in Clause 4 shall not apply in the following limited circumstances: (i) deployments or distributions where the total number of end users (defined as individual natural persons with direct access to the application) does not exceed fifty (50) within any rolling thirty (30) day period; (ii) cases in which the licensee is an official contributor to the codebase—with a substantive code change successfully merged into the main branch of the official codebase maintained by the copyright holder—who has obtained specific prior written permission for branding adjustment from the copyright holder; or (iii) where the licensee has obtained a duly executed enterprise license expressly permitting such modification. For all other cases, any removal or alteration of the "Open WebUI" branding shall constitute a material breach of license.
Materials governed by prior licenses retain those original license
terms, as specified in LICENSE_HISTORY.
6. All code, modifications, or derivative works incorporated into this project prior to the incorporation of this branding clause remain licensed under the BSD 3-Clause License, and prior contributors retain all BSD-3 rights therein; if any such contributor requests the removal of their BSD-3-licensed code, the copyright holder will do so, and any replacement code will be licensed under the project's primary license then in effect. By contributing after this clause's adoption, you agree to the project's Contributor License Agreement (CLA) and to these updated terms for all new contributions.
By contributing to this project, you agree to the project's Contributor
@@ -27,58 +27,82 @@ For more information, be sure to check out our [Open WebUI Documentation](https:
## Key Features of Open WebUI ⭐
- 🚀 **Effortless Setup**: Install seamlessly using Docker or Kubernetes (kubectl, kustomize or helm) for a hassle-free experience with support for both `:ollama` and `:cuda` tagged images.
- 🚀 **Effortless Setup**: Install seamlessly via pip, uv, Docker, or Kubernetes (kubectl, kustomize, or helm), with `:ollama` and `:cuda` tagged images available for container deployments.
- 🤝 **Ollama/OpenAI API Integration**: Effortlessly integrate OpenAI-compatible APIs for versatile conversations alongside Ollama models. Customize the OpenAI API URL to link with**LMStudio, GroqCloud, Mistral, OpenRouter, and more**.
- 🤝 **Broad Model & API Integration**: Connect any OpenAI-compatible API alongside local Ollama models. Point the API URL at**LMStudio, GroqCloud, Mistral, OpenRouter, vLLM, and more** to mix and match providers freely.
-🛡️**Granular Permissions and User Groups**: By allowing administrators to create detailed user roles and permissions, we ensure a secure user environment. This granularity not only enhances security but also allows for customized user experiences, fostering a sense of ownership and responsibility amongst users.
-🔐**Granular RBAC & User Groups**: Administrators define detailed roles, groups, and permissions, giving each user exactly the access they need. Secure by default, with tailored experiences per group.
-📱**Responsive Design**: Enjoy a seamless experience across Desktop PC, Laptop, and Mobile devices.
-🧩**Plugin Support**: Extend Open WebUI with **Filters**, **Actions**, **Pipes**, **Tools**, and **Skills**. Connect external services through **MCP**, **MCPO**, and **OpenAPI tool servers**. Build custom integrations, rate limits, approval flows, data connections, and more.
-📱**Progressive Web App (PWA) for Mobile**: Enjoy a native app-like experience on your mobile device with our PWA, providing offline access on localhost and a seamless user interface.
-🤖**Models & Agents**: Wrap any base model with custom instructions, tools, and knowledge to build specialized agents. Supports dynamic variables, per-user/group access control, and community preset imports via [Open WebUI Community](https://openwebui.com/).
-✒️🔢**Full Markdown and LaTeX Support**: Elevate your LLM experience with comprehensive Markdown and LaTeX capabilities for enriched interaction.
-📝**Notes**: A dedicated workspace for content outside conversations. Draft with a rich editor, use AI to rewrite selected text, and attach notes to any chat for full-context injection.
-🎤📹**Hands-Free Voice/Video Call**: Experience seamless communication with integrated hands-free voice and video call features using multiple Speech-to-Text providers (Local Whisper, OpenAI, Deepgram, Azure) and Text-to-Speech engines (Azure, ElevenLabs, OpenAI, Transformers, WebAPI), allowing for dynamic and interactive chat environments.
-📢**Channels**: Real-time shared spaces where your team and AI models collaborate in one timeline. Tag models to draft or critique, with threads, reactions, pins, and access control.
-🛠️**Model Builder**: Easily create Ollama models via the Web UI. Create and add custom characters/agents, customize chat elements, and import models effortlessly through [Open WebUI Community](https://openwebui.com/) integration.
-🧠**Persistent Memory**: The AI remembers facts about you across conversations, carrying context from one chat to the next.
-🐍**Native Python Function Calling Tool**: Enhance your LLMs with built-in code editor support in the tools workspace. Bring Your Own Function (BYOF) by simply adding your pure Python functions, enabling seamless integration with LLMs.
-✅**Live Workflow & Message Flow**: Watch the AI build and work through checklists in real time. Queue messages while the AI is still responding; they send automatically when it's ready.
-💾**Persistent Artifact Storage**: Built-in key-value storage API for artifacts, enabling features like journals, trackers, leaderboards, and collaborative tools with both personal and shared data scopes across sessions.
-📅**Calendar & AI Scheduling**: Built-in personal and shared calendars with month/week/day views, recurring events, color coding, attendees, and reminders. Models manage your schedule conversationally through native function calling.
-📚**Local RAG Integration**: Dive into the future of chat interactions with groundbreaking Retrieval Augmented Generation (RAG) support using your choice of 9 vector databases and multiple content extraction engines (Tika, Docling, Document Intelligence, Mistral OCR, External loaders). Load documents directly into chat or add files to your document library, effortlessly accessing them using the `#` command before a query.
-⏱️**Automations**: Schedule prompts to run on recurring schedules, with runs surfaced on your calendar and each completed run linking back to the chat it produced.
-🔍**Web Search for RAG**: Perform web searches using 15+ providers including `SearXNG`, `Google PSE`, `Brave Search`, `Kagi`, `Mojeek`, `Tavily`, `Perplexity`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `SearchApi`, `SerpApi`, `Bing`, `Jina`, `Exa`, `Sougou`, `Azure AI Search`, and `Ollama Cloud`, injecting results directly into your chat experience.
-📱**Responsive Design & PWA**: Seamless experience across desktop, laptop, and mobile, with a Progressive Web App for native app-like feel and offline access on localhost.
-🌐**Web Browsing Capability**: Seamlessly integrate websites into your chat experience using the `#` command followed by a URL. This feature allows you to incorporate web content directly into your conversations, enhancing the richness and depth of your interactions.
-✒️🔢**Full Markdown and LaTeX Support**: Comprehensive Markdown and LaTeX capabilities for enriched interaction.
-🎨**Image Generation & Editing Integration**: Create and edit images using multiple engines including OpenAI's DALL-E, Gemini, ComfyUI (local), and AUTOMATIC1111 (local), with support for both generation and prompt-based editing workflows.
-🎤📹**Hands-Free Voice/Video Call**: Integrated voice and video calls with multiple Speech-to-Text providers (Local Whisper, OpenAI, Deepgram, Azure) and Text-to-Speech engines (Azure, ElevenLabs, OpenAI, Transformers, WebAPI).
-⚙️**Many Models Conversations**: Effortlessly engage with various models simultaneously, harnessing their unique strengths for optimal responses. Enhance your experience by leveraging a diverse set of models in parallel.
-💾**Persistent Artifact Storage**: Built-in key-value storage API for artifacts, enabling journals, trackers, leaderboards, and collaborative tools with personal and shared data scopes.
-🔐**Role-Based Access Control (RBAC)**: Ensure secure access with restricted permissions; only authorized individuals can access your Ollama, and exclusive model creation/pulling rights are reserved for administrators.
-📚**Local RAG Integration**: Retrieval Augmented Generation backed by 9 vector databases and multiple content-extraction engines (Tika, Docling, Document Intelligence, Mistral OCR, PaddleOCR-vl, external loaders). Supports hybrid search (BM25 + vector) with reranking and full-context mode. Load documents into chat or pull them from your library with the `#` command.
-🗄️**Flexible Database & Storage Options**: Choose from SQLite (with optional encryption), PostgreSQL, or configure cloud storage backends (S3, Google Cloud Storage, Azure Blob Storage) for scalable deployments.
-🔍**Web Search for RAG**: Search the web through dozens of providers including `SearXNG`, `Google PSE`, `Brave Search`, `Kagi`, `Mojeek`, `Tavily`, `Perplexity`, `Firecrawl`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `SearchApi`, `SerpApi`, `Bing`, `Jina`, `Exa`, `Sougou`, `Azure AI Search`, and `Ollama Cloud`, injecting results directly into the conversation.
-🔍**Advanced Vector Database Support**: Select from 9 vector database options including ChromaDB, PGVector, Qdrant, Milvus, Elasticsearch, OpenSearch, Pinecone, S3Vector, and Oracle 23ai for optimal RAG performance.
-🌐**Web Browsing Capability**: Pull websites into chat with the `#` command followed by a URL, or let the model fetch them on its own when needed.
-🔐**Enterprise Authentication**: Full support for LDAP/Active Directory integration, SCIM 2.0 automated provisioning, and SSO via trusted headers alongside OAuth providers. Enterprise-grade user and group provisioning through SCIM 2.0 protocol, enabling seamless integration with identity providers like Okta, Azure AD, and Google Workspace for automated user lifecycle management.
-🎨**Image Generation & Editing**: Create and edit images with multiple engines including OpenAI DALL·E, Gemini, ComfyUI (local), and AUTOMATIC1111 (local), supporting both generation and prompt-based editing.
-☁️**Cloud-Native Integration**: Native support for Google Drive and OneDrive/SharePoint file picking, enabling seamless document import from enterprise cloud storage.
-⚙️**Multi-Model Conversations**: Engage several models at once, harnessing their individual strengths in parallel for the best possible responses.
- 📊 **Production Observability**: Built-in OpenTelemetry support for traces, metrics, and logs, enabling comprehensive monitoring with your existing observability stack.
- 📊 **Usage Analytics & Model Evaluation**: Admin dashboards track message volume, token consumption, and cost across users and models. Evaluate models with a built-in arena, A/B testing, and ELO-based leaderboards.
-⚖️**Horizontal Scalability**: Redis-backed session management and WebSocket support for multi-worker and multi-node deployments behind load balancers.
-🗄️**Flexible Database & Storage**: Choose SQLite (with optional encryption) or PostgreSQL, and store files locally or on S3, Google Cloud Storage, or Azure Blob Storage.
-🌐🌍**Multilingual Support**: Experience Open WebUI in your preferred language with our internationalization (i18n) support. Join us in expanding our supported languages! We're actively seeking contributors!
-🧬**Advanced Vector Database Support**: Pick from 9 vector databases: ChromaDB, PGVector, Qdrant, Milvus, Elasticsearch, OpenSearch, Pinecone, S3Vector, and Oracle 23ai.
-🧩**Pipelines, Open WebUI Plugin Support**: Seamlessly integrate custom logic and Python libraries into Open WebUI using [Pipelines Plugin Framework](https://github.com/open-webui/pipelines). Launch your Pipelines instance, set the OpenAI URL to the Pipelines URL, and explore endless possibilities. [Examples](https://github.com/open-webui/pipelines/tree/main/examples) include **Function Calling**, User **Rate Limiting** to control access, **Usage Monitoring** with tools like Langfuse, **Live Translation with LibreTranslate** for multilingual support, **Toxic Message Filtering** and much more.
-🪪**Enterprise Authentication & Provisioning**: Full LDAP/Active Directory integration, SSO via trusted headers and OAuth providers, and SCIM 2.0 automated provisioning for identity providers like Okta, Azure AD, and Google Workspace.
-🌟**Continuous Updates**: We are committed to improving Open WebUI with regular updates, fixes, and new features.
-☁️**Cloud-Native File Integration**: Native Google Drive and OneDrive/SharePoint file picking for seamless document import from enterprise cloud storage.
- 🔭 **Production Observability**: Built-in OpenTelemetry support for traces, metrics, and logs, plugging into your existing monitoring stack.
- ⚖️ **Horizontal Scalability**: Redis-backed session management and WebSocket support for multi-worker, multi-node deployments behind load balancers.
- 🌐🌍 **Multilingual Support**: Use Open WebUI in your preferred language with i18n support. We're actively seeking contributors to expand language coverage!
- 🌟 **Continuous Updates**: We're committed to improving Open WebUI with regular updates, fixes, and new features.
- 🛡️ **Transparent Security Process**: Security reports are triaged, fixed, and published as open advisories through a documented responsible-disclosure process. See our [Security Policy](https://github.com/open-webui/open-webui/security).
Want to learn more about Open WebUI's features? Check out our [Open WebUI documentation](https://docs.openwebui.com/features) for a comprehensive overview!
## The Open WebUI Ecosystem 🌐
Open WebUI is the core, surrounded by companion apps and infrastructure that extend what your AI can do, where it can reach, and how you run it:
- 💻 **Open WebUI Computer** ([open-webui/computer](https://github.com/open-webui/computer)): A standalone, mobile-first computer and coding agent that runs on the machine you own. Files, terminal, and git in a browser tab, reachable from your phone. Connect it into Open WebUI as a model, or reach it from Telegram, WhatsApp, and more.
- ⚡ **Open Terminal** and **Terminals (Enterprise)** ([open-webui/open-terminal](https://github.com/open-webui/open-terminal) & [open-webui/terminals](https://github.com/open-webui/terminals)): A self-hosted computing environment that plugs into Open WebUI, giving the AI a place to write code, run it, read output, fix errors, and iterate inside the chat. Terminals gives you per-user isolated containers with separate credentials, resource limits, and network rules. Automatic lifecycle management on Docker or Kubernetes.
- 🔄 **oikb** ([open-webui/oikb](https://github.com/open-webui/oikb)): Feed your Knowledge Bases from 45+ sources (GitHub, Confluence, ServiceNow, Salesforce, Jira, Slack, SharePoint, Notion, and more), keeping the tools your team already uses continuously in sync.
- 🖥️ **Native Desktop App** ([open-webui/desktop](https://github.com/open-webui/desktop)): Run Open WebUI as a native app on macOS, Windows, and Linux. System-wide Spotlight chat bar with screenshot capture, push-to-talk voice, and optional fully-local inference via a built-in llama.cpp engine.
Want to learn more? Check out our [Open WebUI documentation](https://docs.openwebui.com) for more details!
---
We are incredibly grateful for the generous support of our sponsors. Their contributions help us to maintain and improve our project, ensuring we can continue to deliver quality work to our community. Thank you!
@@ -172,8 +196,6 @@ After installation, you can access Open WebUI at [http://localhost:3000](http://
We offer various installation alternatives, including non-Docker native installation methods, Docker Compose, Kustomize, and Helm. Visit our [Open WebUI Documentation](https://docs.openwebui.com/getting-started/) or join our [Discord community](https://discord.gg/5rJgQTnV4s) for comprehensive guidance.
Look at the [Local Development Guide](https://docs.openwebui.com/getting-started/development) for instructions on setting up a local development environment.
### Troubleshooting
Encountering connection issues? Our [Open WebUI Documentation](https://docs.openwebui.com/troubleshooting/) has got you covered. For further assistance and to join our vibrant community, visit the [Open WebUI Discord](https://discord.gg/5rJgQTnV4s).
@@ -224,6 +246,10 @@ This project contains code under multiple licenses. The current codebase include
If you have any questions, suggestions, or need assistance, please open an issue or join our
[Open WebUI Discord community](https://discord.gg/5rJgQTnV4s) to connect with us! 🤝
## Security 🛡️
If you believe you've found a security vulnerability, or something that shouldn't be disclosed publicly, please [reach out confidentially through our responsible disclosure program on GitHub](https://github.com/open-webui/open-webui/security). We accept reports only through GitHub, not through any other platform. Thank you for helping us keep Open WebUI secure!
returnf'[ERROR: {fallback}]'iffallbackelse'Something went wrong :/'
returnf'[ERROR: {err}]'
classMESSAGES(str,Enum):
DEFAULT=lambdamsg="":f"{msgifmsgelse''}"
MODEL_ADDED=lambdamodel="":f"The model '{model}' has been added successfully."
MODEL_DELETED=(
lambdamodel="":f"The model '{model}' has been deleted successfully."
)
DEFAULT=lambdamsg='':f'{msgifmsgelse""}'
MODEL_ADDED=lambdamodel='':f"The model '{model}' has been added successfully."
MODEL_DELETED=lambdamodel='':f"The model '{model}' has been deleted successfully."
classWEBHOOK_MESSAGES(str,Enum):
DEFAULT=lambdamsg="":f"{msgifmsgelse''}"
USER_SIGNUP=lambdausername="":(
f"New user signed up: {username}"ifusernameelse"New user signed up"
)
DEFAULT=lambdamsg='':f'{msgifmsgelse""}'
USER_SIGNUP=lambdausername='':f'New user signed up: {username}'ifusernameelse'New user signed up'
classERROR_MESSAGES(str,Enum):
def__str__(self)->str:
returnsuper().__str__()
DEFAULT=(
lambdaerr="":f'{"Something went wrong :/"iferr==""else"[ERROR: "+str(err)+"]"}'
DEFAULT=_error_message
ENV_VAR_NOT_FOUND='Required environment variable not found. Terminating now.'
CREATE_USER_ERROR='Oops! Something went wrong while creating your account. Please try again later. If the issue persists, contact support for assistance.'
DELETE_USER_ERROR='Oops! Something went wrong. We encountered an issue while trying to delete the user. Please give it another shot.'
EMAIL_MISMATCH='Uh-oh! This email does not match the email your provider is registered with. Please check your email and try again.'
EMAIL_TAKEN='Uh-oh! This email is already registered. Sign in with your existing account or choose another email to start anew.'
USERNAME_TAKEN='Uh-oh! This username is already registered. Please choose another username.'
PASSWORD_TOO_LONG=(
'Uh-oh! The password you entered is too long. Please make sure your password is less than 72 bytes long.'
)
ENV_VAR_NOT_FOUND="Required environment variable not found. Terminating now."
CREATE_USER_ERROR="Oops! Something went wrong while creating your account. Please try again later. If the issue persists, contact support for assistance."
DELETE_USER_ERROR="Oops! Something went wrong. We encountered an issue while trying to delete the user. Please give it another shot."
EMAIL_MISMATCH="Uh-oh! This email does not match the email your provider is registered with. Please check your email and try again."
EMAIL_TAKEN="Uh-oh! This email is already registered. Sign in with your existing account or choose another email to start anew."
USERNAME_TAKEN=(
"Uh-oh! This username is already registered. Please choose another username."
)
PASSWORD_TOO_LONG="Uh-oh! The password you entered is too long. Please make sure your password is less than 72 bytes long."
COMMAND_TAKEN="Uh-oh! This command is already registered. Please choose another command string."
FILE_EXISTS="Uh-oh! This file is already registered. Please choose another file."
COMMAND_TAKEN='Uh-oh! This command is already registered. Please choose another command string.'
FILE_EXISTS='Uh-oh! This file is already registered. Please choose another file.'
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."
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_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)."
INCORRECT_PASSWORD=(
"The password provided is incorrect. Please check for typos and try again."
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.'
)
INVALID_TRUSTED_HEADER="Your provider has not provided a trusted header. Please contact your administrator for assistance."
EXISTING_USERS="You can't turn off authentication because there are existing users. If you want to disable WEBUI_AUTH, make sure your web interface doesn't have any existing users and is a fresh installation."
UNAUTHORIZED="401 Unauthorized"
ACCESS_PROHIBITED="You do not have permission to access this resource. Please contact your administrator for assistance."
ACTION_PROHIBITED=(
"The requested action has been restricted as a security measure."
UNAUTHORIZED='401 Unauthorized'
ACCESS_PROHIBITED=(
'You do not have permission to access this resource. Please contact your administrator for assistance.'
)
ACTION_PROHIBITED='The requested action has been restricted as a security measure.'
FILE_NOT_SENT="FILE_NOT_SENT"
FILE_NOT_SENT='FILE_NOT_SENT'
FILE_NOT_SUPPORTED="Oops! It seems like the file format you're trying to upload is not supported. Please upload a file with a supported format and try again."
NOT_FOUND="We could not find what you're looking for :/"
USER_NOT_FOUND="We could not find what you're looking for :/"
API_KEY_NOT_FOUND="Oops! It looks like there's a hiccup. The API key is missing. Please make sure to provide a valid API key to access this feature."
API_KEY_NOT_ALLOWED="Use of API key is not enabled in the environment."
API_KEY_NOT_ALLOWED='Use of API key is not enabled in the environment.'
MALICIOUS="Unusual activities detected, please try again in a few minutes."
MALICIOUS='Unusual activities detected, please try again in a few minutes.'
PANDOC_NOT_INSTALLED="Pandoc is not installed on the server. Please contact your administrator for assistance."
INCORRECT_FORMAT=(
lambdaerr="":f"Invalid format. Please use the correct format{err}"
)
RATE_LIMIT_EXCEEDED="API rate limit exceeded"
PANDOC_NOT_INSTALLED='Pandoc is not installed on the server. Please contact your administrator for assistance.'
INCORRECT_FORMAT=lambdaerr='':f'Invalid format. Please use the correct format{err}'
RATE_LIMIT_EXCEEDED='API rate limit exceeded'
MODEL_NOT_FOUND=lambdaname="":f"Model '{name}' was not found"
OPENAI_NOT_FOUND=lambdaname="":"OpenAI API was not found"
OLLAMA_NOT_FOUND="WebUI could not connect to Ollama"
CREATE_API_KEY_ERROR="Oops! Something went wrong while creating your API key. Please try again later. If the issue persists, contact support for assistance."
API_KEY_CREATION_NOT_ALLOWED="API key creation is not allowed in the environment."
MODEL_NOT_FOUND=lambdaname='':f"Model '{name}' was not found"
OPENAI_NOT_FOUND=lambdaname='':'OpenAI API was not found'
OLLAMA_NOT_FOUND='WebUI could not connect to Ollama'
CREATE_API_KEY_ERROR='Oops! Something went wrong while creating your API key. Please try again later. If the issue persists, contact support for assistance.'
API_KEY_CREATION_NOT_ALLOWED='API key creation is not allowed in the environment.'
EMPTY_CONTENT="The content provided is empty. Please ensure that there is text or data present before proceeding."
EMPTY_CONTENT='The content provided is empty. Please ensure that there is text or data present before proceeding.'
DB_NOT_SQLITE="This feature is only available when running with SQLite databases."
DB_NOT_SQLITE='This feature is only available with SQLite databases.'
INVALID_URL=(
"Oops! The URL you provided is invalid. Please double-check and try again."
INVALID_URL='The URL you provided is invalid. Please double-check and try again.'
WEB_SEARCH_ERROR='Something went wrong while searching the web.'
OLLAMA_API_DISABLED='The Ollama API is disabled. Please enable it to use this feature.'
FILE_TOO_LARGE=lambdasize='':(
f"Oops! The file you're trying to upload is too large. Please upload a file that is less than {size}."
)
WEB_SEARCH_ERROR=(
lambdaerr="":f"{erriferrelse'Oops! Something went wrong while searching the web.'}"
DUPLICATE_CONTENT='Duplicate content detected. Please provide unique content to proceed.'
FILE_NOT_PROCESSED=(
'Extracted content is not available for this file. Please ensure that the file is processed before proceeding.'
)
OLLAMA_API_DISABLED=(
"The Ollama API is disabled. Please enable it to use this feature."
)
INVALID_PASSWORD=lambdaerr='':erriferrelse'The password does not meet the required validation criteria.'
FILE_TOO_LARGE=(
lambdasize="":f"Oops! The file you're trying to upload is too large. Please upload a file that is less than {size}."
# ### commands auto generated by Alembic - please adjust! ###
op.drop_column("file","updated_at")
op.drop_column("file","data")
op.drop_column("file","hash")
op.drop_column('file','updated_at')
op.drop_column('file','data')
op.drop_column('file','hash')
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Timothy Jaeryang Baek