Commit Graph
18833 Commits
Author SHA1 Message Date
G30andGitHub afeab2e2a7 fix: clear a pending channel reply when switching to another channel (#31376) 2026-09-30 19:59:32 +04:00
bb8e986545 i18n: translate missing Simplified Chinese settings labels (#31519)
Co-authored-by: Alex0AI <206435355+Alex0AI@users.noreply.github.com>
2026-09-30 19:59:24 +04:00
G30andGitHub d4d04dca0d fix: keep a cloned chat in a shared folder the user can write to (#31370) 2026-09-30 19:59:11 +04:00
Classic298andGitHub c7caa1421d fix: tool HTML embeds vanish when the HTML contains entities like &quot; (#31390)
Tools that return inline HTML showed no embed at all when the HTML contained an entity such as `&quot;` or `&#34;`, and HTML containing `&amp;` or `&lt;` showed up with those turned into real characters. The chat view unescaped HTML entities in a tool's embeds, arguments and file links, which is only right for chats saved by older versions, so on current chats it changed the tool's own HTML before displaying it. Embeds, tool arguments and file links now show exactly what the tool returned, and older chats render as before.

Fixes #28085
2026-09-30 19:50:11 +04:00
Classic298andGitHub fab58bd35f fix: ejecting a model ignores AIOHTTP_CLIENT_SESSION_SSL (#31391)
Ejecting a loaded model from the model selector failed with a certificate error on llama.cpp and Ollama connections served over HTTPS with a self-signed or internal CA, even though chatting with the same connection worked. The unload request skipped the AIOHTTP_CLIENT_SESSION_SSL setting and always verified against the default system certificates, so setting it to a CA bundle or to false had no effect there. It now uses that setting, the same way chat requests to the connection already do.

Fixes #31371
2026-09-30 19:49:59 +04:00
Classic298andGitHub d09ab78fce fix: typing with an input method breaks in the empty chat input while a follow-up suggestion is shown (#31393)
After a reply, the first suggested follow-up question sits as grey ghost text in the empty chat input. Typing into it with an input method (Chinese, Japanese or Korean) broke the word being composed: on iOS the input lost focus after the first character, and in Chromium the first letter was left behind, so typing "ni" and picking "你" gave "n你". Clearing the ghost text rebuilt the input's line of text while the keyboard was still composing the word. The ghost text is now drawn over the empty input without being written into it, so input methods work again, and Tab to accept it and the follow-up buttons under the reply behave as before.

Fixes #31372
2026-09-30 19:49:49 +04:00
Classic298andGitHub a5176f4cda fix: Google MCP connections drop about an hour after signing in (#31395)
MCP tool servers that sign in through Google, such as Google's hosted Gmail, Drive and Calendar servers, never got a refresh token, because Google only issues one when the sign-in explicitly asks for offline access. When the one-hour access token ran out the refresh failed and the connection was removed, so every user had to sign in again every hour. When the server's sign-in page is Google's, the sign-in now asks for offline access and a fresh consent, so the token renews on its own. Other providers get the same sign-in request as before, and existing Google connections pick this up the next time the user signs in.

Fixes #28319
2026-09-30 19:49:25 +04:00
Classic298andGitHub a29c969fc7 fix: SCIM group members are returned with "$ref": null (#31529)
Every member listed in a SCIM group response came back with "$ref": null, so identity providers had no link from a group member to that user's SCIM resource. Members now carry the URL of the user they point to, on both the single group and the group list responses.

Fixes #31525
2026-09-30 19:47:11 +04:00
G30andGitHub e276d33352 fix: keep a note's pasted images when the note is rebuilt from HTML (#31513) 2026-09-30 19:47:05 +04:00
Classic298andGitHub 909a2075d3 fix: Scheduled Tasks calendar shows extra runs for automations with a run count (#31604)
An automation whose schedule ends after a fixed number of runs (COUNT in its RRULE) showed extra future runs in the Scheduled Tasks calendar after each run. With COUNT=3, the calendar kept showing three upcoming runs after the first and second run, although only the remaining ones execute. The calendar now counts runs from the schedule's own start date, the same way automations are actually run, so it shows exactly the runs that are still going to happen. The 5000-entry display limit now only counts entries inside the visible range, so very frequent schedules that started shortly before it no longer show up short or empty.

Fixes #31600
2026-09-30 19:43:41 +04:00
Classic298andGitHub 32e532b459 fix: copy last code block shortcut copies the Artifacts pane instead of the last code block (#31613)
With the Artifacts pane open, the Copy Last Code Block shortcut put the pane's full HTML document on the clipboard instead of the last code block in the chat. The pane opens on its own whenever a reply contains an HTML block, so the shortcut was wrong in every such chat.

The shortcut clicks the last Copy button of chat code blocks on the page, and the Artifacts Copy button carried the same marker class. The pane renders after the messages, so it always won. The marker class is now removed from the Artifacts button. It had no styling attached, so the button looks and works as before.

Fixes #31476
2026-09-30 19:43:24 +04:00
Classic298andGitHub d7a74450b5 fix: question and answer vanish from the chat when a background sub-agent finishes before the answer (#31557)
When a background sub-agent finished while the model was still writing the answer that started it, the sub-agent's report was placed as a reply to the previous answer instead. Once the answer ended, the chat switched to that other branch of the conversation, which hid the latest question and answer, and the model replied to the report without seeing that question and answer. The report now follows the answer that started the sub-agent (or the newest completed reply after it), so the conversation stays on one branch.

Fixes #31507
2026-09-30 19:43:13 +04:00
Classic298andGitHub f98ca224c5 perf: use the faster JSON encoder by default (#31616)
ENABLE_ORJSON has shipped as an option since v0.11.0 (2026-07-27), five releases and two months ago, and orjson is already installed with every instance. The only two problems ever found with it (rare line break characters splitting a stream, and extra encoding options being ignored) were fixed in v0.11.1 and nothing has come up since. The regression suite at https://github.com/open-webui/tests now runs 222 tests with the option on and off side by side, on SQLite, Postgres, several workers sharing one Redis with some on and some off, and in the browser: chats, completions for every provider format, tool calls, citations, all workspace and admin data, exports and imports, notes and live socket updates behave the same, and every API response is byte for byte identical. The only differences were in how non-English text gets saved to the database, where the standard encoder is the one with bugs (missed searches and too small size limits). Turning it on by default gives every instance the speedup measured in #27583 (live socket updates encode 17x and decode 3x faster), and ENABLE_ORJSON=false keeps the old encoder.
2026-09-30 19:42:53 +04:00
G30andGitHub 3d46a59b2d fix: turn a large channel paste into a file when Paste Large Text as File is on (#31366) 2026-09-30 19:42:46 +04:00
Classic298andGitHub bff0492b5f fix: every log line is exported twice to the OpenTelemetry collector when traces and logs are both on (#31528)
With ENABLE_OTEL, ENABLE_OTEL_TRACES and ENABLE_OTEL_LOGS all on, the collector received each log line twice, once with code location attributes and once without, because the logging instrumentation for traces now attaches its own log exporter next to Open WebUI's. It now keeps trace context on log lines without adding that second exporter, so each log line reaches the collector once, and OTEL_PYTHON_LOG_AUTO_INSTRUMENTATION=false is no longer needed as a workaround.

Fixes #31524
2026-09-30 19:36:43 +04:00
Classic298andGitHub 2f6addf351 fix: new chat from the search dialog cuts off or changes text containing #, & or + (#31592)
Starting a new chat from the search dialog with "Start a new conversation" sent a different message than the one typed: everything from a # or & onwards was dropped and + turned into a space. The typed text now reaches the new chat unchanged.

Fixes #31469
2026-09-30 19:35:38 +04:00
Classic298andGitHub bee06b08ba fix: jina-colbert-v2 reranker fails to load and turns hybrid search off (#31532)
Choosing jinaai/jina-colbert-v2 as the reranking model failed on the current transformers release with "'HF_ColBERT' object has no attribute 'all_tied_weights_keys'", and saving the Documents settings quietly switched hybrid search back off. The ColBERT reranker now finishes loading, reranks search results and hybrid search stays on after saving.

Fixes #31522
2026-09-30 19:34:20 +04:00
G30andGitHub e5b57540f6 fix: use the moved chat's pinned state when a folder chat is dropped on Chats or Pinned (#31368) 2026-09-30 19:33:19 +04:00
Classic298andGitHub 710b9f1e2c fix: exact matches score as the worst result on Weaviate (#31531)
With Weaviate as the vector database, a chunk identical to the query (distance 0) was treated as having no distance and got a relevance score of 0. Perfect matches could land at the bottom of the results or fall below the relevance threshold. They now score 1 as expected.

Fixes #31527
2026-09-30 19:29:59 +04:00
Classic298andGitHub 52533c5675 refac: calendar event tools use the calendar's access check (#31537)
Editing or deleting a calendar event through the chat tools now checks access to the event's calendar the same way the calendar API does.
2026-09-30 19:29:19 +04:00
G30andGitHub 8600ab9fa7 fix: keep a table cell's line breaks inside its row when converting to Markdown (#31539) 2026-09-30 19:25:38 +04:00
Classic298andGitHub 4987711391 fix: DEFAULT_LOCALE is ignored on a user's first visit (#31551)
Since v0.11.4 a new visitor always got their browser's language even when an admin set DEFAULT_LOCALE, because the browser language was remembered as if the user had picked it, so the configured default never applied. The configured default now applies on the first visit again, as it did up to v0.11.3. A language the user picks in Settings and a ?lang= link still take precedence, and instances without DEFAULT_LOCALE keep following the browser language.

Fixes #31548
2026-09-30 19:25:16 +04:00
G30andGitHub 5338aeca89 fix: render tilde-fenced code blocks as code blocks (#31543) 2026-09-30 19:25:01 +04:00
Classic298andGitHub 288bf91f73 fix: tool prompts time out when the user's tab is on another instance (#31620)
With WEBSOCKET_MANAGER=redis and several instances or workers, an instance only subscribed to Redis once a browser tab had connected to it. Until then, when a tool or Function asked the user something (a confirmation or an input dialog) and the user's tab was connected to another instance, the user's reply never reached the tool and it waited until it timed out. Every instance now subscribes at startup, so the reply arrives whichever instance the tab is on.
2026-09-30 19:24:03 +04:00
Classic298andGitHub 6d409da9d2 refac: apply the Notes permission to live note editing (#31552)
Opening a note for live collaborative editing now follows the same Notes permission as the rest of the Notes feature.
2026-09-30 19:19:53 +04:00
Classic298andGitHub 321a24dfea fix: non-English text is missed by searches and counted six times against size limits (#31615)
With ENABLE_ORJSON off (the default), non-English letters were saved to the database as escape codes, so "Ü" was stored as \u00dc. Searches that ignore upper and lower case compare against that saved text, so they missed any match that differs only in the case of a non-English letter: filtering models by the tag "Überblick" found nothing on Postgres, and searching automations for "отчёт" missed a prompt containing "Отчёт" on SQLite and Postgres. The 100,000 character size limit for user and chat variables counted the escape codes too, so Cyrillic or Chinese variables were refused as too large (or chat variables silently came out empty in the system prompt) at about a sixth of that size. Non-English text is now saved as written, which is how it is already saved with ENABLE_ORJSON on, so nothing changes for those instances, and the limit counts real characters. Anything saved before this keeps the escape codes until it is next edited.
2026-09-30 19:19:29 +04:00
Classic298andGitHub 8b1ae1d331 fix: renaming a folder with Enter saves it twice (#31584)
Renaming a folder in the sidebar and pressing Enter sent the rename to the server twice and showed "Folder updated successfully" twice, because Enter saved the name and closing the text field saved it again. Enter now just closes the text field, which saves the new name once and shows one confirmation.

Fixes #31582
2026-09-30 19:10:48 +04:00
Classic298andGitHub 2062231f9c fix: apply the usual login check when the app loads its settings (#31621)
Loading the app's settings now uses the same login check as every other request, so a session that is no longer valid gets the logged-out settings.
2026-09-30 19:09:34 +04:00
Classic298andGitHub d3dde3609d fix: audit log records new passwords in plain text (#31622)
With request auditing turned on, the audit log only masked fields named exactly "password". The new password from a password change, and passwords entered in admin settings such as YaCy or Jupyter, were written to the log as-is. Any field whose name ends in "password", in any letter case, is now replaced with asterisks.
2026-09-30 19:09:03 +04:00
Classic298andGitHub b4ebd0d62f refac(hardening): apply the same safety checks to generated image downloads (#31623)
When an image generation backend returns a link instead of the image itself, the download now goes through the same safety checks used for other external image downloads. Links on the configured ComfyUI address are still trusted as before, so a ComfyUI server on a local network keeps working.
2026-09-30 19:08:48 +04:00
Classic298andGitHub 6a2aad92f2 fix: SSO login failures show the email/password error (#31629)
When signing in through an OAuth/OIDC provider failed, for example because the provider denied access, the account had no email or its email domain was not allowed, the login page told the user their email or password was wrong, even though they never typed one. Every such failure now shows "Sign-in with your identity provider failed. Please contact your administrator for assistance." The text is the same for every cause so it does not reveal which check failed, and the exact reason is still written to the server log as a warning.

Fixes #31627
2026-09-30 19:07:49 +04:00
Classic298andGitHub b6dfa3799d fix: chat links in finished and failed webhook notifications open a 404 (#31572)
The link in "chat finished" and "chat failed" webhook notifications was missing the `/c/` part of the chat address, so clicking it opened a 404 page. The link is sent as `/c/<chat id>` again and opens the chat.

Fixes #31565
2026-09-30 19:05:45 +04:00
Classic298andGitHub c488f60e8b fix: temporary chats save sub-agent conversations on the server (#31573)
In a temporary chat, when the model handed a task to a sub-agent, the sub-agent's conversation with the task and its answer was saved on the server, although a temporary chat should leave nothing behind. The model is no longer offered sub-agents in temporary chats.

Fixes #31567
2026-09-30 19:05:21 +04:00
Classic298andGitHub 3d70d43a1c fix: members can be added to or removed from a direct message through the API (#31575)
The person who started a direct message could add or remove people through the API, although the app only offers this in group channels. Someone added this way could read the whole earlier conversation, and because the original pair no longer matched the conversation, their next message opened a second, empty direct message. Changing the members of a direct message now answers with a 403.

Fixes #31570
2026-09-30 19:03:32 +04:00
Classic298andGitHub fa66b0f306 fix: background sub-agent and timer replies only show in the open chat after a page reload (#31576)
When a sub-agent running in the background finished, or a timer the model set went off, the result and the model's follow-up reply were saved but did not show in the chat the user had open. They only appeared after a manual page reload. They now show in the open chat as soon as they arrive.

Fixes #31566
2026-09-30 19:03:01 +04:00
Classic298andGitHub 9b45bedeaf fix: check the Channels permission when a channel automation runs (#31577)
An automation that posts into a channel now only runs when the user who created it has the Channels permission.
2026-09-30 19:02:37 +04:00
Classic298andGitHub 028dab8f1f fix: apply the Channels permission to real-time channel messages (#31578)
Real-time channel messages are now only delivered to users who have the Channels permission set in the admin user permission settings.
2026-09-30 19:02:21 +04:00
Classic298andGitHub e26da43076 fix: automation still shows "Last run Never" after "Run now" (#31583)
* fix: automation still shows "Last run Never" after "Run now"

Running an automation with "Run now" added the run to its history, but the automation page and the Automations list kept showing "Last run Never" (or the time of the last scheduled run). Only scheduled runs recorded a last run time. A manual run now records it too, so the automation page and the Automations list show the time of the run you just started.

Fixes #31580

* fix: show the new last run time right after Run now

The automation page now takes the automation the server returns after Run now, so the last run time updates on the spot and no reload is needed.
2026-09-30 19:01:42 +04:00
Classic298andGitHub 6419eca3a2 chore: bump pycrdt to 0.14.8 (#31636)
Bumps pycrdt from 0.13.1 to 0.14.8.
2026-09-30 19:00:23 +04:00
Classic298andGitHub c1f245845c fix: MCP tool calls ignore AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER (#31641)
A tool call to an MCP server had no time limit, so a server that hung kept the chat waiting until a reverse proxy or the server itself closed the connection, even with AIOHTTP_CLIENT_TIMEOUT_TOOL_SERVER set. The call now stops after the configured number of seconds and the model sees the timeout as a tool error, which is how OpenAPI tool servers already behave. When the variable is unset it falls back to AIOHTTP_CLIENT_TIMEOUT, and with neither set (or a value of 0 or below) MCP tool calls still have no time limit.

Fixes #31640
2026-09-30 19:00:20 +04:00
spoofyandGitHub 3819dc2793 i18n: ms-MY language translations (#31560)
* i18n: complete and review Malaysian Malay translations

* i18n: clarify Malay time labels and technical wording
2026-09-29 20:25:27 +04:00
Classic298andGitHub ad01bca2b0 fix: repeating calendar events are missing when they began before the visible dates (#31606)
A repeating event that was still running when the visible dates began was left out of the calendar. A weekly event from 23:00 to 01:00, for example, did not show on the following day, while the same event without a repeat did. Repeating events are now shown whenever any part of them falls within the visible dates.

Fixes #31605
2026-09-29 20:25:18 +04:00
G30andGitHub 176d31d1db fix: turn Code Interpreter off when opening another chat from the sidebar (#31500) 2026-09-28 07:59:54 +04:00
Classic298andGitHub d4b9d19645 fix: system prompt and compacted context are lost after approving a tool call (#31501)
With tool approval set to ask, the request sent to the model after approving a tool call left out the system prompt. In a compacted chat it also left out the conversation summary and sent the whole history again. The request after approval now has the same system prompt and compacted context as the one before it, plus the tool call and its result. The system prompt is picked the same way as for any other message: the chat Controls prompt, else your personal Settings prompt, else the admin default. A system prompt sent only in an API request is not kept by the server, so it is still missing after approval.

Fixes #31499
2026-09-28 07:59:43 +04:00
Classic298andGitHub b00745c8d0 fix: approved tool results are lost and approved tools can run twice (#31502)
With tool approval set to ask, the result of an approved tool call was dropped from the chat once the reply finished, so the model no longer saw it in later turns. When the model then asked for a second tool, the first call went back to waiting for approval, and approving it again ran the tool a second time. Approved results now stay in the chat and each tool runs once.

Related to #31499
2026-09-28 07:59:20 +04:00
Classic298andGitHub 8667dd7121 fix: actions in a model's menu in the model selector do nothing (#31503)
Picking Edit, Keep in Sidebar, Copy Link or Delete from the menu next to a model in the model selector did nothing: the selector closed and the action never ran. These actions work again. Edit and Delete now also close the selector, so the first click in the settings window or the delete confirmation that opens is not lost.
2026-09-28 06:16:07 +04:00
G30andGitHub c8a3e10e13 fix: let Enter through when the message box's suggestion list has nothing to show (#31506) 2026-09-28 06:15:56 +04:00
Timothy Jaeryang Baek 82b5181372 refac 2026-09-28 06:15:24 +04:00
Classic298andGitHub ef67cc3fac fix: first click in a dialog opened from a menu does nothing (#31496)
The chat header menu, the folder menu, the note menu, the knowledge Add Content and folder row menus and the Actions menus in Models, Archived Chats and Personalization stayed open behind the dialog they opened, so the first click inside that dialog only closed the menu. These menus now close when an item is picked, like the sidebar chat menu. The note menu's Formatting switch, the chat tags and submenus such as Download still keep the menu open.

Fixes #31493
2026-09-28 01:52:26 +04:00
Classic298andGitHub b6f9679985 refac: only open web, mail, phone and relative file links (#31491)
File attachment links and the file links in code execution results now go through the same link check as links in chat messages, so only web, mail, phone and relative links are opened.
2026-09-28 01:52:15 +04:00