Pull Request Explorer
| Title | Author | Size | AI | Cycle Time | Review | Merged |
|---|---|---|---|---|---|---|
| fix: apply the custom model fallback to chats sent from the web UI | Classic298 | XS | No | 152.3h | - | Oct 01, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: subscribe members added to a channel to its live feed | Classic298 | XS | No | 163.2h | - | Oct 01, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: send the query embedding as a vector in external pgvector retrieval | Classic298 | XS | No | 163.2h | - | Oct 01, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: editing a knowledge file on Elasticsearch keeps its old text searchable | Classic298 | XS | No | 67.9h | - | Oct 01, 2026 |
SummaryBug FixFix stale Elasticsearch chunks after knowledge file edit Ensures updated knowledge files replace old text in Elasticsearch, improving search accuracy and preventing duplicate content. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: prompt cache misses after a model calls tools one after another | Classic298 | XS | No | 40.2h | - | Oct 01, 2026 |
SummaryBug FixFix prompt cache misses for consecutive tool calls Ensures each tool call is sent as a separate message, preserving conversation context and preventing cache mismatches. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: sub-agents in folder chats are not limited to the folder's knowledge | Classic298 | XS | No | 42.8h | - | Oct 01, 2026 |
SummaryBug FixFix sub-agents folder knowledge limitation Ensures sub‑agents inherit the folder’s knowledge base, preventing them from accessing unrelated knowledge and aligning behavior with the parent chat. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.10
Tech Stack
Languages:
Python
|
||||||
| i18n: complete and review existing Hindi translations | EllaFr | XL | AI | 4.3h | - | Oct 01, 2026 |
SummaryFeatureComplete and review Hindi translations Users can now see a fully localized Hindi interface, reducing English fallback and improving user experience. Health Assessment
Large
Low
Low
AI Details
Usage:
AI Assisted
Category:
Code AI
Confidence:
0.95
Tech Stack |
||||||
| fix: uploaded files lose Chinese punctuation and quotes | Classic298 | XS | No | 9.3h | - | Oct 01, 2026 |
SummaryBug FixFix Chinese punctuation loss in file uploads This change preserves full-width punctuation and curly quotes when extracting text from uploaded files, ensuring accurate previews and model input. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: audit log shows passwords that contain a double quote | Classic298 | S | No | 7.1h | - | Oct 01, 2026 |
SummaryBug FixFix audit log masking for passwords with quotes Ensures sensitive passwords are fully masked in audit logs, preventing accidental exposure of credentials. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: don't bring up the keyboard when menus open or close on mobile | silentoplayz | XS | AI | 3.6h | - | Sep 30, 2026 |
SummaryBug FixFix mobile keyboard pop-up on menu interactions Prevents unwanted keyboard pop-ups on mobile, improving user experience during menu interactions. Health Assessment
Small
Low
Low
AI Details
Usage:
AI Reviewed
Category:
Code AI
Confidence:
0.85
Tech Stack
Frameworks:
Svelte
|
||||||
| feat(refac): Milvus hybrid search without a second text-only copy of every collection | Classic298 | XL | No | 1.7h | - | Sep 30, 2026 |
SummaryFeatureMilvus hybrid search without duplicate collections Consolidates vectors, text, and BM25 indices into a single Milvus collection, reducing storage overhead and migration time while maintaining search quality. Health Assessment
X-Large
Medium
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: Playwright web loader returns only the site menu for pages with more than one <main> element | Classic298 | XS | No | 1.0h | - | Sep 30, 2026 |
SummaryBug FixFix Playwright loader for multi-<main> pages Ensures Playwright web loader captures full page content when multiple <main> tags exist, preventing loss of content on certain pages. This improves content extraction accuracy for pages with complex layouts. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| feat: native hybrid search for Milvus and Milvus multitenancy | Classic298 | L | No | 0.9h | - | Sep 30, 2026 |
SummaryFeatureAdd native hybrid search for Milvus multitenancy Adds hybrid BM25 and vector search to Milvus, improving relevance and scalability for multi-tenant knowledge bases while minimizing storage overhead. Health Assessment
Large
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: shared chats can't open new files attached together with a file already in the chat | Classic298 | XS | No | 0.4h | - | Sep 30, 2026 |
SummaryBug FixFix shared chat file attachment issue Ensures that new files attached to a chat are correctly recorded and visible when the chat is shared, improving collaboration and user experience. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: iPhone HEIC photos are rejected by vision models | Classic298 | S | No | 0.5h | - | Sep 30, 2026 |
SummaryBug FixFix HEIC photo rejection in vision models Ensures iPhone HEIC/HEIF photos are converted to JPEG before upload, preventing vision model errors and improving user experience. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Typescript
|
||||||
| fix: artifact preview closes right after opening when a filter writes the reply | Classic298 | XS | No | 0.3h | - | Sep 30, 2026 |
SummaryBug FixFix artifact preview auto-close on filter output This fix ensures that artifact previews remain visible when filters output partial content, preventing abrupt closure and data loss. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: artifact preview sometimes closes the moment it auto-opens | Classic298 | XS | No | 0.3h | - | Sep 30, 2026 |
SummaryBug FixFix artifact preview auto-close bug Ensures artifact preview remains open when auto-opened, improving user experience in chat. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.95
Tech Stack
Frameworks:
Svelte
|
||||||
| fix: open the highlighted chat when Enter is pressed in the search dialog | silentoplayz | XS | No | 152.0h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: channel mention notification shows raw mention markup with the user id | Classic298 | XS | No | 28.5h | - | Sep 30, 2026 |
SummaryBug FixFix raw mention markup in notifications Corrects notification display to show user-friendly mentions instead of raw markup, improving clarity for users. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Frameworks:
Svelte
|
||||||
| fix: send max_completion_tokens for Bedrock-prefixed OpenAI models | Classic298 | XS | AI | 152.3h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: edit a repeating event's series instead of moving its start to the clicked occurrence | silentoplayz | XS | No | 153.3h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: find read-only shared notes when searching the chat's note picker | silentoplayz | XS | No | 153.8h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: report invalid image Additional Parameters JSON instead of leaving Save spinning | silentoplayz | S | AI | 132.2h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: re-enable the connection dialog's Save after an invalid Headers error | silentoplayz | XS | AI | 132.6h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: clear a pending channel reply when switching to another channel | silentoplayz | XS | AI | 132.7h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| i18n: translate missing Simplified Chinese settings labels | Alex0AI | S | No | 56.9h | - | Sep 30, 2026 |
SummaryFeatureTranslate missing Simplified Chinese settings labels Adds missing Simplified Chinese labels to settings, improving localization and user experience for Chinese-speaking users. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.80
Tech Stack |
||||||
| fix: keep a cloned chat in a shared folder the user can write to | silentoplayz | XS | AI | 135.7h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: tool HTML embeds vanish when the HTML contains entities like " | Classic298 | XS | No | 127.5h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: ejecting a model ignores AIOHTTP_CLIENT_SESSION_SSL | Classic298 | XS | No | 127.4h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: typing with an input method breaks in the empty chat input while a follow-up suggestion is shown | Classic298 | M | No | 127.2h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: Google MCP connections drop about an hour after signing in | Classic298 | XS | No | 126.8h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: SCIM group members are returned with "$ref": null | Classic298 | XS | No | 56.1h | - | Sep 30, 2026 |
SummaryBug FixFix SCIM group members $ref null Ensures SCIM group responses include valid user URLs, enabling identity providers to correctly resolve group members. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: keep images pasted into a note when the note is cut and pasted or edited by a model | silentoplayz | XS | AI | 58.7h | - | Sep 30, 2026 |
SummaryBug FixFix image persistence in note editing Ensures pasted images remain in notes after editing or model modifications, preventing data loss and improving user experience. Health Assessment
Small
Low
Medium
AI Details
Category:
Code AI
Confidence:
0.95
Tech Stack
Languages:
Typescript
|
||||||
| fix: Scheduled Tasks calendar shows extra runs for automations with a run count | Classic298 | S | No | 26.7h | - | Sep 30, 2026 |
SummaryBug FixFix calendar showing extra automation runs The calendar now accurately reflects remaining automation runs, preventing confusion and ensuring users see only upcoming tasks. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.75
Tech Stack
Languages:
Python
|
||||||
| fix: copy last code block shortcut copies the Artifacts pane instead of the last code block | Classic298 | XS | No | 21.3h | - | Sep 30, 2026 |
SummaryBug FixFix copy last code block shortcut behavior Users now copy the intended code block instead of the entire Artifacts pane, improving productivity. Health Assessment
Small
Low
Low
AI DetailsTech Stack |
||||||
| fix: question and answer vanish from the chat when a background sub-agent finishes before the answer | Classic298 | XS | No | 42.8h | - | Sep 30, 2026 |
SummaryBug FixFix chat branch misplacement on sub-agent completion Ensures chat remains on correct branch after sub-agent finishes, preventing loss of recent question and answer. Improves user experience by keeping conversation flow consistent. Health Assessment
Small
Low
Medium
AI DetailsTech Stack
Languages:
Python
|
||||||
| perf: use the faster JSON encoder by default | Classic298 | XS | No | 20.2h | - | Sep 30, 2026 |
SummaryFeatureEnable faster JSON encoder by default Switching to orjson improves API response speed and socket update performance, reducing latency for all users. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.80
Tech Stack
Languages:
Python
|
||||||
| fix: turn a large channel paste into a file when Paste Large Text as File is on | silentoplayz | S | AI | 136.0h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: every log line is exported twice to the OpenTelemetry collector when traces and logs are both on | Classic298 | XS | No | 56.0h | - | Sep 30, 2026 |
SummaryBug FixFix duplicate log export to OpenTelemetry Prevents duplicate log entries, reducing noise and storage costs in monitoring. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: new chat from the search dialog cuts off or changes text containing #, & or + | Classic298 | XS | No | 28.3h | - | Sep 30, 2026 |
SummaryBug FixFix chat message truncation in search dialog Ensures messages typed in the search dialog are sent unchanged, preserving special characters. Health Assessment
Small
Low
Low
AI DetailsTech Stack |
||||||
| fix: jina-colbert-v2 reranker fails to load and turns hybrid search off | Classic298 | S | No | 55.8h | - | Sep 30, 2026 |
SummaryBug FixFix ColBERT v2 reranker load failure Ensures the ColBERT v2 reranker loads correctly and keeps hybrid search enabled, improving search reliability. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: use the moved chat's pinned state when a folder chat is dropped on Chats or Pinned | silentoplayz | XS | AI | 135.7h | - | Sep 30, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix: exact matches score as the worst result on Weaviate | Classic298 | XS | No | 55.8h | - | Sep 30, 2026 |
SummaryBug FixFix exact matches score as worst result on Weaviate Ensures perfect matches in Weaviate vector database receive maximum relevance score, preventing them from being ranked lower or filtered out. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.75
Tech Stack
Languages:
Python
|
||||||
| refac: calendar event tools use the calendar's access check | Classic298 | S | No | 54.7h | - | Sep 30, 2026 |
SummaryRefactorRefactor calendar event tool access checks Ensures calendar event editing/deleting via chat tools respects calendar permissions, aligning with API behavior. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.80
Tech Stack
Languages:
Python
|
||||||
| fix: keep a note table cell's second line inside its row in the note's Markdown | silentoplayz | XS | AI | 52.9h | - | Sep 30, 2026 |
SummaryBug FixFix table cell line breaks in Markdown Ensures table formatting is preserved in note downloads and chat, improving user experience. Health Assessment
Small
Low
Low
AI Details
Usage:
AI Reviewed
Category:
Code AI
Confidence:
0.95
Tech Stack |
||||||
| fix: DEFAULT_LOCALE is ignored on a user's first visit | Classic298 | XS | No | 49.5h | - | Sep 30, 2026 |
SummaryBug FixFix DEFAULT_LOCALE ignored on first visit Ensures admin-set default locale applies to new users, improving consistency. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Frameworks:
Svelte
|
||||||
| fix: show code blocks fenced with ~~~ as code blocks | silentoplayz | XS | AI | 52.1h | - | Sep 30, 2026 |
SummaryBug FixFix rendering of ~~~ fenced code blocks Ensures code blocks with tilde fences display correctly, improving readability and user experience. Health Assessment
Small
Low
Low
AI Details
Usage:
AI Reviewed
Category:
Code AI
Confidence:
0.85
Tech Stack
Frameworks:
Svelte
|
||||||
| fix: tool prompts time out when the user's tab is on another instance | Classic298 | XS | No | 18.3h | - | Sep 30, 2026 |
SummaryBug FixFix tool prompt timeout on multi-instance Ensures user responses reach tools regardless of instance, preventing timeouts. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| refac: apply the Notes permission to live note editing | Classic298 | S | No | 46.2h | - | Sep 30, 2026 |
SummaryRefactorApply Notes permission to live note editing Ensures live collaborative note editing respects the same permission rules as other note operations, improving security consistency. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||
| fix: non-English text is missed by searches and counted six times against size limits | Classic298 | XS | No | 20.1h | - | Sep 30, 2026 |
SummaryBug FixFix non-English text search and size limit issues Ensures non-English characters are stored and counted correctly, improving search accuracy and preventing false size limit errors for multilingual content. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Python
|
||||||