release.2026-07-02T15-09-23Z.2097
Feature Updates
- Improved Teable Agent resource link recognition: Agent now better recognizes Teable table, app, automation, and record links.
Bug Fixes & Optimizations
-
Fixed organization permission inheritance: Members now correctly receive permissions from departments or user groups, including parent departments.
-
Fixed date formula evaluation issues:
IS_BEFORE,IS_AFTER, andIS_SAMEnow return correct booleans in logical functions. -
Improved status/date formula stability: Time-based process states are now more reliable.
-
Fixed record insertion after attachment upload: Reduced “Attachment not found” errors when inserting records after uploads.
-
Fixed inaccessible or undeletable error tables: Tables in error states can now be removed for easier recovery.
-
Improved chat response speed: Cached information is prioritized to reduce wait time.
-
Improved chat attachment handling: Files sent through guided input are now preserved, displayed, and passed into chat.
-
Fixed attachment upload blocking in queued chats: Improved upload stability in queued chat scenarios.
-
Fixed duplicate AI replies: Reduced duplicate responses after the first instruction.
-
Fixed aggregation result loading failures: Tables with mixed-case field names now load summary results more reliably.
-
Fixed saving shared Bases as copies: Users can now save copies when shared links allow “Save as copy”.
-
Improved app card display in AI conversations: Related app cards now appear below messages with context preserved.
-
Improved CLI help and documentation: Clarified scope, base parameters, skill import, GitHub skill paths, login config, and SQL table naming.
-
Improved Agent Computer startup experience: Enhanced app preview prewarming for faster, more stable app opening.
-
Unified SQL query experience: SQL query permissions are now consistent across UI and CLI.
-
Improved table content display in AI Chat: “Add to chat” now shows field names instead of column positions.
-
Improved AI Chat selection markers: Row and column selections are clearer, helping AI understand referenced ranges.