Changelog 2026-08

Hebrew Support and Multilingual Experience Optimization

Feature Updates

• Added Hebrew interface support: Community and Enterprise now localize settings, preferences, calendars, announcements, static pages, and SDK.

Bug Fixes & Improvements

• Improved right-to-left interfaces: Enhanced icons, fields, panels, forms, menus, selectors, gradients, and keyboard navigation in Arabic and Hebrew.
• Improved multilingual text editing: Input direction follows Arabic and Hebrew interfaces while content direction remains automatically detected.
• Completed Arabic translations: Added missing Arabic text across pages and controls, reducing mixed English content for fuller localization.
• Optimized self-hosted license mobile layout: Improved small-screen display for easier credential viewing and management.
• Improved relationship graph accuracy: Shows only context-relevant relationships and excludes unrelated parent trees, clarifying data associations for administrators.

Agent queue, compute, and stability optimizations

Feature Updates

• Added Agent queues: Queue or guide 30 messages across tabs and reconnects; App Builder and Agent-to-Agent show status, senders, and controls.

Bug Fixes & Improvements

• Improved Rollup filter accuracy: Fixed OR filters including unrelated linked records, ensuring expected Rollup results.
• Improved Workflow stability: Optimized snapshots and run lists at scale, reducing loading and archived-data calculation errors.
• Improved copied-table sharing settings: Copied views are unshared by default, preventing accidental exposure while preserving source-table links.
• Improved Chat reliability: Removed invalid Bot Chat records to reduce errors.
• Improved table schema update stability: Enhanced transaction failure recovery to prevent updates from getting stuck.
• Improved view access for restricted collaborators: Views remain accessible despite unauthorized filter fields, which stay hidden; operations use masked data.
• Optimized deleted-table navigation: Prevented hangs when reopening deleted tables; inaccessible tables now redirect to an available Base table.
• Improved hybrid-mode calculation accuracy: Lookup and Formula fields now refresh after large insertions, preventing blank or stale values.
• Improved bulk-write data consistency: Incomplete calculated-field updates now retry under high concurrency, improving imports and API bulk writes.
• Improved calculated field dependency accuracy: Fixed Lookup fields incorrectly appearing as final write nodes in calculation lineage graphs.
• Improved complex-formula table stability: Fixed hangs or timeouts when creating records with nested array formulas containing multiple user fields.
• Optimized Workflow trigger table selector: New tables now appear faster in Workflow trigger table lists.
• Improved bulk update performance: Skips unnecessary recalculation when Number fields have no dependent calculated fields.
• Improved permission-based view operations: Fixed blank grids and validation errors; group titles no longer expose restricted data.

Content Creation, Table Import, and Optimization

Feature Updates

• New Artifacts: Create, preview, revise, version, protect, and share HTML or Markdown from chats with short links and revocable permissions.
• Native Google Sheets import: Securely import public or private worksheets via onboarding, AI chat, or CLI, with type detection, progress, and failure reports.

Bug Fixes & Improvements

• Improved member selector: Shared forms, kanban, plugins, and editable links can search members with access through nested or parent departments.
• Improved computed field and bulk update stability: Large recalculations recover safely after interruptions, reducing operational impact and resuming incomplete updates.
• Improved calculation progress and errors: Restored progress accurately and prevented misleading “Internal server error” messages during Airtable imports or recalculation.
• Fixed SWITCH formula errors: SWITCH now returns mixed types, including strings, numbers, booleans, datetimes, link fields, and lookup fields.
• Improved Base duplication reliability: Fixed naming conflicts and improved consistency when duplicating or deleting records and linked data.
• Improved view loading and sorting: Fixed personal view loading and stabilized sorting in masked views for smoother browsing.
• Improved record history: Deleted linked records now show “Record deleted” instead of outdated titles, clarifying historical changes.
• Improved computed field dependency view: Lookup fields are no longer omitted, providing a more complete picture of calculation dependencies.
• Improved share-to-earn points claims: Strengthened eligibility and weekly quota checks, with clearer failure feedback.
• Optimized Google Sheets imports: Improved narrow-screen and multilingual usability; handled blanks, problematic headers, empty sheets, authorization issues, and partial imports.
• Improved CLI documentation and validation: Clarified aliases, scopes, paths, deletion, skills, Agent Computer availability, providers, and configuration errors.

Mobile experience, computing, and stability optimizations

Bug Fixes & Improvements

• Optimized calculation lineage view: Shows only fields affecting downstream updates and clarifies status when no tasks are pending.
• Improved calculated field update stability: Makes formulas using numeric columns and text operations more reliable and accurate.
• Fixed lost field settings: Field updates and conversions now preserve options, including Markdown display settings for Long Text fields.
• Improved mobile plan purchasing: Fixed incomplete scrolling in plan dialogs on small screens, enabling full plan access and checkout.
• Improved deleted-table consistency: Deleted tables are now reliably logged and immediately searchable for quick status confirmation.
• Optimized mobile view toolbar: View options now use touch-friendly bottom drawers, with improved narrow-screen selectors and filters.
• Improved locked-view feedback: Clearer touch prompts explain why unavailable toolbar actions are restricted.
• Optimized onboarding: Clarified exit and skip flows, highlighted “I’ll write it myself,” refined recommendations, and updated all translations.
• Streamlined post-onboarding flow: Blank Bases open full-screen AI chat; existing Bases open their first table or mobile chat drawer.
• Improved mobile onboarding stability: Prevented blank resource pages and interrupted chats after AI-generated content redirects.
• Fixed linked record responses: Many-to-one links now return a single linked record instead of an array.
• Improved conditional lookup accuracy: Corrected field reference direction parsing for lookups within the current table.
• Optimized archived data loading: Stops unnecessary scans after enough results and prevents timeouts on large archive pages.
• Fixed linked lookups in shared views: Lookup fields now correctly identify tables referenced by linked record fields, preventing load failures.
• Improved grouped views for restricted users: Loads accessible records and skips or degrades grouping when fields are unreadable or hidden.
• Optimized mobile plans and pricing: Improved tabs, plan cards, billing details, and side menus for easier plan and billing management.
• Improved multilingual labels: Fixed localization for sorting, grouping, and search labels, with correct singular and plural forms.
• Optimized calculated field lineage graph: Corrected dependencies and connections, with automatic layouts for clearer field relationships.
• Improved query recommendations and Deep Analysis: Shows IDs beside Space, Base, and table names, and includes saved table views.
• Fixed linked record field filtering: Keeps selected records readable and supports switching from “is” to “contains” without filtering errors.
• Improved table deletion and recovery: Redirects collaborators to valid pages and provides recovery options, preventing broken links and repeated errors.

CLI, Import, and Stability Updates

Feature Updates

• Added CLI import location selection: Use --folder-id to import data into a target folder for flexible automation and archiving.
• Added CLI device-code login: Securely authenticate with teable auth login --device-code in SSH, containers, cloud IDEs, CI, private deployments, and custom domains.
• Added persistent Agent Computer configurations: Preserve, list, restore, update, and remove tool configurations and runtime states across environment rebuilds.

• Added folder imports: Import CSV, Excel, and Airtable data directly into a selected folder from its Add menu.

Fixes & Improvements

• Improved import location stability: Imports now reliably target specified folders when folders and tables update simultaneously.
• Improved large Excel imports: Added real-time row progress, fewer timeouts, persistent results, and per-sheet row-limit summaries.

• Improved record links: Links copied from Kanban, Gallery, and Calendar now open correctly, including shared-view and linked-record scenarios.
• Improved Lookup and Rollup field reliability: Fixed linked-record-title “Contains” filters, stale conversion metadata, and failures from missing or inconsistent dependencies.
• Improved AI image-generation compatibility: Prevented aspect ratios from producing resolutions unsupported by GPT Image 2, increasing success rates.
• Improved record deletion reliability: Records can still be deleted when undo history cannot be fully saved.
• Optimized record-copy notifications: Copying records with assigned users no longer triggers incorrect new-assignment notifications.
• Improved post-deletion view states: Automatically switch to another view, or show a clear empty state when none remain.
• Improved calculated field update compatibility: Fixed table-update failures when recalculating formulas combining date or time values with text.
• Improved multi-field search accuracy: Results now follow active view filters, with consistent pagination, grouping, and summaries.
• Fixed tracking-field sorting: “Last modified time” and “Last modified by” now sort by their displayed tracking values.
• Improved calculation task stability: Strengthened Worker recovery and retries; Airtable imports now wait for table configuration before calculations resume.
• Improved Table API compatibility and permissions: Optimized copy/paste, export, calendar, search, comments, history, archiving, undo/redo, trash, and related operations.

Teable Now Supports Arabic, Italian, and Spanish

More languages for more teams worldwide.

Multilingual, performance, and stability upgrades

Feature Updates

• Added Arabic support: Added Arabic UI translations; the interface currently remains left-to-right.
• Added Spanish and Italian support: Enterprise Edition now includes complete translations in both languages.

Fixes & Optimizations

• Improved multilingual experience: Refined all translations, reducing mixed-language content and English fallbacks for greater consistency and readability.
• Improved localization and messaging: Refined plurals, validation messages, dates, calendars, relative times, and errors for greater accuracy and local relevance.
• Improved grouped view performance: Accelerated loading and operations when grouping large tables by low-cardinality fields for smoother browsing.
• Improved link field pasting: Fixed blocked text pasting while the link picker is open.
• Improved record list, grouping, and aggregation sorting: Accelerated numeric, date, user, and formatted-date fields without changing order or null placement.
• Fixed Cuppy permissions in permission-matrix databases: Collaborators can use granted CRUD permissions, while unauthorized actions remain blocked with clearer errors.
• Improved post-onboarding page loading: Added a skeleton screen to reduce perceived waiting.
• Fixed mobile Airtable import database selection: Users can scroll normally to browse and select the complete database list.
• Improved post-signup routing: Template and AI generation destinations continue correctly instead of mistakenly entering onboarding.
• Improved onboarding continuity: Prompts and attachments persist through refreshes and signup redirects, preventing duplicate chats and enabling seamless continuation.
• Improved new-user Base generation: Free users retain prompts and attachments; failures show a blank space instead of indefinite loading.
• Improved field search: Results, counts, pagination, sorting, grouping, and aggregation respect filters and permissions, refreshing after changes.
• Improved AI Agent output stability: Streaming recovers from network or gateway interruptions without duplicated or corrupted text.
• Strengthened AI Agent credential protection: User-provided task credentials are more secure and immediately usable after submission.
• Improved app and Base links: Generated links use better formatting, reducing invalid or inaccessible links.
• Improved SQL query API: Standardized successful-status responses and added compressed responses for large results, improving reliability and efficiency.
• Fixed attachment downloads: Safari preserves original files, and compatible object storage services return correct file types.
• Improved Grid View stability: Fixed crashes when rapidly modifying fields or inserting fields from the column menu.
• Improved linked-table deletion reliability: Deletion completes when display columns are unavailable, reducing schema-update failures and manual recovery.
• Improved large-table performance: Accelerated record-list reading, filtering, and sorting for smoother large-scale data handling.
• Clarified “Copy records” permission: Distinguishes copying selected data to the clipboard from creating record duplicates.
• Improved Ask User stability: Submissions immediately end waiting, timeout handling is consistent, and malformed formatting no longer crashes pages.
• Improved linked-record reliability: Fixed deletion in copied or imported Bases, with cleanup and messaging for cross-Base, trashed-table, and deleted-field links.
• Improved Excel import compatibility: Supports title banners, leading blank rows, non-first-row headers, and duplicate or similar column names.
• Fixed date grouping display: Records group correctly when field and server time zones differ, preventing misplacement or hiding.

Calculations, Relationships, and Performance Optimization

Bug Fixes and Experience Improvements

• Improve Formula and Lookup field update stability: Ensure reliable backfilling despite storage type mismatches, reducing failures and stalls.
• Optimize mobile Space page layout: Keep navigation clear and usable when the header contains extensive information.
• Improve desktop Base access: Make the Base entry more prominent for faster access.
• Improve numeric Lookup field update stability: Prevent invalid values from causing update or conversion failures.
• Optimize Lookup field configuration updates: Improve stability when adjusting display settings, reducing unnecessary failures.
• Optimize Lookup and Rollup performance for large tables: Accelerate conditional processing, reducing timeouts and update failures.
• Fix numeric Lookup field display errors: Apply default number formatting when absent, preventing page crashes.
• Improve linked table update stability: Skip trashed referenced tables while continuing updates for other active tables.
• Fix usage period display: Correct billing period and timezone handling across usage dialogs, charts, and history for all plans.
• Optimize App Builder preview guidance: Reopen chat after selection and guide users to accessible Preview instead of unavailable local URLs.
• Improve Excel import stability: Handle duplicate or similar column names to prevent table creation failures in complex files.
• Fix Link field recalculation: Required single-record Link fields recalculate when records are empty but display values remain, refreshing dependent fields.
• Improve browser translation stability: Reduce errors from rewritten page content and stabilize onboarding in unsupported-language regions.
• Optimize table operations with many fields: Speed up record creation, duplication, and form submission for smoother data entry.
• Improve table creation recovery: Reliably resume creation when retrying after an invalid state, reducing repeated failures.
• Fix app publication status syncing: After unpublishing, promptly update the sidebar badge and share icon across collaborators and page refreshes.
• Fix automation percentage copying: Direct percentage field mapping now preserves original values instead of reducing them 100-fold.
• Optimize bulk deletion and redo performance: Accelerate large operations while preserving Recycle Bin restoration and clearing capabilities.
• Optimize Formula field loading and calculation: Reduce unnecessary requests and improve progress visibility and stability during large cascading calculations.
• Improve invalid Formula value handling: Avoid pointless retries and provide clearer failure messages for easier troubleshooting.
• Fix shared app saving: Saved apps appear immediately; duplicates no longer fail; copied folders are ordered last with conflict-free names.
• Improve real-time table loading errors: Show centered, localized notifications with clearer explanations for missing tables or restricted access.

Onboarding, performance, and stability improvements

Feature Updates

• Added missing options when creating linked records: Add missing options directly without leaving the current workflow.

• Added first-time guidance for new Cloud users: Step-by-step profile setup, data imports, AI-recommended automations and apps, and team invitations.

Bug Fixes & Improvements

• Improved mobile record comments: Comments now fill the drawer, respect safe areas, and hide record navigation when opened.
• Improved Markdown readability: Refined spacing, list styles, and heading hierarchy for clearer structure and easier reading.
• Improved field schema safety: Protected system-managed columns can no longer be deleted, reducing accidental data loss.
• Fixed bidirectional one-to-one Link deletion: Removing Link fields no longer damages shared table schemas or record access, keeping data stable.
• Improved multi-row pasting: Reliably paste over 17 rows; stops if target records are missing or inaccessible, preventing incorrect writes.
• Improved field type conversion performance: Accelerated conversions that preserve values, such as converting Single Select fields to Text.
• Improved invalid record link handling: Removes invalid parameters and loads tables when records are invalid or inaccessible, preventing server errors.
• Fixed App Builder task statuses: Completed generation tasks now update correctly instead of remaining “Running.”
• Improved cross-base Lookup and Rollup performance: Optimized conditional calculations and refreshes results promptly after source record changes.
• Improved App Builder GitHub integration: Added repository validation, clearer errors, sensitive-file protection, and automatic cleanup of stale merged branches.
• Fixed date sorting in grouped views: Dates spanning multiple years now sort correctly, ensuring accurate grouping.
• Improved calculated field update reliability: Preserved value types for Lookup, Formula, Link, and Rollup during schema and data changes.
• Improved inline table creation from records: Required fields are validated before saving, preventing unusable tables from incomplete records.
• Improved field deletion in large tables: Faster deletion retains undo support for smoother high-volume operations.
• Improved wide-table view duplication: Reduced unnecessary data processing to accelerate copying views with many fields.
• Fixed duplicate-value messages for unique fields: APIs now return the conflicting field and localized guidance for faster resolution.
• Improved automation date condition stability: Fixed crashes and made date filters more reliable in triggers and condition editors.
• Improved sign-up and login experience: New users land on sign-up, with clearer switching between account creation and login.

• Improved large-table duplication performance: Duplication is approximately 48% faster for tables with 50,000 records and 20 fields.
• Improved performance for bases with many fields: Accelerated form submissions and record duplication for smoother operations.
• Optimized calculated and Rollup field creation and duplication: Reduced processing delays in large tables for faster configuration and copying.
• Fixed column configuration in collapsed groups: Double-click or right-click columns when all groups in a grouped grid view are collapsed.
• Improved Lookup field reliability: Fixed update triggers for legacy Lookup fields and streamlined conversion to basic fields for stable updates.

Build AI Features in App Builder

Apps built with App Builder can now use AI directly, without requiring your own AI API key.

1. Add AI Directly to Your App

2. No AI API Key Setup

Shared Skills and Performance Upgrades

Feature Updates

• Added shared Skills management: Install Skills in a Space, Base, or personal account, with configurable access permissions.
• Added asynchronous compute task processing: Enable on demand to maintain stability during task surges or bulk writes.
• Upgraded context usage display: A ring indicator below the input shows usage percentage and progress when clicked.

Fixes & Improvements

• Improved high-write-volume table stability: Enhanced bulk-write handling to reduce timeouts and service outages during peak loads.
• Fixed context compression status display: The divider persists after refresh, usage updates promptly, and compression failures display accurately.
• Optimized mobile top bar layout: Resized and repositioned the top bar and chat button to prevent overlap with the statistics bar.
• Fixed mobile dark mode display: Corrected the chat button color for a clearer, more consistent interface.
• Improved shared Skills experience: Accelerated discovery and enhanced runtime isolation between Skills for greater efficiency and stability.
• Optimized field change performance: Faster field conversions and deletions reduce wait times when restructuring complex tables.
• Optimized computed field updates: Reduced duplicate calculations during cascading updates, improving performance for tables with extensive linked data.
• Improved multi-field text search: Accelerated searches across multiple fields while maintaining accurate filters and matches.

Multi-app generation, bulk downloads, and stability improvements

Feature Updates

• Tab selection added: Composer shows hints and accepts highlighted Skills or Slash Commands with Tab, preserving focus and mobile selection.
• Multi-app generation progress added: Chat now shows separate progress for each app generated from one prompt.
• Batch file downloads by chat added: Displayed files are grouped by chat and can be downloaded in batches.

Bug Fixes & Optimizations

• Calculations stabilized: Large cells no longer block unrelated Formula, Lookup, or Rollup updates; Compute Activity retains errors until successful refresh.
• Record creation near table limits optimized: Safe writes are preserved, while operations that cannot complete safely are rejected.
• Date-time field backfill fixed: Schema updates no longer fail when calculated backfill values involve date-time fields.
• Record writing and related calculation performance optimized: Accelerated inserts, Upserts, and cascading Lookup or Formula calculations in highly linked tables.
• Personal access token permissions strengthened: Scoped tokens now strictly enforce configured action and resource permissions, preventing unauthorized access.
• Calculation field processing reliability improved: Improved frequent updates and stale dependencies, reducing failures and repeated retries.
• Search results optimized: Private or functional pages such as login, settings, developer, and Base appear less often.
• Question card option display fixed: Fixed mismatches between highlighted and submitted options, ensuring the final choice displays correctly.
• Attachment and avatar loading stability improved: Improved caching for public images, private previews, and updated avatars, ensuring more reliable loading.
• CSV import failure messages optimized: Failures now show the original cause instead of remaining in progress or displaying unrelated errors.
• Table opening and statistics fixed: Resolved inaccessible tables and restored statistics and aggregation features.
• AI text and image automation errors improved: More accurate feedback now covers input, capability, attachment, response, rate-limit, and timeout failures.
• Linked record write validation strengthened: Required links are validated during writes, and display values remain intact when titles refresh.
• Record copying fixed: Records containing linked or Lookup fields with date-time comparisons can now be copied successfully.

• Lookup and calculated field updates stabilized: Large Lookup groups and broad cascades no longer trigger server errors during bulk updates.

Enhanced performance, real-time responsiveness, and stability

Feature Updates

• Set minimum scheduled task interval: New or updated automations and email polling require 10-minute intervals; existing schedules remain unchanged.

Bug Fixes & Improvements

• Improve bulk selection performance: Deleting, clearing, copying, and pasting tens of thousands of selected records is now faster and smoother.
• Reduce duplicate member notifications: Merge duplicates across undo, redo, bulk moves, imports, and table copies while preserving valid assignment alerts.
• Optimize field type conversion: Convert Single Select fields to Text faster in large tables, with undo and redo support.
• Accelerate view duplication in large tables: Improved performance reduces waiting time.
• Improve view stability with field permissions: Ignore inaccessible fields in saved filters or sorts, preventing errors while enforcing permissions.
• Fix statistics after archiving: Statistics, counts, and rollups now update promptly, with more reliable archiving, undo, and redo.
• Improve calculated field cascading: More reliably generate chained lookup values while improving overall calculation performance.
• Optimize member field operations: Copying, pasting, or filling clears unmatched former base or space collaborators instead of failing.
• Improve record search: Faster searches and more accurate highlighting align with records and fields on the current page.
• Fix email trigger configuration: Settings now save and update correctly when input values are cleared.
• Improve real-time view updates: View creation, duplication, restoration, updates, deletion, filtering, grouping, and sorting now appear without refreshing.
• Stabilize operations during table loading: Field edits, AI assistance, and schema changes avoid repeated connection and “Table not found” errors.
• Fix occasional table creation freezes: Tables with initial records now create reliably without blocking later imports or table creation.

Brand customization and stability improvements

Feature Updates

• App Builder branding: Published apps support custom favicons, titles, and descriptions for tabs, metadata, and previews; default: Teable Build favicon.
• Extended Automation Script runtime: Nodes can now run up to 180 seconds, supporting more complex, time-consuming automation tasks.

Bug Fixes & Improvements

• Improved Grid grouping and sorting reliability: Changes sync instantly across view settings, pages, and collaborators, including rapid consecutive edits.
• Improved sidebar after AI table creation failures: Inaccessible entries are removed on refresh, preventing confusion and disruption.
• Improved Grid linked-record labels: Truncated labels no longer overflow adjacent cells, and ellipsis styling is now consistent.
• Fixed records grouped by user: Eliminated duplicate “Add record” rows, renumbering, and misplacement; collaborator and multi-user fields now group consistently.
• Improved user-name display: Last Modified By and Editor show names instead of IDs across record view, API, and collaborator fields.
• Improved AI Chat reconnection: Responses recover more reliably after brief network interruptions, reducing unexpected streaming failures.
• Improved multi-level Lookup updates: Chained Lookup fields now synchronize latest values accurately during concurrent linked-record updates.
• Improved user field parsing and validation: Recognizes restricted collaborators, phone numbers, and Enterprise department members; rejects unrelated or deleted users.
• Improved calculation performance under high load: Better handles busy databases, reducing slow or stalled writes during large linked-record updates.
• Fixed AI Chat duplicates: Prevents duplicate assistant rows when sending after stopping; message state remains consistent during resending and recovery.

Announcing Data Archiving: Unlimited Rows, Now Possible

Don’t need some records right now, but don’t want to delete them? Archive them. They remain available and no longer count toward your Row Limit.

1. Archive Records Instead of Deleting Them

2. View Archived Records Anytime

3. Archived Records Don’t Count Toward Your Row Limit

Base Migration and Stability Improvements

Bug Fixes & Improvements

• Expanded Base import support: Import Bases without tables and with only Webhooks, enabling more automation migration scenarios.
• Improved Base import/export reliability: Preserve field descriptions and AI configurations during export and re-import, reducing repeated setup after migration.
• Improved record creation stability: Optimized workspace row counting and refresh, reducing delays/failures; unlinked tables no longer consume row quotas.
• Fixed calculation status updates: Errors now update after permanent Base/table deletion or storage migration, keeping administrators’ attention counts accurate.
• Improved high-load write stability: Optimized writes during busy periods, preventing long waits or stalls and smoothing record creation and updates.
• Optimized Lookup and cascading updates: Improved multi-level Lookup and external-link query/update efficiency, accelerating complex relational data processing.

Computational Stability and Performance Optimization

Fixes & Improvements

• Improved task stability in read-only mode: Pauses calculations during write conflicts and automatically resumes after database recovery.
• Fixed computed field updates in migrated Bases: Legacy records now recalculate correctly after imports or migrations.
• Optimized record operation performance: Improved query and processing efficiency for smoother large-scale record operations.
• Improved Outbox failure statistics: Fixed inaccurate counts for missing tasks while preserving failure history for troubleshooting.
• Improved field operation stability: Optimized conflict handling during creation, duplication, and deletion to prevent calculation interruptions.
• Improved computed field stability: Deleted tables no longer cause repeated failures, while restored tables automatically resume updates.
• Optimized computed Lookup performance: Accelerated backfills on large tables, reduced timeouts, and enabled resuming interrupted updates.
• Optimized computed and lookup field updates: Accelerated small-scale and multi-link changes while maintaining stability for large data tasks.
• Improved computed field processing stability: Optimized calculations during database load, reducing unnecessary retries and increasing update reliability.
• Optimized record operation performance: Accelerated reading, creation, and deletion, especially for large tables and bulk operations.

Failure Record Management and Stability Improvement

Feature Updates

• Added recovered delivery failure management: Hides recovered failures by default, with counts and a visibility toggle for easier troubleshooting.

Bug Fixes & Improvements

• Improved conditional rollup editor stability: Fixed crashes after hard refresh when calculating date field minimums or maximums.
• Optimized conditional rollup loading and format compatibility: Prevented errors while link fields load or saved settings mismatch.
• Improved link field cleanup reliability: Optimized cleanup of invalid link fields and missing database objects, reducing deletion failures and inconsistencies.
• Improved SMTP error messages: Shows clearer causes and provider rejection details after delivery failures, accelerating email troubleshooting.
• Improved generation task stability: Prevented tasks from stalling when linked bases are missing or deleted.
• Fixed image loading on shared pages: Cover images, logos, and plugin logos now load correctly from absolute URLs.
• Optimized mobile login experience: Supports device safe areas and keeps login/registration switching controls visible and accessible.

• Improved table filter stability: Fixed filters not saving or applying in new views or views without existing conditions.
• Optimized calculated field refresh failure handling: Reduces redundant retries and reports failures sooner for faster troubleshooting.
• Fixed incorrect prompts when editing Long Text fields: Prevented batch confirmation dialogs without AI use when editing or toggling Markdown.

Brand Customization and Performance Optimization

Feature Updates

• Expanded calculation activity access: View details via Teable API Keys or personal access tokens, respecting existing space and table permissions.

• Custom app branding: Teable Build and App Builder support custom favicons up to 5 MB, browser titles, and page descriptions.

Bug Fixes & Improvements

• Improved large-table read performance: Faster sorting, grouping, searching, and paginated views, with accurate pagination for smoother, reliable browsing.
• Improved legacy column compatibility: Views with legacy column metadata now open normally without manual cleanup.
• Fixed number field creation: Number fields now create successfully and display correctly in grid view, even with incomplete field information.
• Fixed filter view notifications: Prevented false alerts when opening table filter views containing Lookup fields referencing Link fields.
• Improved automation reliability: Dependency changes from Lookup, Rollup, or Link fields trigger automations reliably; same-table calculation changes still do not.

• Improved calculation field stability: Fixed update interruptions and repeated calculation failures involving conditional Rollup, Lookup, and field-reference filter configurations.

Deployment, AI, and Stability Improvements

Bug Fixes & Improvements

• Simplified self-hosted deployment: Docker starts without preconfigured keys while existing encrypted data remains accessible. Production deployments should use dedicated keys.

• Improved Pro upgrade guidance: All roles see upgrade options; non-owners selecting “Upgrade to Pro” are prompted to contact space owners.
• Improved view filter stability: Incomplete filter conditions no longer disappear during setup, making filter editing more reliable.
• Fixed linked record search in shared views: Searches follow Link field settings and table permissions for accurate, authorized results.
• Improved unavailable table handling: Better handling of unavailable related tables reduces errors and improves reliability.
• Improved conditional update performance: More efficient processing makes related operations respond faster.

• Unified calculated field statuses: Single-field and field-list APIs consistently return pending statuses for Formula, Lookup, Rollup, and other calculated fields.
• Improved conditional Lookup recovery: Restore fields from Trash with matching conditions and resolved results intact, preventing configuration or data loss.
• Improved AI Chat stability: Queued messages resume after stopped responses; better temporary-failure and busy-state handling reduces interruptions.
• Improved usage limits: APIs, bulk editing, imports, exports, and streaming now share localized notices and accurately enforce row quotas.
• Faster AI Chat image uploads: Attachments complete independently, so slow or failed uploads no longer block others or subsequent conversations.
• Improved cache configuration detection: Configuring a Redis URI automatically enables Redis; select other services through the cache provider configuration.
• Fixed incomplete linked record display: The linked record picker shows more available records, improving accuracy especially for cross-Base links.
• Improved conditional Lookup and Rollup compatibility: Unsupported field-reference comparisons no longer disrupt reading or updating unrelated records.
• Fixed empty table duplication: Tables without records can now be duplicated reliably for quick structure reuse.
• Improved saved view loading: Fixed occasional “Socket Error” failures when loading records in saved views with scalar Lookup filters.

Announcements, Archiving, and Performance Upgrades

Feature Updates

• Targeted in-app announcements: Schedule, localize, and withdraw multilingual announcements for specific users via banners, Toasts, modals, or sidebar cards.
• Recycle Bin snapshots: View complete deleted-record snapshots and filter by resource type, operator, deletion time, creator, or creation time.
• Business record archiving: Browse, restore, permanently delete, or export archived records to CSV, with longer retention and lower active-table load.
• Improved published app management: Adds prominent quick access, clearer unpublished-change alerts, and better version history and rollback.

Bug Fixes & Improvements

• More stable table filtering: Fixes errors when deleting the last filter, clearing filters, or opening views with incomplete date filters.
• Improved filter synchronization: Toolbar summaries update immediately, and filter changes sync more reliably across connected clients.
• Improved AI generation cancellation: Stopping chat, App Builder, or Agent generation is handled normally, preventing errors in other windows.
• More reliable checkpoint generation: Improves checkpoint reliability under high load, reducing state-save failures on busy databases.
• Faster large-table calculations: Improves calculated field processing in highly linked tables, reducing write stalls, memory pressure, and 503 errors.
• More accurate calculated fields: Fixes dependency propagation across comparisons, linked records, Lookup, Rollup, circular dependencies, and partial recalculations.
• Faster Lookup updates: Shortens Lookup propagation after single-record updates, displaying linked data sooner.
• Optimized conditional Rollup calculations: Avoids unnecessary large-scale recalculations in complex dependencies, improving responsiveness and reducing resource usage.
• More stable field type conversion: Fixes prolonged freezes when changing field types in tables containing calculated fields.
• Consistent time handling: Fixes time offsets in scheduled tasks, metadata updates, and database workflows across non-UTC deployments.
• Improved AI workflow creation: Places generated nodes near source nodes and lets creation commands specify folders.
• Stronger sharing permission controls: Sharing controls honor user and access token permissions, disabling unavailable options with explanations.
• Improved field creation auditing: Field creation logs now include field names, types, and options for clearer schema auditing.
• Fixed orphaned linked field deletion: Linked fields can now be deleted after their referenced tables are removed.
• Broader core data compatibility: Improves formulas, Lookup, Rollup, filtering, sorting, grouping, search, aggregation, link validation, field lifecycles, and Base duplication.
• Consistent null writes: Cleared text, checkbox, attachment, user, link, and multi-value fields save as null, with reliable required-field validation.
• Improved linked-record writes: Link updates accept single-link and multi-link inputs, supporting integrations, imports, and cardinality changes.
• Stricter date validation: Strict writes reject invalid dates; conversions save them as null instead of silently changing them.
• Improved date field conversion: Text-to-date and formula-to-date conversions skip invalid dates, preventing isolated errors from blocking entire conversions.
• Fixed Rating field conversion: Prevents invalid stray values and standardizes decimal rounding for consistent ratings.
• Stronger record and field permissions: Enforces table, row, and field permissions for record operations, including masked fields and disabled grants.
• Fixed user avatar compatibility: Updating records for users without custom avatars no longer fails; initial-based default avatars remain available.
• Fixed grouped view loading: Adds required grouping field data, preventing some grouped views from failing to open.
• Improved view stability: Improves reliability for creation, updates, sharing, imports, duplication, real-time updates, plugin views, and concurrent editing.
• Improved self-hosted registration: New users reliably join eligible auto-join Spaces as Viewers, including Spaces with SSO enrollment.
• Improved unsubscribe page usability: Pages remain usable after preferences update successfully, even if a later refresh fails.
• Fixed duplicate automation runs: Prevents repeated action chains and duplicate messages, Webhooks, or emails.
• Improved Formula automation triggers: Source-field changes updating Formula results now trigger automations correctly without duplicates or loops.
• More reliable bulk deletion undo and redo: Reliably restores or repeats large deletions, reducing recovery errors and unexpected timeouts.
• Optimized large imports and table duplication: Reduces unnecessary initial history and database load without affecting routine edit history.
• Improved mobile attachment previews: Widens spreadsheet viewing, improves horizontal scrolling, prevents navigation overlap, and correctly fits rotated images.
• Faster Base loading and navigation: Reduces duplicate requests and safely opens previously visited, pinned, or eligible default Bases.
• Fixed invalid Base links: Invalid direct links return to the standard entry flow, avoiding redirect loops and unnecessary errors.
• Improved field validation messages: Improves localized required and unique-field errors, identifying duplicate fields in bulk and selection-based operations.
• Stronger secret management: Improves secret handling and secure credential rotation, reducing risks from long-lived credentials.
• Improved real-time grid loading and updates: Reduces synchronization bandwidth and improves loading and responsiveness on slower networks.

App Builder Now Supports Visual Editing

App Builder now lets you refine your app directly in the preview, making small changes faster and helping reduce Credits used during optimization.

1. Edit Page Text Directly

2. Select an Element and Tell AI What to Change