Significant-Gravitas/AutoGPT autogpt-platform-beta-v0.11.0 on GitHub

🔥 What's New?

⚠️ Breaking Changes

#16459 - Payment enforcement is now the backend setting ENABLE_PLATFORM_PAYMENT (default off) instead of the enable-platform-payment feature flag. A self-hosted install that charges must set it; FORCE_FLAG_ENABLE_PLATFORM_PAYMENT and NEXT_PUBLIC_FORCE_FLAG_ENABLE_PLATFORM_PAYMENT no longer do anything (by @Pwuts)
#16474 - The platform API answers 413 to a request body over MAX_REQUEST_BODY_MB (default 10 MiB) on any route that is not an upload route or webhook ingress. Raise it, or configure media storage, if you pass larger files inline (by @Pwuts)
#16522 - Behind a proxy chain whose X-Forwarded-For carries more than one address, list your proxies in the new TRUSTED_PROXIES setting (IP addresses or CIDR ranges). Left unset, per-visitor limits get no client IP instead of the header's leftmost entry. An entry that is not an IP address or CIDR range stops the frontend from starting (by @Pwuts)
#15853 - LaunchDarkly support is removed and every feature flag is read from PostHog. LAUNCH_DARKLY_SDK_KEY, FEATURE_FLAG_BACKEND, NEXT_PUBLIC_LAUNCHDARKLY_ENABLED, NEXT_PUBLIC_LAUNCHDARKLY_CLIENT_ID and NEXT_PUBLIC_FEATURE_FLAG_BACKEND are ignored; without PostHog every flag is its default (by @kcze)
#16026 - AutoPilot conversations are no longer traced to Langfuse, which is now used only for prompt management. AutoPilot traces go to PostHog (by @kcze)

New Features

#16310 - Add Parallel Decisions blocks (web-grounded yes/no, choice, score) (by @Torantulino)
#16326 - Show real changelog entries in the login What's new panel (by @Torantulino)
#16316 - Show popular services first in the Connect a service dialog (by @Torantulino)
#16306 - Offer Connect an integration from the @ mention list (by @Torantulino)
#16279 - Redesign the "Confirm your email address" email (by @Torantulino)
#16268 - One Microsoft connection for mail, calendar, files, Teams, Copilot and Work IQ (by @ntindle)
#15087 - Add Gemini 3.5 Flash, 3.5 Flash Lite and 3.1 Flash Lite to the LLM catalog (by @Bentlybro)
#16407 - Link feedback to the person's PostHog replay (by @kcze)
#16398 - Keep the onboarding role when its step is answered, not at the end (by @kcze)
#16527 - Ask self-hosted users for a GitHub star after five answered prompts (by @ntindle)
#14545 - Rate limit the v1 external API (by @Jia-ben00)
#16459 - Payment enforcement is the ENABLE_PLATFORM_PAYMENT setting (by @Pwuts)
#16522 - Resolve the client IP through operator-listed trusted proxies (by @Pwuts)
#16554 - Seed previews with a small synthetic marketplace (by @ntindle)
#16569 - Route load-test accounts to the dummy engine or a scripted model (by @ntindle)
#15733 - Wake Otto after opt-in background subtasks (by @ntindle)

Enhancements

#16352 - Compact spend approval card with a positive tone (by @0ubbe)
#16280 - Clarify app connections in Otto (by @0ubbe)
#16314 - Sign-up panel describes a team of AI experts, not agent building (by @Torantulino)
#16394 - Design-system guardrails as lint rules with a per-file baseline (SECRT-3097 wave 0) (by @Abhi1992002)
#13676 - Skip the truncate() probe grid when the payload already fits (by @dexhunter)
#16225 - Say "Done" at the end of the latest turn (by @kcze)
#16226 - Plain-language compaction copy and a clock on long-running steps (by @kcze)
#16227 - Turn notifications on from the chat banner in one click (by @kcze)
#16228 - Thumbs up/down work on a reply that just finished (by @kcze)
#16250 - Give a Discord channel mention the recent channel messages as context (by @Bentlybro)

UI/UX Improvements

#16401 - Design tokens, block-menu chips, LegacyButton refactor (by @Abhi1992002)
#16404 - Design tokens, Text and type scale; zero strict violations in copilot (by @Abhi1992002)
#16399 - Shadcn names alias the roles; Text and Button internals on tokens; one dark selector (by @Abhi1992002)
#16406 - Design tokens, Text, LegacyButton; zero strict violations in admin, team (by @Abhi1992002)
#16405 - Design tokens, Text, Status badge, profile sidebar, LegacyButton; zero strict violations in library, profile (by @Abhi1992002)
#16400 - Contextual, layout, renderers and auth on design tokens; zero strict violations (by @Abhi1992002)
#16403 - Design tokens, Text, type scale, search chips and sort menu; zero strict violations in settings, marketplace, artifacts (by @Abhi1992002)
#16402 - Atoms, molecules, organisms, tour and public pages on design tokens; zero strict violations (by @Abhi1992002)
#16411 - Delete the legacy layer; design-system lint is an error everywhere (SECRT-3153, DS4b, DS3a tail) (by @Abhi1992002)
#16429 - CSS-first theme on Tailwind 4 (by @Abhi1992002)
#16428 - Icons through the Icon atom (by @Abhi1992002)
#16513 - Text, Icon and Button variants instead of call-site restyles (by @Abhi1992002)
#16556 - One text-size scale; Tailwind's sizes reset (SECRT-3152) (by @Abhi1992002)
#16511 - Collapse colours to one role token set (SECRT-3154) (by @Abhi1992002)
#16256 - Show every onboarding role on a phone (by @Pwuts)

Bug Fixes

#16053 - Websocket server reads through the database manager, so sockets open and live run updates stream (by @Pwuts)
#16267 - Microsoft 365 Copilot sign-in codes send the app's secret (by @ntindle)
#16303 - Show every paused run's approval in the chat, not only the newest (by @Pwuts)
#16265 - Keep MCP tools' structured results (by @ntindle)
#16311 - Sign-in tokens whose algorithm doesn't match their key get a 401, not a 500 (by @Pwuts)
#15492 - Score memory candidates in one wave, not two (by @ntindle)
#16349 - Keep list_workspace_files in the model's tool list (by @0ubbe)
#15162 - Execute the validated graph version in run_agent (by @siye566)
#16342 - Keep a menu open when the composer mounts under it (by @Pwuts)
#16360 - Report a tool's refusal to an MCP client as a failed call (by @Pwuts)
#16351 - Keep 50-75 KB tool results reachable from bash_exec (by @0ubbe)
#15298 - Keep distinct list items with equal keys apart in dedupe/difference/intersection (by @ntindle)
#16350 - Point agents at workspace storage for uploaded and saved files (by @0ubbe)
#16369 - Keep a plain chat's sandbox across long gaps and say when it is replaced (by @0ubbe)
#16365 - Redirect bash_exec on any reference to the host tool-results folder (by @0ubbe)
#16340 - Say a public mail provider's related accounts are linked by a shared IP (by @Pwuts)
#15386 - Keep users signed in through network blips and deploys (by @ntindle)
#16410 - Resync Mistral Large 4 display price with OpenRouter (by @ntindle)
#12896 - Prevent Select controlled/uncontrolled switch warning (#11093) (by @SymbolStar)
#13234 - Handle empty jina search results (by @RykerFeng)
#13244 - Clarify invalid models and bound retry token budgets (by @RykerFeng)
#15027 - Prevent RangeError from invalid date in timezone formatting (by @sentry[bot])
#15841 - Send memory recall's reranker through OpenRouter's priority tier (by @ntindle)
#15849 - Read a new chat's first-message context while the turn sets up (by @ntindle)
#15866 - Recall a new chat's memories while its query is built (by @ntindle)
#16370 - Run file-writing tool calls in the model's order (by @0ubbe)
#16312 - Reload once when a chunk fails to load after a deploy (by @posthog-eu[bot])
#16383 - Handle a missing currentPath in the validateSession server action (by @sentry[bot])
#16396 - Delete analytics and ad cookies when consent is declined or withdrawn (by @kcze)
#16432 - Give mirrored Sentry messages an exception PostHog can group (by @kcze)
#16435 - Stop two Sentry report tests depending on delivery order (by @kcze)
#15079 - Say why a bot turn failed, and replay a codex resend after a client timeout (by @Bentlybro)
#16262 - Normalize Ollama message payloads (by @Bentlybro)
#16264 - Handle invalid PDF artifact responses (by @Bentlybro)
#15330 - Bound store review score to 1-5 (by @ntindle)
#15600 - A declined card on auto top-up says why instead of failing with a 500 (by @Bentlybro)
#16440 - Restore dynamic nested output handles (by @ntindle)
#16443 - Update expert roster tests to the shipped skills catalog (by @Bentlybro)
#16419 - Move to Better Auth 1.7.12 ahead of its Nov 3 security release (by @ntindle)
#16448 - Bill DataForSEO and Exa at the platform's 150 credits per USD (by @ntindle)
#16444 - Apply attempt limits to the login and reset-password actions (by @Pwuts)
#16466 - Chat links to app pages without a leading slash open instead of showing [blocked] (by @kcze)
#16467 - Search the Favorites tab on the server (by @kcze)
#16468 - Recognise a duplicate chat message id from Prisma's real error (by @kcze)
#15326 - Make OAuth authorization code redemption single-use under races (by @ntindle)
#15788 - Count each run once in mv_agent_run_counts (by @ntindle)
#16478 - Preserve precision when converting integer strings (builds on #16415) (by @ntindle)
#15783 - Keep /link/ pages out of the onboarding redirect (by @ntindle)
#16494 - Limit video frame size and VideoConcatBlock input count (by @Bentlybro)
#16496 - Limit Text Replace and Combine Texts output to 10M characters (by @Bentlybro)
#15083 - First-turn thread context, other people's mentions, repeat and no-op setup cards (by @Bentlybro)
#16470 - Stream virus scans with backpressure and bound Sentry's frame locals (by @Pwuts)
#16472 - Finish the REST API start-up when a data migration fails (by @Pwuts)
#16474 - Limit request body size on the platform API (by @Pwuts)
#16499 - Render workspace previews in a child process with a time and memory limit (by @Bentlybro)
#16487 - Read folder listings and run credentials once, and keep folder-write reads on their transaction (by @Pwuts)
#16476 - Read each graph link and node once, and bound run inputs (by @Bentlybro)
#16481 - Run direct block execution off the server loop and bound block work (by @Bentlybro)
#16473 - Share one Redis connection per shard per WebSocket and cap subscriptions (by @Pwuts)
#16490 - Bound graph, review and skill request handling and response compression (by @Bentlybro)
#16483 - Bound how many named inputs a preset and a run can hold, and how many presets an account holds (by @Pwuts)
#16177 - Show plan guidance for memory chat paywalls (by @posthog-eu[bot])
#16497 - Cap the bytes archive members unpack to in previews, skill uploads and spreadsheet reads (by @Bentlybro)
#16510 - Load every page of library workflows so the briefing counts the whole library (by @kcze)
#16518 - Write PostHog once-per-tab markers only when the event is sent (by @kcze)
#16520 - Name users by ID, not email, in analytics views and ops alerts (by @kcze)
#16524 - Update the swap proxy to mitmproxy 12.3.1 and stop decoding Brotli bodies (by @kcze)
#16523 - A relaunched turn reaches the model instead of dying on our own 409 (by @Pwuts)
#16424 - Two chat turns starting at once on one executor pod both get their sandbox (by @Pwuts)
#16260 - Pin the consult escaping's quote handling and input cuts (by @ntindle)
#16079 - Pass value1 through when ConditionBlock omits the false value (by @saivinaypotluri)
#15553 - Reject lossy If Input Matches comparisons (by @cindykrafft)
#16453 - Reuse credential snapshots for setup cards (by @ntindle)
#15535 - Guard desktop stream fullscreen requests (by @sentry[bot])
#16545 - Read the chat list's page and its total side by side (by @ntindle)
#16548 - Read the subscription status's independent parts side by side (by @ntindle)
#16551 - Read the visitor's session beside the listing prefetches (by @ntindle)
#15911 - Validate stored workflow input depth (by @troyan-dy)
#16145 - Preserve explicit Responses tool strictness (by @sxl-lucky7)
#16549 - Read a new run's user, settings and workspace together (by @ntindle)
#16547 - Read the marketplace, experts, library and run pages' independent queries side by side (by @ntindle)
#16546 - Read AI connections, a chat's share state and an expert's workspace side by side (by @ntindle)
#16534 - Answer a refused graph activation over the DatabaseManager at once (by @Pwuts)
#15659 - Preserve empty anyOf object inputs (by @sentry[bot])
#16552 - Close a deleted session's browser only in the process that runs it (by @ntindle)
#16550 - Register a node output's links into different nodes side by side (by @ntindle)
#15742 - Resolve a request's personal org and its membership in one read (by @ntindle)
#16417 - Keep docker-compose.yml read-only inside the container (by @ntindle)
#16462 - Stop the preferences update from changing the account's email (by @ntindle)
#16465 - Credit a paid checkout once however many fulfilments race for it (by @ntindle)
#16571 - Stop the brain-dump loop check rejecting real Indic-script and long answers (by @ntindle)
#16493 - Render templates in a worker process under memory and time limits (by @Pwuts)
#16581 - Normalise key-name separators in shared-chat redaction (by @Pwuts)
#16344 - Judge a typed brain dump by whether it answers, not by speech rules (by @Pwuts)
#16409 - Decide the workspace storage quota where each file's row commits (by @Pwuts)
#16495 - Apply size limits while reading sandbox, tool and block output (by @Pwuts)
#16557 - Answer an error raised on purpose over RPC as itself, at once (by @Pwuts)
#16577 - Bump better-auth to 1.7.16 for the Nov 3 security release (by @ntindle)
#16334 - A marketplace listing version that doesn't exist answers 404, not 500 (by @Pwuts)
#16388 - Ten X blocks accept the X connections users actually have (by @Pwuts)
#16390 - A run finishing during shutdown no longer aborts the executor's graceful stop (by @Pwuts)
#16488 - Limit schedules and executor workers per account, and refuse crons that never fire (by @Pwuts)
#15638 - Make DELETE /graphs/{id} work instead of 500ing on FK violation (by @ntindle)
#16358 - Let Platform MCP calls read the caller's own workspace files (by @Pwuts)
#16485 - Bound stored request text, and the reads that load it back (by @Pwuts)
#16570 - Harden resuming approved block calls and sandboxes (by @ntindle)
#16389 - browser_screenshot keeps every unnamed capture and says why a save failed (by @Pwuts)
#16391 - Hiring an expert yourself completes the wallet's marketplace task (by @Pwuts)
#16191 - Move workflows on retired Claude aliases to the same family's successor (by @Pwuts)

Documentation

#16020 - Changelog v0.9.0 (by @itsababseh)
#16304 - Tell v1 migrators to create a new API key, since keys can't be edited (by @Pwuts)

Dependencies & Maintenance

#16266 - Add a helper that syncs selected env keys between machines (by @ntindle)
#16296 - Bump dotenv from 17.2.3 to 18.0.8 in /autogpt_platform/frontend (by @dependabot[bot])
#16301 - Bump uuid from 11.1.1 to 14.0.3 in /autogpt_platform/frontend (by @dependabot[bot])
#16295 - Bump @sentry/nextjs from 10.27.0 to 11.8.0 in /autogpt_platform/frontend (by @dependabot[bot])
#16298 - Bump ai from 6.0.134 to 7.0.143 in /autogpt_platform/frontend (by @dependabot[bot])
#16371 - Bump the production-dependencies group across 1 directory with 8 updates (by @dependabot[bot])
#16297 - Bump lucide-react from 0.552.0 to 1.59.0 in /autogpt_platform/frontend (by @dependabot[bot])
#16283 - Bump ruff from 0.17.1 to 0.17.2 in /autogpt_platform/backend (by @dependabot[bot])
#16282 - Bump ruff from 0.17.1 to 0.17.2 in /autogpt_platform/autogpt_libs (by @dependabot[bot])
#16300 - Bump react-day-picker from 9.11.1 to 10.0.2 in /autogpt_platform/frontend (by @dependabot[bot])
#16281 - Bump node from 1681862 to 55fb5f5 in /autogpt_platform/single-container (by @dependabot[bot])
#16286 - Bump certifi from 2025.7.14 to 2026.10.18 in /autogpt_platform/autogpt_libs (by @dependabot[bot])
#16289 - Bump motion from 13.2.0 to 14.4.0 in /autogpt_platform/frontend (by @dependabot[bot])
#16291 - Bump shiki from 3.21.0 to 4.5.0 in /autogpt_platform/frontend (by @dependabot[bot])
#16000 - Bump msw from 2.11.6 to 2.15.0 in /autogpt_platform/frontend (by @dependabot[bot])
#16001 - Bump eslint-plugin-storybook from 9.1.5 to 10.6.3 in /autogpt_platform/frontend (by @dependabot[bot])
#16433 - Update posthog to 7.80.3 (by @kcze)
#15853 - Remove LaunchDarkly and read every flag from PostHog (by @kcze)
#16026 - Remove Langfuse tracing of AutoPilot conversations (by @kcze)
#14369 - Upgrade conflict label action to Node.js 24 (by @QiuLsG)
#16070 - Cache the Prisma CLI and retry its download in backend and full-stack CI (by @Pwuts)
#16418 - Run Classic's Python checks only on changes under classic/ (by @Pwuts)
#16508 - Run the Playwright browser jobs on ubuntu-24.04 (by @Pwuts)
#16423 - Start the e2e backend services after the frontend image is built (by @Pwuts)
#16477 - Deploy workflows dispatch the commit they migrated, not a branch name (by @Pwuts)
#16489 - Drop the MCP tests' stored-credential fixture (by @Pwuts)
#16544 - Guard every test against real iframe navigation (by @Pwuts)
#16463 - Count only trusted !deploy/!undeploy comments in the PR-close scan (by @ntindle)
#16232 - Keep compose FalkorDB's memory graph across container recreation (by @Pwuts)

🎉 Thanks to Our Contributors!

A huge shoutout to everyone who contributed to this release. Special thanks to our new contributors:

As well as our returning contributors:

📥 How to Get This Update

To update to this version, run:

git pull origin autogpt-platform-beta-v0.11.0

Or download it directly from the Releases page.

For a complete list of changes, see the Full Changelog.

📝 Feedback and Issues

If you encounter any issues or have suggestions, please join our Discord and let us know!

Significant-Gravitas/AutoGPT autogpt-platform-beta-v0.11.0 🚀 AutoGPT Platform v0.11.0 on GitHub