Share link + scheduling + security upgrade
Release history
AutoGPT releases
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
All releases
22 shown
Security upgrade + bug fixes + new features
- #12958 - Admin CSV exports — credit transactions + copilot weekly usage (by @majdyz)
- #12978 - Connect onboarding SubscriptionStep to Stripe Checkout (by @0ubbe)
- #13002 - Dynamic max_budget_usd for SDK; parity budget hint for baseline (by @majdyz)
Full changelog
🚀 Release autogpt-platform-beta-v0.6.59
Date: May 2026
🔥 What's New?
New Features
- #12942 - Settings v2 billing page (subscription + automation credits) (by @Abhi1992002)
- #12780 - Tier-based workspace file storage limits (by @ntindle)
- #12958 - Admin CSV exports — credit transactions + copilot weekly usage (by @majdyz)
- #12978 - Connect onboarding SubscriptionStep to Stripe Checkout (by @0ubbe)
- #12985 - Vary daily-limit popup CTA by plan (by @0ubbe)
- #12995 - Adopt mentioned Discord threads in copilot (by @Bentlybro)
- #13002 - Dynamic max_budget_usd for SDK; parity budget hint for baseline (by @majdyz)
- #13004 - Yearly Stripe billing — pricing + frontend hookup (by @majdyz)
- #12624 - Frontend platform linking page (
/link/{token}) (by @Bentlybro) - #13009 - Persist activity-status LLM cost to PlatformCostLog (by @majdyz)
- #13007 - Add empty state with build / marketplace CTAs to library (by @Abhi1992002)
- #13016 - Refresh wallet checklist walkthrough videos (by @kcze)
- #13000 - Add
get_platform_infotool for tier-aware AutoPilot (by @ntindle)
Enhancements
- #12973 - Update daily limit popup to new design (by @ntindle)
- #12979 - Creator dashboard polish + MCP setup + settings link cleanup (by @Abhi1992002)
UI/UX Improvements
- #12972 - Update subscription screen plan copy (by @itsababseh)
- #12970 - Change onboarding checklist labels (by @Torantulino)
- #12976 - Settings v2 cleanup — sidebar, account route, headers, wallet (by @Abhi1992002)
- #12982 - Remove Compare plans link from billing YourPlanCard (by @Abhi1992002)
- #12984 - Settings v2 QA pass — dialog padding, billing tab restore, auto-refill defaults (by @Abhi1992002)
- #12996 - Remove country selector from onboarding step 4 (by @Torantulino)
- #13011 - Remove stale 'Reset daily limit' flash, always show 'Go to billing' (by @Abhi1992002)
- #13012 - Split UsageLimits into UsagePopover and UsageLimitReachedCard (by @0ubbe)
- #12999 - Update auto-refill warning text on billing settings page (by @itsababseh)
Bug Fixes
- #12947 - Cancel-banner copy + clearer 422 on currency mismatch (by @majdyz)
- #12952 - Gate credits-add pop-up behind
ENABLE_PLATFORM_PAYMENT(by @0ubbe) - #12956 - Bump Replicate per-second rate to cover up to A100-80GB (by @majdyz)
- #12950 - Run copilot executor in foreground (by @Bentlybro)
- #12948 - Bundle of chat stream stability fixes (PK dedup, race, compaction, errors, chips) (by @majdyz)
- #12959 - Reserve TOP_UP for real Stripe checkouts; refunds use grant_credits (by @majdyz)
- #12964 - Parse Prisma raw timestamptz strings in copilot weekly usage (by @majdyz)
- #12974 - Align 'Complete onboarding' wallet task and normalize reward amounts (by @kcze)
- #12969 - Remove BetaUserCredit monthly credit refill (by @Torantulino)
- #12980 - Persist onboarding store + skip Stripe on LOCAL (by @0ubbe)
- #12981 - Default new users to NO_TIER instead of PRO (by @ntindle)
- #12990 - Fail-closed on Redis unavailable to prevent USD-cap bypass (by @majdyz)
- #12994 - Stop bot setup prompts in unrelated threads (by @Bentlybro)
- #13003 - Bill Pro→Max upgrades immediately with payment failure rollback (by @majdyz)
- #12992 - Re-prompt on thinking-only finish; route storage-limit through DB-manager (by @majdyz)
Refactoring & Maintenance
- #12968 - Remove 4 earn-credits onboarding checklist items (by @Torantulino)
- #12951 - Use --json bucket instead of awk text-column parsing (by @majdyz)
- #12960 - Migrate direct hook tests to component-level tests (by @0ubbe)
🙏 Thanks to Our Contributors!
A huge shoutout to everyone who contributed to this release. Special thanks to our new contributors:
- @itsababseh
As well as our returning contributors:
- @majdyz
- @0ubbe
- @Abhi1992002
- @Bentlybro
- @ntindle
- @Torantulino
- @kcze
📥 How to Get This Update
To update to this version, run:
git pull origin autogpt-platform-beta-v0.6.59
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!
- Settings v2 pages for API keys, integrations, profile, and preferences management
- Copilot-bot integration for Python/discord.py
- Web Push notifications via VAPID for background message delivery
Full changelog
🚀 Release autogpt-platform-beta-v0.6.58
Date: April 2026
🔥 What's New?
New Features
- #12890 - Add message timestamps + accurate thought-for time in Copilot (by @majdyz)
- #12894 - Dynamic BlockCostType (SECOND/ITEMS/COST_USD/TOKENS) + E2B/FAL migration (by @majdyz)
- #12903 - Add MAX tier + LD-configurable pricing + hide unconfigured tiers (by @majdyz)
- #12880 - Inline picker-backed inputs via run_block + accept AgentInputBlock subclasses (by @anvyle)
- #12885 - Add Settings v2 page layout behind SETTINGS_V2 flag (by @Abhi1992002)
- #12907 - Settings v2 API keys page (by @Abhi1992002)
- #12909 - Dynamic COST_USD billing + close 8 cost-leak surfaces (by @majdyz)
- #12910 - LD-configurable rate-limit multipliers + relative UI display (by @majdyz)
- #12911 - Settings v2 integrations page + provider description SDK (by @Abhi1992002)
- #12826 - Add Claude Opus 4.7 model support (by @Otto-AGPT)
- #12924 - Add settings v2 profile page (by @Abhi1992002)
- #12925 - Preferences v2 page (by @Abhi1992002)
- #10014 - Filter store creators with approved agents (by @itsababseh)
- #12900 - Redis Cluster client support (by @majdyz)
- #12934 - Creator dashboard page for settings v2 (by @Abhi1992002)
- #12618 - Copilot-bot (Python / discord.py) (by @Bentlybro)
- #12723 - Add Web Push notifications via VAPID for background delivery (by @kcze)
- #12935 - Add subscription plan selection step to onboarding (by @itsababseh)
- #12943 - Gate onboarding SubscriptionStep behind ENABLE_PLATFORM_PAYMENT (by @0ubbe)
- #12933 - Subscription credit grants + paywall gate + dialog UX + cross-pod cache (by @majdyz)
Enhancements
- #12893 - Charge Ayrshare per-post + align Bannerbear/Jina floors (by @majdyz)
UI/UX Improvements
- #12937 - Small fixes and stagger animations on settings pages (by @Abhi1992002)
Bug Fixes
- #12887 - Disable ask_question tool in Copilot pending UX rework (by @majdyz)
- #12886 - Sync safety net for Redis-induced zombie sessions in Copilot (by @majdyz)
- #12892 - Raise baseline tool-round limit to 100 + graceful finish hint in Copilot (by @majdyz)
- #12889 - Tame Kimi compaction storm + tunable threshold + Langfuse cost backfill (by @majdyz)
- #12874 - Resolve autopilot beta blockers (SECRT-2266/2267/2268/2269) (by @ntindle)
- #12904 - Surface non-zero E2B exits as real results, not sandbox errors (by @majdyz)
- #12905 - Use {schema_prefix} in raw SQL migrations instead of hardcoded 'platform.' (by @majdyz)
- #12912 - Audit + correct stale LLM/block rates + migrate generic ReplicateModelBlock to COST_USD (by @majdyz)
- #12919 - Atomic upsert + drop eager include for notifications (by @majdyz)
- #12918 - Preserve interrupted SDK partial work on final-failure exit in Copilot (by @majdyz)
- #12921 - Filter null query params in buildUrlWithQuery (by @SymbolStar)
- #12920 - Redirect www to non-www with 308 to preserve request method (by @SymbolStar)
- #12926 - Surface empty-success ResultMessage as stream error in Copilot (SECRT-2252) (by @majdyz)
- #12927 - Pause idle timer during pending tools in Copilot (by @majdyz)
- #12932 - Normalize model name based on actual transport, not config shape (by @majdyz)
- #12813 - Fix streaming reconnect races, hydration ordering, and reasoning split in Copilot (by @0ubbe)
- #12944 - Skip OS push for onboarding payloads (by @kcze)
Refactoring & Maintenance
- #12883 - Migrate Ayrshare to standard managed-credential flow (by @majdyz)
- #12908 - Clearer naming + docs for managed-cred sweep in integrations (by @majdyz)
- #12914 - Use $REPO_ROOT in pr-test skill instead of hardcoded absolute path (by @majdyz)
- #12915 - Consolidate 6 LD flags into 2 JSON flags (by @majdyz)
- #12917 - Consolidate 4 model-routing LD flags into 1 JSON flag in Copilot (by @majdyz)
🙏 Thanks to Our Contributors!
A huge shoutout to everyone who contributed to this release. Special thanks to our new contributors:
- @SymbolStar
As well as our returning contributors:
- @majdyz
- @ntindle
- @anvyle
- @Abhi1992002
- @Otto-AGPT
- @itsababseh
- @0ubbe
- @Bentlybro
- @kcze
📥 How to Get This Update
To update to this version, run:
git pull origin autogpt-platform-beta-v0.6.58
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!
- Agent Briefing Panel and scheduled agents in library
- Subscription tier billing via Stripe Checkout
- xAI Grok 4.20 models via OpenRouter
Full changelog
🚀 Release autogpt-platform-beta-v0.6.57
Date: April 2026
🔥 What's New?
New Features
- #12764 - Add Agent Briefing Panel (by @0ubbe)
- #12727 - Subscription tier billing via Stripe Checkout (by @majdyz)
- #12620 - Add xAI Grok 4.20 models from OpenRouter (by @Torantulino)
- #12818 - Surface scheduled agents in library & copilot briefings (by @0ubbe)
- #12841 - Queue follow-up messages on busy sessions (UI + run_sub_session + AutoPilot block) (by @majdyz)
- #12865 - Defer paid-to-paid subscription downgrades + cancel-pending flow (by @majdyz)
- #12864 - Real OpenRouter cost + cost-based rate limits (percent-only public API) (by @majdyz)
- #12870 - Stream extended_thinking on baseline via OpenRouter (by @majdyz)
- #11235 - Add system diagnostics and execution management dashboard (by @ntindle)
- #12699 - AI chat panel for the flow builder (by @majdyz)
- #12615 - Platform server linking API for multi-platform CoPilot (by @Bentlybro)
- #11741 - Add AutoGPT logo to share page and zip download for outputs (by @ntindle)
- #12871 - Kimi K2.6 fast default + 4-config matrix + coalesced reasoning + web_search tool (by @majdyz)
- #12876 - Register 13 paid blocks + document credit/microdollar wallet boundary (by @majdyz)
- #12873 - Live baseline streaming + render flag + Sonar web_search + simulator cost tracking + reconnect fixes (by @majdyz)
- #12879 - TodoWrite for baseline copilot (by @majdyz)
- #12878 - SDK fast tier defaults to Kimi K2.6 via OpenRouter + vendor-aware cost + cross-model fix (by @majdyz)
- #12881 - LaunchDarkly per-user model routing (by @majdyz)
Enhancements
- #12804 - Unified transcript context — extract_context_messages, mode-gated --resume, compaction-aware gap-fill (by @majdyz)
- #12737 - Queue follow-up messages on busy sessions (by @majdyz)
- #12882 - Moonshot module + cache_control widening + partial-messages default-on + title cost (by @majdyz)
UI/UX Improvements
- #12690 - Allow zooming out more in the builder (by @chernistry)
- #12815 - Small UI fixes — sort menu bg, name update auth, stats grid overflow, pulse chips (by @0ubbe)
- #12854 - Compute monthly spend for AgentBriefingPanel (by @0ubbe)
- #12805 - Preserve agent name in AgentExecutor node title after reload (by @joemunene-by)
Bug Fixes
- #12799 - Default copilot sonnet to 4.6 (by @ntindle)
- #12802 - Disable fallback model to fix prod CLI rejection (by @majdyz)
- #12807 - Add check for GitHub account connection status (by @Torantulino)
- #12796 - Fix initial load missing messages + forward pagination for completed sessions (by @majdyz)
- #12814 - Replace dedup lock with idempotent append_and_save_message (by @majdyz)
- #12823 - Remove debug console.log statements (by @slepybear)
- #12831 - Revert forward pagination, add visibility guarantee for blank chat (by @majdyz)
- #12832 - Add gh auth status check to Tool Discovery Priority section (by @Torantulino)
- #12853 - Persist reasoning, split steps/reasoning UX, fix mid-turn promote stream stall (by @majdyz)
- #12862 - Dedupe tool row, lift bash_exec timeout, Stop+resend recovery (by @majdyz)
- #12867 - Release session cluster lock on completion (by @majdyz)
- #12866 - Reduce SDK + baseline prompt cache waste (by @majdyz)
- #12835 - Forward sentry-trace and baggage across API proxy (by @ntindle)
- #12856 - Artifact panel fixes (by @ntindle)
- #12834 - Allow multiple compactions per turn (by @ntindle)
Documentation
- #12819 - Convert in-code comments to MkDocs admonitions in block-sdk-gui (by @slepybear)
Dependencies & Maintenance
- #12812 - Make session-file helper functions public to fix Pyright warnings (by @majdyz)
- #12715 - Security upgrade next from 15.4.10 to 15.4.11 (by @ntindle)
🎉 Thanks to Our Contributors!
A huge shoutout to everyone who contributed to this release. Special thanks to our new contributors:
- @chernistry
- @slepybear
- @joemunene-by
As well as our returning contributors:
- @majdyz
- @ntindle
- @Torantulino
- @0ubbe
- @Bentlybro
📥 How to Get This Update
To update to this version, run:
git pull origin autogpt-platform-beta-v0.6.57
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!
- MemoryEnvelope metadata model
- Scoped memory retrieval
- Memory hardening
Full changelog
What's Changed
- fix(backend/copilot): pre-create assistant msg before first yield to prevent last_role=tool by @majdyz in https://github.com/Significant-Gravitas/AutoGPT/pull/12797
- feat(backend): MemoryEnvelope metadata model, scoped retrieval, and memory hardening by @ntindle in https://github.com/Significant-Gravitas/AutoGPT/pull/12765
Full Changelog: https://github.com/Significant-Gravitas/AutoGPT/compare/autogpt-platform-beta-v0.6.55...autogpt-platform-beta-v0.6.56
- P0 guardrails and security hardening in CoPilot
- LLM prompt caching with cross-user optimization via SystemPromptPreset
- Per-model cost breakdown and cache token tracking in platform admin
- Graphiti temporal knowledge graph memory for CoPilot
- Extended thinking execution mode for OrchestratorBlock
- Generic managed credential system with AgentMail auto-provisioning
- New /orchestrate skill for Claude Code agent fleet supervision
- SmartDecisionMakerBlock renamed to OrchestratorBlock
- Workflow import from n8n, Make.com, Zapier
- Parallel block execution via infrastructure-level pre-launch
- Dry-run execution mode with LLM block simulation
- CoPilot credit charging and token rate limiting with usage UI
- Support prompt-in-URL for shareable prompt links
- AutoPilotBlock for invoking AutoPilot from graphs
- @@agptfile file-ref protocol for tool call inputs
- Builder file inputs switched from base64 to workspace uploads
- Attach uploaded images and PDFs as multimodal vision blocks
- File upload to copilot chat
- E2B cloud sandbox with persistent execution and output truncation
- Agent-browser multi-step automation tools
- Legacy builder removed and replaced with new flow editor
- MCP tool block with OAuth and tool discovery
- Claude Agent SDK integration for CoPilot
- Folder organization system for agents
- Enforce disabled flag on blocks in graph validation
- Claude Opus 4.6 model support
- ClamAV scanning for local file paths
- Extended thinking enabled for Claude models
- Prevent reflected XSS in OAuth callback route
- User Workspace for persistent CoPilot file storage
- CoPilot speech-to-text via Whisper model
- Redis-based SSE reconnection for long-running operations
- ClaudeCodeBlock for executing tasks via Claude Code in E2B sandbox
- External Agent Generator service integration with clarification questions UI
- Long-running tool results persistence to survive SSE disconnects
- Unified hybrid search with BM25 reranking for store, docs, and blocks
- Agent generator tools for chat copilot
- Block documentation auto-generation system
- Reddit OAuth2 integration with advanced blocks
- WordPress Get All Posts block and Publish Post draft toggle
- Human-in-the-loop review system for blocks requiring approval
- OAuth API and Single Sign-On implementation
- Agent mode for SmartDecisionMakerBlock with autonomous loops
- Auto-save Draft Recovery with IndexedDB persistence