Choose the plan that's right for you

Upgrade, downgrade, or change plans anytime. All plans include core features to get you started.

Not sure yet? Try it—Instant demo—No signup

Plans

Monthly billing | Yearly billing | Save 15%

Pro

1x usage
$42.50/ month
billed annually
For individuals getting started with dependable everyday automation.

• Access to virtually any leading AI model
• Agents work non-stop in the background
• Visual agent builder and chat-based agent creation
• File-aware agents that can work with your documents
• End-to-end agent management and run visibility
• Scheduled and event-based triggers

Get Pro

Cancel anytime

Max

8.5x usage
$272.00/ month
billed annually
For users who are serious about getting more work done.

Includes everything in Pro

• 8.5x Pro usage
• 5x file storage
• Early access to features before anyone else
• Expand our extensive integration library
• Priority support
• Help drive the roadmap for new features

Upgrade to Max

Cancel anytime

Team

Coming soon
Contact us
For teams that need collaboration, controls, and custom usage.

• Multi-user workspaces
• Admin controls & roles
• Centralized billing & seat management
• Collaboration & approvals
• Security & compliance options
• Secure access for shared integrations

Contact sales

Need something custom? Talk to sales.

AI Agents & AutoPilot

Highlighted feature. AutoPilot Chat: natural-language chat that creates, runs, debugs, and edits agents end-to-end.

• Agent Generation from Prompt: describe an agent in plain English; AutoPilot asks clarifying questions, builds, and saves it.
• Dry-Run Self-Repair Loop: newly-generated agents are simulated, errors caught, and fixed automatically.
• Personalized Home & Prompt Suggestions: quick-action prompts tailored to your signup answers and recent activity.
• AutoPilot Memory: persistent temporal knowledge graph remembers facts, preferences, and context across sessions.
• Parallel Tool Execution: AutoPilot fires independent tools concurrently instead of one-at-a-time.
• Context Compaction: long chats stay coherent via automatic compression and gap-filling.
• Reasoning Display: view extended-thinking traces from AutoPilot.
• Task Planning: AutoPilot tracks and updates structured task lists mid-job.
• Multimodal Inputs: drop images and PDFs into chat and have a vision model read them.
• Voice In & Out: speak to AutoPilot naturally with your voice and listen to its response via natural-sounding TTS voices.
• File Upload in Chat: upload files; AutoPilot reads, writes, edits, and references them across turns.
• Cloud Sandbox: persistent isolated environment for running code, files, and tools; auto-resumes on demand and syncs files back to your workspace.
• Highlighted feature. Teach AutoPilot Skills: AutoPilot learns your processes over time, building "Skills" for reuse whenever needed.
• GitHub CLI in AutoPilot: call gh commands inside chat with commits signed by your GitHub profile identity.
• Highlighted feature. Browser Automation: multi-step browser tasks via Stagehand and a Chromium-backed agent browser.
• Highlighted feature. MCP Tool Support: discover and execute Model Context Protocol tools from any compatible MCP server.
• Web Search: built-in web search plus Perplexity Sonar and WebFetch retrieval.
• SQL Analytics Tool: query your data with a read-only SQL block from inside chat.
• AutoPilot Block: invoke AutoPilot from inside any agent.
• Export Chat as Markdown: download your conversation history with one click.
• AutoPilot Notifications: push, web, and follow-up notifications keep you in the loop.

Builder

• Highlighted feature. Visual Agent Builder: drag-and-drop graph builder for composing agents from blocks.
• Auto-Save & Draft Recovery: IndexedDB-backed recovery of unsaved work.
• Undo / Redo with Keyboard Shortcuts: full builder history with batch undo for cascading edits.
• Copy, Paste & Multi-Select: copy blocks, edges, and multi-node selections between flows.
• Inline Node-Title Editing: rename any node directly on the canvas.
• Block Menu Search: fuzzy search over names, descriptions, and capabilities with relevance ranking.
• Friendly Block Names & Docs: technical block names mapped to plain-language labels with searchable docs.
• Inline Field Validation: graph errors surface on the offending node field before you run.
• Rich Node Outputs: render JSON, code, markdown, LaTeX, images, and files in node output panels.
• Expandable Block Outputs: drill into any block's output without leaving the canvas.
• Run From Builder: execute the agent you're editing without leaving the page.
• Task Presets: save task configurations as reusable presets.
• Open Task in Builder: jump from any task history entry back to the agent that produced it, with all execution steps preserved.
• Highlighted feature. Sub-Agents as Blocks: drop one agent into another as a sub-graph; credentials and approvals flow through.
• Agent Import / Export: export agents to file and import them back, with sub-agents intact.
• Rich Input Widgets: file picker, Google Drive picker, table input, JSON editor, multi-select, AnyOf/OneOf forms.
• Smart Type Coercion: structured data automatically parsed and coerced across block schema mismatches.
• Inline Credential Prompts: missing credentials surface and resolve when you run, not after a failure.
• Auto-Select Credentials: builder picks the right credential when only one fits.
• Builder Tutorial: interactive walkthrough for new users.

Execution, Scheduling & Triggers

• Real-Time Execution Stream: WebSocket-driven live task updates with step counts and activity summaries.
• Task Cost Display: see cost for every task and every block.
• Live Activity Status: AI-generated, plain-English summary of what an agent is doing right now.
• Correctness & Accuracy Scoring: execution-level quality scores with alerting on regressions.
• Task History with Infinite Scroll: paginated, searchable history with exact-timestamp tooltips.
• Stop & Cascade Stop: cancel any task; stopping a parent cascades into nested sub-agents.
• Continue Aborted Tasks: resume broken tasks instead of starting over.
• Parallel Block Execution: independent blocks run concurrently inside an agent.
• Cron Scheduling: schedule any agent on a cron with timezone awareness and validation.
• Run Now on Schedules: fire a scheduled agent on demand from the schedule page.
• Version-Pinned Schedules: schedule a specific agent version, not always the latest.
• Recommended Schedules: agent-aware suggestions for when to run.
• Late & Stuck Execution Alerts: automatic alerting on missed cron windows or tasks exceeding one day.
• Trigger On Anything: broad trigger framework for event-driven flows.
• Generic Webhook Trigger: fire any agent from an arbitrary HTTPS webhook payload.
• GitHub Webhook Triggers: purpose-built blocks for GitHub events (push, PR, issue, release).
• Manual Webhook Setup: generate and copy a webhook URL for any triggered agent.
• Triggered Agents in Library: library view surfaces and manages trigger-driven agents.

Blocks & Integrations

• Highlighted feature. 200+ Integration Blocks: pre-built connectors for the most popular SaaS and developer tools (full catalog at end of doc).
• Communication Suite: Slack, Discord, Telegram, Email/SMTP, Gmail.
• Developer Suite: GitHub (repos, PRs, issues, files, checks, statuses), Linear.
• Google Workspace: Sheets, Docs, Calendar, Gmail, plus a Google Drive file picker.
• Productivity Tools: Airtable, Notion, Todoist.
• CMS & Publishing: WordPress, Medium.
• CRM & Sales: HubSpot, Apollo, EnrichLayer, AgentMail.
• Social Media: Twitter/X plus Ayrshare posting to TikTok, YouTube, Instagram, LinkedIn, Facebook, and more.
• Web Search & Research: Exa, Perplexity, Jina, Wolfram Alpha, DataForSEO.
• Web Scraping: Firecrawl, Jina extraction, ScreenshotOne, Bannerbear overlays.
• Browser Automation: Stagehand, Chromium-backed agent browser tools.
• AI & Dev Tool Blocks: MCP tools, Claude Code, E2B sandbox.
• Image, Video & Music Generation: Ideogram, Nano Banana, Flux, VEO3, Fal, Replicate, MusicGen, Revid.ai.
• Vector & Memory: Pinecone vector storage, Mem0 long-term memory.
• Data Transformation Blocks: text encode/replace, XML parser, spreadsheet reader, list/dict manipulation, word counts, time/date formatting.
• AI Reasoning Blocks: AI Condition, Structured Response, AI Conversation, AI List Generator, Fact Checker.
• Key-Value Storage Block: persistent KV store accessible from any agent to share data across tasks and agents.
• HTTP Request Block: generic web requests with multipart upload, host-scoped auth, and safe-redirect protection.
• SQL Query Block: read-only SQL access to your platform data, including custom analytics views.
• Block Development SDK: build, register, and ship your own custom blocks.
• Auto-Generated Block Docs: every block documented automatically and surfaced in builder search.

LLM Providers & Models

• Highlighted feature. All Major LLM Families: frontier models from Anthropic, OpenAI, Google, xAI, Meta, Mistral, Cohere, and more.
• Claude Family: Opus 4.7, 4.6, 4.5; Sonnet 4.6, 4.5; Haiku 4.5.
• GPT Family: GPT-5.2, GPT-5.1, GPT-5; plus GPT-OSS open-source models.
• Gemini Family: Gemini 3.1 Pro Preview, Gemini 3 Flash, and the 2.5 line.
• Grok (xAI): Grok 3, Grok 4.20, plus latest updates.
• Llama Family: Llama 4 Maverick, Scout, and other Llama variants.
• Kimi / Moonshot: Kimi K2.6 and Moonshot models.
• Mistral Family: Mistral Large 3, Medium 3.1, Small 3.2, Codestral.
• Cohere Command A: full Command A family of models.
• Microsoft Phi-4: Phi-4 reasoning model.
• Perplexity Sonar: Sonar and Sonar Reasoning Pro for real-time search-grounded answers.
• Z.ai GLM: GLM 5, GLM 5 Turbo, GLM 4.7, and more (7 variants).
• DeepSeek: DeepSeek V3 and DeepSeek R1.
• Qwen: Qwen3 235B Thinking and Qwen3 Coder.
• v0 by Vercel: v0 code-generation models.
• Per-Task LLM Budget Caps: set max_budget_usd to bound LLM spend on any task.
• Token & Cost Tracking: per-model token consumption, cache hits, and dollar cost on every task.
• Prompt Caching: cache LLM prompts to reduce cost and latency on repeat or long conversations.

Library, Marketplace & Sharing

• Agent Library: personal library of agents, presets, and recent tasks.
• Folders for Library Agents: nested folders for organizing large agent collections.
• Favorite Agents: pin frequently-used agents for fast access.
• Fork Agents: duplicate any library agent to customize and make it your own.
• Highlighted feature. Agent Marketplace: public marketplace of community-built agents with ratings, task counts, and previews.
• Run & Save Marketplace Agents: execute any published agent, or add it to your library, without forking.
• Marketplace Search & Filters: full-text, hybrid, and faceted search across agents and creators.
• Import from n8n, Make.com, Zapier: bring existing automations from other tools onto AutoGPT as agents.
• Featured Creators & Agents: curated discovery on the marketplace home.
• Creator Profiles: public profile pages with bio, avatar, social links, and published agents.
• Creator Dashboard: manage submissions, listings, analytics, and review status.
• Agent Submissions: submit agents to the marketplace with output demos, instructions, and rich media.
• AI-Generate Listing Images & Thumbnails: auto-generated visuals for marketplace cards.
• Highlighted feature. Public Share Links: share a task's results with anyone via public URL.
• Prompt-in-URL Sharing (AutoPilot): pre-fill an AutoPilot chat prompt through a shareable link.
• Output ZIP Download: download a task's output files together as a single ZIP archive.
• Marketplace Update Alerts: notifications when library agents you use receive an upstream update.

Auth, Credentials & Security

• Google Sign-In: single-click Google authentication.
• Email & Password Login: password authentication with secure reset flows.
• Single Sign-On (OAuth): enterprise OAuth 2.0-based SSO.
• Password Strength Requirements: 12-character minimum enforced for new and updated passwords.
• CAPTCHA on Auth Forms: Turnstile-backed CAPTCHA on login, signup, and password reset.
• API Key Management: generate, view, rotate, and revoke API keys for programmatic access.
• Credentials Manager: central UI to view, edit, and revoke OAuth, API-key, and basic-auth credentials for 3rd party Integrations.
• Multi-Provider Credentials per Service: pick from multiple auth methods for the same service.
• Optional Credentials: blocks can mark credentials optional for execution paths that are optional.
• Host-Scoped HTTP Credentials: HTTP request credentials limited to specific hostnames.
• Highlighted feature. Managed Credentials — No Accounts to Create: we supply the API keys for most integrations, so you never sign up for another service; just connect the ones you already use (Gmail, HubSpot, etc.).
• Inline Credential Setup: resolve missing credentials when you run or schedule, instead of failing.
• Human-in-the-Loop Approvals: place approval gates anywhere in an agent (including sub-agents); execution pauses for human review, then resumes.
• Safe URL Redirects & DNS-Rebinding Protection: HTTP blocks validate redirects and prevent open-redirect attacks.
• Credential Scope Filtering: credentials filtered by compatible services and security scope.
• Admin Impersonation: admins can impersonate users for debugging, with audit trails.

Storage, Files & Workspaces

• Persistent User Workspace: durable file storage that survives across chat sessions and tasks.
• Workspace File Tools: list, read, write, edit, and folder-manage workspace files from any agent.
• Large File Uploads: files up to 256 MB supported.
• Antivirus Scanning: ClamAV scans every uploaded and workspace file before use.
• Multimodal File Support: image and PDF inputs fed straight into vision-capable blocks.
• Rich Media Previews: preview cards for images, video, audio, and structured outputs.
• Builder File Inputs from Workspace: agents pick files from your workspace, not just base64 blobs.
• Google Drive Picker: pick files from Google Drive as agent inputs.

Billing, Credits & Usage

• The model: automations (running agents/workflows) are pay-as-you-go via the credit wallet — the same on every plan. The subscription (Pro vs Max) gates AutoPilot Chat usage, and that's the one real volume difference between plans.
• Highlighted feature. AutoPilot Chat Allowance: how much AutoPilot Chat you can use; Max's allowance is 8.5× Pro's.

Standard

8.5×

Unlimited (BYO LLM keys)

• Highlighted feature. Support Level: how you get help when something goes wrong.

Email

Priority + Onboarding

Community / GitHub

• Pay-As-You-Go Automations: pay only for the automation runs you use, billed from your credit wallet.
• BYO compute
• Wallet & Top-Up: pre-paid credit wallet with on-demand top-up.

Auto-Refill Wallet: automatic refills with configurable minimums.

Execution Cost Visibility: see dollar cost per task, per node, per block.

Notifications & Communication

• Real-Time WebSocket Notifications: in-app notifications delivered live.
• Web Push Notifications: VAPID-based browser push, even when the tab is closed.
• Email Notifications: reliable delivery with retry and deadletter handling.
• BYO SMTP
• Daily Email Digest: opt-in roll-up of notifications into a single daily email.
• BYO SMTP
• Submission Review Notifications: creators notified of marketplace submission decisions.

Self-Host Admin Tools

Rows in this section are admin/operations tools that exist for whoever runs the platform. On Pro and Max we run it — so these are internal AGPT-staff tools not exposed to customers. On Self-Host you run it — so these become real features for the self-hoster.

• Execution Diagnostics Dashboard: system diagnostics and execution oversight for platform admins.
• Execution Analytics Dashboard: execution metrics, performance, and trends.
• User Spending Dashboard: per-user spending and billing health.
• User Search & Rate-Limit Controls: find users and adjust per-user limits.
• Credit & Balance Admin Tools: grant credits, add dollars, and adjust balances.
• CSV Exports: export credit transactions and usage reports for finance.
• Platform Cost Dashboard: admin visibility into model costs, token usage, and cache efficiency.

APIs, SDKs & Developer Platform

• Highlighted feature. Public REST API: programmatic access to agents, tasks, blocks, and integrations.
• Create Agents via API: POST /graphs deploys an agent without opening the UI.
• Execution Completion API: wait for and retrieve agent execution results from the API.
• External Integration Management API: manage credentials and connections via API.
• Store Search & Usage API: public endpoints for marketplace discovery and access.
• OpenAPI Schema: auto-generated OpenAPI spec for the full public API.
• Generated SDK Clients: language SDKs auto-generated from the OpenAPI spec.
• Platform Linking API: multi-product platform linking for embedded integrations.
• CORS Support: regex-based origin allowlisting.
• Streaming API Responses: server-sent events / streaming proxy for long-running endpoints.
• Highlighted feature. Self-Hosting via Docker Compose: single-command self-host with bundled services.
• Self-Host Setup Scripts: guided scripts for getting a self-hosted instance running.
• Self-Host Platform Support: supported on Podman, WSL, and Raspberry Pi in addition to mainstream Linux.