# CrewAI — Open-source multi-agent framework with a managed platform for deploying and governing agent workflows. > Source: The Agents Index — https://theagentsindex.com/crewai (structured, researched, re-verified) > Facts last verified: 2026-09-13 CrewAI is an open-source Python framework for building crews of role-playing AI agents and event-driven flows. A managed platform, CrewAI AMP, adds a visual studio, deployment, tracing and governance on top of that same framework. | Fact | Value | | --- | --- | | Website | https://www.crewai.com | | Price | ENTERPRISE Custom Scale agentic AI across the organization, with the governance to keep it in production. | | Free tier | Yes | Basic plan, $0 and self-serve: 2 automations, 50 included workflow executions per month with a hard maximum of 50 and no purchasable overage. Includes Crew Studio, GitHub integration, tracing, OpenTelemetry, guardrails, human-in-the-loop input, cron scheduling and community support. Enterprise connectors, dedicated VPC, SSO and role-based access control are Enterprise-only. ## What it does **Writes code** · Partly — Agents you build can author and run Python when code execution is switched on, but this is a capability of agents inside the framework rather than a product that ships code for you. > CrewAI Agents now have the powerful ability to write and execute code, significantly enhancing their problem-solving capabilities. > — https://docs.crewai.com/en/learn/coding-agents · Coding Agents page, Introduction paragraph **Debugs and fixes** · Partly — A code-executing agent receives the exception from its own generated code and retries, capped by max_retry_limit; nothing documents diagnosing or fixing a failure in an existing repository. > If executed code raises an exception, the agent will receive the error message and can attempt to correct the code or provide alternative solutions. > — https://docs.crewai.com/en/learn/coding-agents · Coding Agents page, Important Considerations, item 2 Error Handling **Ships it** · Partly — It ships the automation itself rather than landing code changes in your repository: crewai deploy create pushes a crew or flow to AMP. > Deploy the crew or flow to CrewAI AMP. > — https://docs.crewai.com/en/concepts/cli · Available Commands, section 10 Deploy **Done-for-you service** · Partly — You run the software yourself, but the Enterprise plan sells forward-deployed engineers, onboarding and training separately, so people-delivered build help exists as a paid extra. > Forward deployed engineering and training, available a la carte > — https://www.crewai.com/pricing · Plans for every stage, Enterprise plan card, fourth bullet ## How much it does unattended **Runs autonomously** · Yes — The vendor positions crews as autonomous workers that interact with enterprise applications unattended, with hierarchical delegation and cron scheduling available on both plans. > Build a crew of AI agents that autonomously interact with enterprise applications and use tools to automate workflows and tasks, with or without writing code. > — https://www.crewai.com/agent-management-platform · CREW AI AMP section, closing paragraph **Multi-agent** · Yes — Multi-agent coordination is the core abstraction: crews of specialised agents delegate between themselves, with MCP for tooling and A2A for cross-agent discovery. > CrewAI transforms a set of AI agents into a crew of AI agents that supporting protocols such as MCP for tooling and A2A for agent collaboration and discovery > — https://www.crewai.com/open-source · COLLABORATION section, Agents that collaborate **Agent permissions** · Partly — Code execution is off by default in the framework and the platform adds human-in-the-loop checkpoints and guardrails, but the governance layer described here is an Enterprise plan feature. > Every agent runs inside a scalable governance layer with role-based access, audit trails, human-in-the-loop checkpoints, and compliance controls baked in from day one. > — https://www.crewai.com/agent-management-platform · Agent Management Platform page, step 04 Govern ## Which models it runs on **Claude** · Yes — Anthropic is one of the natively integrated providers, configured with your own key, and Claude models are also reachable through Bedrock and Snowflake Cortex. > CrewAI provides native integration with Anthropic through the Anthropic Python SDK. > — https://docs.crewai.com/en/concepts/llms · Provider Configuration Examples, Anthropic accordion **GPT** · Yes — OpenAI has a native SDK integration covering both Chat Completions and the Responses API, and gpt-4o-mini is the framework's documented default model. > CrewAI provides native integration with OpenAI through the OpenAI Python SDK. > — https://docs.crewai.com/en/concepts/llms · Provider Configuration Examples, OpenAI accordion **Open models** · Yes — Local and open-weight serving is documented end to end, with step-by-step Ollama setup plus Hugging Face, Together, Groq and local NVIDIA NIM examples via LiteLLM. > This flexibility allows you to choose the most suitable model for your specific needs, whether you prioritize performance, cost-efficiency, or local deployment. > — https://docs.crewai.com/en/learn/llm-connections · Connect to any LLM page, Conclusion, after the Using Local Models with Ollama steps **Model choice** · Yes — You choose the model per agent, per task or per crew; five providers have native SDKs and everything else routes through LiteLLM or any OpenAI-compatible base URL. > CrewAI connects to LLMs through native SDK integrations for the most popular providers (OpenAI, Anthropic, Google Gemini, Azure, and AWS Bedrock), and uses LiteLLM as a flexible fallback > — https://docs.crewai.com/en/learn/llm-connections · Connect CrewAI to LLMs, opening paragraph ## Where you use it **In your editor** · Partly — There is no CrewAI editor extension; instead the vendor publishes a skill pack that teaches Claude Code, Cursor and Codex to scaffold CrewAI projects inside your editor. > They teach coding agents—like Claude Code, Cursor, and Codex—how to scaffold Flows, configure Crews, use tools, and follow CrewAI patterns. > — https://docs.crewai.com/en/skills · Skills page, first paragraph under the heading **On the command line** · Yes — A first-class terminal surface covers scaffolding, running, training, testing, replaying and deploying crews and flows, including login against enterprise SSO. > The CrewAI CLI provides a set of commands to interact with CrewAI, allowing you to create, train, run, and manage crews & flows. > — https://docs.crewai.com/en/concepts/cli · CLI page, Overview **In your pipeline** · Partly — Deployments connect to GitHub repositories and agents can file issues and cut releases, but the documented action list contains no pull-request review or CI job step. > Enable your agents to manage repositories, issues, and releases through GitHub. Create and update issues, manage releases, track project development, and streamline your software development workflow > — https://docs-platform.crewai.com/platform/en/integrations/github · GitHub Integration page, Overview, above the Available Actions list **In a browser** · Yes — Both plans include the hosted Crew Studio visual editor at app.crewai.com, where workflows are built, tested and run without writing code. > Studio (visual editor) Yes Yes > — https://www.crewai.com/pricing · Pricing comparison table, Build and integrate, Studio (visual editor) row (Basic, Enterprise) ## Whose machine it runs on **Self-hosted** · Yes — The open-source framework installs and runs on your own machines on any tier, while self-hosting the managed platform (VPC, own infrastructure, Factory) is an Enterprise option. > Deploy on CrewAI cloud, your own VPC, or your own infrastructure > — https://www.crewai.com/pricing · Plans for every stage, Enterprise plan card, second bullet **Open source** · Yes — The framework is genuinely open source under a permissive MIT-style licence in the vendor's own repository; the AMP platform around it is closed and commercial. > Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction > — https://github.com/crewAIInc/crewAI/blob/main/LICENSE · LICENSE file, grant paragraph (repository linked from the crewai.com footer as GitHub project) ## What it costs to run **How it meters** · Partly — Metering is per workflow execution rather than per seat, and the free tier states its allowance, but Enterprise volumes are sized case by case with unpriced overage. > Included executions 50 Sized to workflow > — https://www.crewai.com/pricing · Pricing comparison table, Agentic Workflow, Included executions row (Basic, Enterprise) **Free tier** · Yes — A standing free Basic plan is self-serve from the sign-up page, with the visual editor, GitHub integration and a monthly execution allowance, no sales call required. > Visual editor and AI copilot GitHub integration 50 workflow executions/month > — https://www.crewai.com/pricing · Plans for every stage, Basic plan card (price cell reads Free), feature bullets **API access** · Yes — Every deployed crew exposes REST kickoff and status endpoints behind a bearer token, and a separate platform-wide v1 API manages automations, users and teams though it is still beta. > You can also kickoff crews programmatically using the CrewAI AMP REST API. > — https://docs-platform.crewai.com/platform/en/guides/kickoff-crew · Kickoff Crew guide, Method 2: Using the API **MCP server** · Yes — Both plans can publish a built workflow as an MCP server so other agents can call it, and agents can equally consume any external MCP server as tooling. > Export as MCP server Yes Yes > — https://www.crewai.com/pricing · Pricing comparison table, Build and integrate, Export as MCP server row (Basic, Enterprise) **Bring your own key** · Yes — Inference runs on your own provider accounts: you supply each provider key, or manage LLM connections centrally in the platform, and Enterprise can host the runtime in your cloud. > Once you specify the model you are using, you will need to provide the configuration (like an API key) for each of the model providers you use. > — https://docs.crewai.com/en/concepts/llms · Setting up your LLM, introductory paragraph ## Buying it for a team **A company can buy it** · Yes — Accounts are organisations with members, teams, roles and billing under organisation settings, and the Enterprise plan is sold to the organisation rather than the individual. > Manage billing, plans, and organization configuration > — https://docs-platform.crewai.com/platform/en/features/rbac · Feature Permissions Matrix, organization_settings row, Description column **Pooled budget** · Partly — The published allowance is stated per plan per month rather than per member, so usage is drawn from one plan-level pool, but no per-seat or per-team budget split is documented. > 50 workflow executions/month > — https://www.crewai.com/pricing · Plans for every stage, Basic plan card, third bullet **Admin controls** · Partly — Admins get feature-level roles plus entity permissions over individual automations, env vars, LLM connections and repos, but the pricing table lists RBAC as Enterprise-only. > In addition to organization‑wide roles, CrewAI supports fine‑grained entity-level permissions that restrict access to individual resources. > — https://docs-platform.crewai.com/platform/en/features/rbac · Automation-level Access Control (Entity Permissions), opening paragraph **Audit log** · Partly — The vendor claims audit trails inside the governance layer and revoked role grants are retained for audit, but audit logging is not a listed row on either plan and governance is Enterprise-gated. > Every agent runs inside a scalable governance layer with role-based access, audit trails, human-in-the-loop checkpoints, and compliance controls baked in from day one. > — https://www.crewai.com/agent-management-platform · Agent Management Platform page, step 04 Govern **Single sign-on** · Partly — SAML 2.0 and OIDC against Entra ID, Okta, Auth0 or Keycloak are supported, including CLI device flow, but the pricing table gives SSO to the Enterprise plan only. > SAML 2.0 and OAuth 2.0 / OIDC protocol support > — https://docs-platform.crewai.com/platform/en/features/sso · Single Sign-On page, Key Capabilities list, first bullet _Not established: Seat model_ ## What happens to your code **Opt out of training** · Partly — Framework telemetry excludes prompts and agent data by default and is disabled with one environment variable, but no page states whether platform execution data trains vendor models. > Users can disable CrewAI telemetry by setting CREWAI_DISABLE_TELEMETRY to true, 1, yes, or on (any case). > — https://docs.crewai.com/en/telemetry · Telemetry page, paragraph above the Examples code block **Data residency** · Partly — Enterprise buyers can keep execution in their own VPC or infrastructure, but no region menu is published for the managed cloud and the trust centre lists AWS in the US as cloud subprocessor. > Deploy on CrewAI cloud, your own VPC, or your own infrastructure > — https://www.crewai.com/pricing · Plans for every stage, Enterprise plan card, second bullet **Getting out** · Partly — Automation source can be pulled back as a ZIP over the API and traces can be shipped to your own OpenTelemetry collector; nothing published covers rollover of unused executions. > Downloads an Automation source archive as a ZIP file. > — https://docs-platform.crewai.com/api/v1/reference/automations/export-an-automation-zip · Export an Automation ZIP endpoint, description line **Certifications** · Yes — The trust centre publishes a SOC 2 Type 2 programme with reports on request, plus a HIPAA audit report and pen test, and Enterprise adds FedRAMP High and SAM deployment options. > We maintain a SOC 2 Type 2 certified security program with comprehensive controls across our infrastructure, endpoints, and identity management. > — https://trust.crewai.com/ · Trust Center overview, Welcome to the CrewAI Trust Center paragraph ## Platform record (as of 2026-08-30) Reported by the code-hosting platform, not by the vendor. | Fact | Value | | --- | --- | | Licence | MIT | | Stars | 57822 | | Forks | 8287 | | Last push | 2026-08-28 | ## Provenance 1 of 23 sources behind the facts above are independent of the vendor. ## Coverage elsewhere None recorded. Every fact above comes from the vendor and is not independently corroborated yet. ## What changed - 2026-09-13 — Capabilities · Commercial terms · Free tier · Price - 2026-09-02 — Capabilities · Commercial terms · Free tier - 2026-08-31 — Capabilities · Commercial terms · Free tier - 2026-08-23 — Alternatives · Documentation link · One-liner · Repository · Summary - 2026-08-16 — One-liner · Summary ## Correcting this record Anything marked "Not established" is a question we have not answered — it is not a No. If you represent CrewAI, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth. Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.