CrewAI
An open-source Python framework for orchestrating role-playing, autonomous multi-agent crews, with an enterprise console.
What it is
CrewAI is an open-source (MIT) Python framework for building and orchestrating multi-agent workflows — "crews" and "flows" — with an enterprise platform for deploying, managing, and monitoring agents on CrewAI or private infrastructure.
At a glance
7/7 facts sourced- Pricing
- Basic is free (visual editor, AI copilot, GitHub integration, 50 workflow executions/mo). Enterprise is custom-priced (request a trial); no public figure.source ↗
- Hosting model
- Hybrid — the open-source library is self-hosted via CLI; the Enterprise console offers managed deployment on CrewAI or private infrastructure.source ↗
- Memory & state
- A unified Memory class where the LLM infers scope and importance on save; retrieval is ranked by similarity, recency, and importance, defaulting to LanceDB with custom backends supported.source ↗
- MCP support
- Nativesource ↗
- Integrations
- Connectors including Gmail, Google Drive, Outlook, Teams, OneDrive, HubSpot, Slack, and Salesforce, plus tools to call existing CrewAI automations or Amazon Bedrock Agents.source ↗
- Model providers
- Native SDKs for OpenAI, Anthropic, Google, Azure, AWS Bedrock, and Snowflake Cortex; many more via LiteLLM (Llama, Mistral, Groq, Ollama, Perplexity, OpenRouter, and others).source ↗
- Open source
- Yessource ↗
Editorial scorecard
Overall 4.0 / 5
Our opinion — the same 0–5 rubric is applied to every platform, including Alfe. See the rubric.
Strengths
- Open source (MIT) and self-hostable.
- Broad LLM coverage — six native SDKs plus many providers through LiteLLM.
- Native MCP support across Stdio, SSE, and Streamable HTTP transports.
Trade-offs
- The free plan is capped at 50 workflow executions per month.
- No public pricing for the Enterprise tier.
- Founding date and headquarters are not disclosed on the primary sites.
Our take
Editorial opinion — not a hands-on test resultIn our opinion, CrewAI is a good fit for teams that think in terms of role-based agents collaborating on a task and want an open-source Python core they can self-host, with a managed console available when they need it. Its MCP support and LLM breadth are notable strengths; the main watch-items are the capped free tier and the lack of public enterprise pricing.
Sponsored
Alfe is a managed agent-hosting platform. If you want an agent hosted and running without wiring infrastructure yourself, see Alfe → (or read its dossier).
FAQ
- Is CrewAI open source?
- Yes — the CrewAI framework is MIT-licensed.
- Does CrewAI support MCP?
- Yes, natively — via the mcps field or MCPServerAdapter over Stdio, SSE, and Streamable HTTP transports.
Compare CrewAI
Sources
- CrewAI homepage · accessed 2026-07-01
- CrewAI pricing · accessed 2026-07-01
- CrewAI MCP overview · accessed 2026-07-01
- CrewAI on GitHub (MIT license) · accessed 2026-07-01
The Agent Examiner editorial · last updated 2026-07-01. Facts are drawn from each vendor’s own public materials; scores are our editorial opinion (methodology).
Product and company names are trademarks of their respective owners and are used here only to identify the products under discussion.