CrewAI
Multi-agent framework and enterprise platform for building, running and governing teams of role-based agents
CrewAI is a platform for building and running multi-agent automations, built on the open-source CrewAI Python framework. Developers and business teams define role-based agents that work together on long-horizon tasks, while platform teams govern them centrally.
Key features:
- Build crews in a no-code visual editor with an AI copilot that exports to Python, or code-first through the API.
- Find automation candidates with Discovery, which ranks opportunities from tickets, chats and workflows.
- Trace every LLM call, tool call and memory read with cost accounting and OpenTelemetry.
- Add human approval gates, guardrails and PII redaction hooks at runtime.
- Export workflows as MCP servers or UI components and chat with them in Slack or Teams.
- Train agents, test multiple LLMs, with evaluation via Arize, Galileo, Datadog and Patronus.
The framework is open source under the MIT license. The hosted platform has a free Basic plan with 50 workflow executions a month; Enterprise adds SSO, RBAC and deployment in CrewAI cloud, a customer VPC or on-premises, with pricing through sales.