← Back to Discover

CrewAI

An open-source Python framework for multi-agent automation, built from Crews of role-based agents and event-driven Flows, with a CLI for scaffolding and running projects.

Agent frameworksAutonomous agents
Visit website ↗

What it does

CrewAI is a Python framework for building multi-agent workflows. The repository describes two complementary pieces: Crews, teams of role-based agents that collaborate with autonomy, and Flows, event-driven workflows with state, branching and routing that can call agents, single LLM calls and ordinary Python code. A CLI installed with uv creates projects where agents and crew settings are defined in JSONC files, and an older Python and YAML layout is still available. Projects run with the crewai run command and need an API key for the model provider you choose. Sequential and hierarchical processes are supported, and the repository lists tools, memory, knowledge, checkpointing, async execution and MCP/A2A support. The framework is released under the MIT License and requires Python 3.10 or later but earlier than 3.14. The repository also describes a commercial suite, CrewAI AMP, that adds managed deployment, observability, governance and enterprise support; one part of it, the Crew Control Plane, can be tried for free.

Capabilities

  • Crews of role-based agents that delegate and collaborate
  • Flows for event-driven control with state, branching and routing
  • CLI that scaffolds a project and runs it with crewai run
  • Tools, memory, knowledge, checkpointing and async execution
  • MCP and A2A support
  • Sequential and hierarchical processes

When to use it

  • You want several specialized agents to work together on one task
  • You need event-driven control with branching around LLM calls
  • You are building Python agent automations and want a scaffolded project to start from