// the find
crewAIInc/crewAI
Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
CrewAI is a Python framework for building multi-agent LLM systems, offering two models: role-based autonomous 'Crews' and event-driven 'Flows' for more deterministic control. It targets developers building agentic automations who want more structure than a raw LangChain script but don't want to hand everything to a hosted platform.
The Crews vs Flows split is a genuinely useful distinction - most competing frameworks pretend everything should be autonomous agent chatter, while CrewAI lets you drop back into plain Python and structured state (pydantic models) when you need determinism. It's a standalone framework, not a LangChain wrapper, so there's less dependency bloat and fewer leaky abstractions. Tooling ecosystem (SerperDevTool, S3, vector search tools, MCP/A2A support) is broad and the CLI scaffolding (crewai create crew) gets you to a working example fast.
The README spends as much time pitching the paid AMP/enterprise control plane as explaining the open-source project, which is a signal to watch for feature gating down the line. The recent JSON-first project scaffold (agents/*.jsonc, crew.jsonc) is a breaking change from the old Python/YAML pattern and most existing tutorials/examples online still use the old format, so expect friction migrating. Telemetry is on by default and only opt-out via an env var, which some teams will flag in a security review. Multi-agent orchestration frameworks like this are still fundamentally prompting patterns with retry/parsing logic around them - debugging why a Crew went off the rails is still opaque without pulling in their paid tracing product.