CrewAI.
Multi-agent coordination: role-based delegation patterns.
What it is.
Brazilian open-source framework focused on multi-agent coordination. Defines agents by role (researcher, writer, reviewer) and orchestrates them through hierarchical or sequential processes. CrewAI Enterprise wraps the open-source library with hosted runtime, audit, and SLA.
Where it fits.
Use cases where the work decomposes naturally into roles — research, content, customer service. Strong for teams that want the agent topology to mirror an existing team structure they already understand.
- Role-based abstraction maps to existing team workflows
- Lightweight compared with LangGraph for simple flows
- Active community and example library
- Less explicit state control than LangGraph
- Enterprise feature gap behind LangChain ecosystem
Frequently asked.
When is CrewAI a better fit than LangGraph?
When the work decomposes cleanly into roles and the team values readability over explicit state-machine control.
Is CrewAI tied to a specific model?
No. LiteLLM under the hood means any major model works — Claude, GPT-5, Gemini, Llama, local.
Is CrewAI Enterprise required for production?
Not strictly, but production deployments above moderate scale typically need its hosted runtime, audit, and SLA features.
