CrewAI
CrewAI is an open-source Python framework for multi-agent workflows. Agents with defined roles work together as a team (a “crew”); event-driven “flows” give the process a fixed structure.
- Project
- CrewAI
- Publisher
- CrewAI
- Repository
- github.com/crewAIInc/crewAI
- Documentation
- docs.crewai.com
- Languages
- Python
- Category
- AI agent frameworks
What is CrewAI?
CrewAI is a standalone Python framework for orchestrating AI agents that is not built on other agent frameworks. It combines high-level abstractions for collaboration between several agents with interfaces for precise control of the workflow. The project is released under the MIT licence; the vendor also offers a commercial platform for deployment, observability and governance.
Core capabilities
- Crews
- Teams of specialised agents with a role, goal, tools and tasks that divide work among themselves.
- Flows
- Event-driven workflows with state, branching and routing into which crews, single model calls and regular Python code are embedded.
- Tools and memory
- Tool integration, memory, knowledge sources, checkpointing, asynchronous execution and support for MCP and A2A.
- Human input
- A workflow can request human input and produce structured outputs.
- Model choice
- The OpenAI API is used by default; other providers and local models, for example via Ollama, can be connected.
When CrewAI fits
CrewAI fits when a task divides naturally into roles – such as research, analysis and drafting – and you want that division of labour to be visible in code. Flows help when parts of the process must be fixed. If you want to model every state transition individually, a graph framework such as LangGraph offers a lower level of abstraction.
Using CrewAI effectively in an AI-first business model
Roles in a crew are quickly defined. What matters is that each role serves a business outcome and that it is clear which actions it may perform on its own and which require approval. This gradation belongs in the operating model before the first crew is built.
Where several agents collaborate, quality becomes harder to trace. Plan tracing and evaluation from the start, and decide who is accountable for the crew in operation when data, models or goals change.
From tool to outcome
In AI-first coaching you clarify which outcomes AI agents should be accountable for in your go-to-market and within which framework. REVAGO then designs, builds and runs your system. You pay for results – not for activity.
Frequently asked questions
What is CrewAI?
CrewAI is an open-source Python framework (MIT licence) for orchestrating AI agents. It provides crews for collaboration between role-based agents and flows for event-driven workflows with precise control.
What is the difference between crews and flows?
Crews rely on autonomous collaboration between several agents with flexible decision-making. Flows define the process in an event-driven way, manage state and allow branching. Both can be combined, for example by having a flow call a crew at one step.
Is CrewAI built on LangChain?
No. According to the project description, CrewAI is a standalone framework with its own building blocks for agents, tasks, crews, flows and tools.
Further reading in the wiki
A neutral overview, not a recommendation of individual providers. REVAGO is not affiliated with the listed projects. Descriptions are based on the projects’ public information; features and licences may change.