Python framework for building autonomous AI agents that research, plan, code and run tasks, with memory, RAG and 100+ LLM providers.
crewAIInc/crewAI
Open-source Python framework for orchestrating role-based autonomous AI agents (Crews) and event-driven workflows (Flows).
- Stars
- 59,907
- Forks
- 8,806
- Open issues
- 762
- Last commit
- 22 hours ago
What is it?
CrewAI is a Python framework for building multi-agent systems, where each agent has a defined role, goal, and backstory and can use tools, memory, and knowledge sources. Agents are grouped into Crews that divide work through tasks and can run sequentially or under a hierarchical manager process.
Flows provide the deterministic side of the framework: event-driven workflows with shared state, conditional branching, routing via decorators such as @start, @listen, and @router, and the logical operators or_ and and_. A Flow can call a single LLM, run a Crew as one of its steps, and combine both in one pipeline.
The project is configured through JSON files (agents/*.jsonc and crew.jsonc) or through Python classes, and the crewai CLI scaffolds, installs, and runs projects. It requires Python 3.10 to below 3.14, supports connecting to OpenAI and local models such as Ollama, and includes MCP and A2A support.
Key features
- 01Role-based agents with goals, backstories, tools, and memory collaborate inside Crews.
- 02Event-driven Flows with shared state, conditional branching, and routing decorators.
- 03Sequential and hierarchical process modes for coordinating tasks.
- 04Project scaffolding, installation, and execution through the crewai CLI.
- 05Connects to OpenAI and local models such as Ollama and LM Studio.
- 06MCP and A2A support for agent integrations.
Use cases
- Automating multi-step research and report writing with specialized agents.
- Building production workflows that mix LLM calls with deterministic Python business logic.
- Prototyping agent teams that delegate subtasks and validate each other's output.
- Adding human review steps and structured outputs to AI-driven processes.
Quick start
- 1
Check that Python is between 3.10 and 3.14
python3 --version - 2
Run the official uv install script from the README (download and pipe to shell)
- 3
Install the CrewAI CLI
uv tool install crewai - 4
Create a new crew project
crewai create crew <project_name>
Things to know
- Requires Python >=3.10 and <3.14.
- Example crews call a hosted LLM such as openai/gpt-4o, so you need your own model API key set in the project's .env file.
- On Windows, the chroma-hnswlib dependency may fail to build without Visual Studio Build Tools.
Topics
FAQ
- What is crewAI?
- Open-source Python framework for orchestrating role-based autonomous AI agents (Crews) and event-driven workflows (Flows). CrewAI is a Python framework for building multi-agent systems, where each agent has a defined role, goal, and backstory and can use tools, memory, and knowledge sources. Agents are grouped into Crews that divide work through tasks and can run sequentially or under a hierarchical manager process.
- Is crewAI free to use? What license does it have?
- crewAI is open source under the MIT license. Check the license terms before using it in commercial products.
- Is crewAI still maintained?
- Actively maintained. The last commit to crewAI was on Oct 20, 2026.
- What can I use crewAI for?
- Automating multi-step research and report writing with specialized agents. Building production workflows that mix LLM calls with deterministic Python business logic. Prototyping agent teams that delegate subtasks and validate each other's output. Adding human review steps and structured outputs to AI-driven processes.
- How do I install crewAI?
- Check that Python is between 3.10 and 3.14: python3 --version Run the official uv install script from the README (download and pipe to shell) Install the CrewAI CLI: uv tool install crewai Create a new crew project: crewai create crew <project_name>
- Where can I get crewAI?
- The source code and installation instructions are on GitHub: https://github.com/crewAIInc/crewAI
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
AutoGPT is an open-source platform for building, running, and scheduling AI agents that carry out multi-step workflows from plain-English descriptions or a visual builder.
A Python framework for building LLM-powered applications and agents, with standard interfaces for models, embeddings, vector stores and third-party integrations.
Self-hosted Node.js and React app for orchestrating teams of AI agents, with org charts, task tracking, budgets and governance.
Self-hosted personal AI agent framework in Python with a browser WebUI, terminal client, chat app channels, tools, memory and MCP support.
A low-level Python framework for building stateful, long-running agents with durable execution, human-in-the-loop and memory.