C
crewAI
Sign in Sign up
● CrewAI OSS

The open source, multi-agent
orchestration framework

Empowering developers to orchestrate high-performing AI agents with ease and scale.

Install CrewAI Run quickstart
main.py
# Agent: AI Agents Senior Researcher
# Task: Identify the next big trend in AI Agents with pros and cons.
class AIAgentResearcher:
def research(self):
# Thought: To identify the next big trend in AI Agents...
search_query = "emerging AI agent technologies 2024 trends"
return results
# Tool Output:
Search results: Title: The State of AI Agents: Lots of Potential...

Loved by AI builders. Trusted by AI leaders.

ABInBev BDO CHG Healthcare docusign experian EXL genpact globo.com
Johnson & Johnson kovecta kt NICE Actimize NTT DATA pepsico pwc
● Why Open Source

CrewAI is built in the open to make multi-agent AI accessible to all. With no-code speed and full-code power, it gives developers and teams the freedom to create, orchestrate, and scale agents without barriers. Backed by a growing community, CrewAI sets the standard for transparent, reliable, and enterprise-ready AI.

You define AI agents and tasks,
CrewAI builds and orchestrates them

Request a demo
main.py
my_crew = Crew(
agents=self.agents,
tasks=self.tasks,
process=Process.sequential,
planning=True,
)

Planning

Crews of AI agents can take advantage of a specialized planning agent which will create a step-by-step plan for all tasks, and share it with the crew.

Build a crew of AI agents with abstraction

Define AI agents

Simply describe an AI agent's role, goal and backstory to get started. Optionally, you can specify its LLM, enable reasoning and memory, provide it tools and more.

  • Define in YAML, code or both
  • Set the LLM and enable reasoning and/or memory
  • Specify tools and knowledge sources

Define tasks

Define the crew

main.py
agent = Agent(
role="Senior Data Scientist",
goal="Analyze and interpret complex datasets",
backstory="With over 10 years of experience...",
llm="gpt-4",
reasoning=True,
memory=True,
tools=[SerperDevTool()],
)

Traces are now available for local crews and flows.

Learn more
main.py
crew = Crew(
agents=[researcher, writer],
tasks=[research_task, writing_task],
process=Process.sequential,
tracing=True, # Enable tracing in your crew
)
✓ Trace batch finalized with session ID: 244c416a...

Guides and tutorials

Getting Started with CrewAI

Watch video

Building Your First Crew

Watch video

Advanced Agent Configuration

Watch video

Deploying to Production

Watch video

For more examples, visit our examples repository and our curated playlist of Youtube tutorials from the community.

Be part of a global
developer community

Connect with like-minded developers in our forum, contribute to the open-source project on GitHub, and collaborate on exciting AI automation projects. Together, we can accelerate the development of cutting-edge AI solutions.

Join our forum Explore GitHub discussions

Ready to get started?

C

CrewAI AMP Cloud

Manage the full AI agent lifecycle — build, test, deploy, and scale — with a visual editor and ready-to-use tools.

Request a demo
C

CrewAI AMP Factory

All the power of AMP Cloud, deployed securely on your own infrastructure — on-prem or private VPCs in AWS, Azure, or GCP.

Meet with us
C

CrewAI OSS

An open-source orchestration framework with high-level abstractions and low-level APIs for building complex, agent-driven workflows.

Read the docs View on GitHub Join the forum