Capability · Framework — agents

CrewAI

CrewAI models AI systems as a crew of role-playing agents working together on tasks. It has a simple YAML- and code-driven setup, a robust built-in tools library, support for sequential and hierarchical processes, and its own CrewAI Flows for event-driven workflows. It's one of the fastest-growing multi-agent frameworks in 2025-26.

Framework facts

Category: agents
Language: Python
License: MIT
Repository: https://github.com/crewAIInc/crewAI

Install

pip install crewai crewai-tools

Quickstart

from crewai import Agent, Task, Crew

researcher = Agent(
    role='Researcher',
    goal='Find key facts about {topic}',
    backstory='You are a diligent researcher.'
)
task = Task(
    description='Research {topic} and summarise in 3 bullets',
    expected_output='3-bullet summary',
    agent=researcher
)
crew = Crew(agents=[researcher], tasks=[task])
result = crew.kickoff(inputs={'topic': 'Mamba architectures'})

Alternatives

LangGraph — lower-level, more controllable
AutoGen — Microsoft's multi-agent chat framework
Agno — Python-native, higher performance
OpenAI Agents SDK — OpenAI's official alternative

Frequently asked questions

What's the difference between CrewAI and LangGraph?

CrewAI is higher-level and role-driven — you describe who the agents are and what they do. LangGraph is lower-level and state-driven — you describe the graph of steps and transitions. CrewAI is easier to start; LangGraph gives more control.

Does CrewAI support hierarchical teams?

Yes. You can set process=Process.hierarchical and assign a manager LLM that delegates tasks among agents. For complex workflows, CrewAI Flows adds event-driven orchestration.

Sources

CrewAI — docs — accessed 2026-04-20
CrewAI — GitHub — accessed 2026-04-20