CrewAI is an open-source framework for orchestrating multi-agent AI workflows. It enables task automation and decision-making by coordinating teams of AI agents. Since CrewAI supports OpenAI-compatible endpoints, you can configure it to use MARA Cloud models.
Prerequisites
- Python 3.10+
- A MARA Cloud API key. See API Keys and URLs to generate one.
Setup
Install CrewAI:
bash
pip install crewaiConfiguration
Define an
LLM object directly, then pass it to your agents:python
from crewai import Agent, Task, Crew, LLM
llm = LLM(
base_url="https://api.cloud.mara.com/v1",
api_key="your-api-key",
model="MiniMax-M2.5",
)
researcher = Agent(
role="Researcher",
goal="Find and summarize the latest AI trends",
backstory="You are an experienced AI researcher.",
llm=llm,
)
task = Task(
description="Research and summarize the top 3 AI trends in 2025.",
expected_output="A concise summary of the top 3 AI trends.",
agent=researcher,
)
crew = Crew(
agents=[researcher],
tasks=[task],
)
result = crew.kickoff()
print(result)Build with CLI template
You can also initialize a new CrewAI project using the CLI and select your MARA Cloud model:
bash
crewai create crew my_new_crewLearn more
- Model Catalog - Browse all available models.
- CrewAI Documentation - Official CrewAI docs.