CrewAI

CrewAI is a cutting-edge framework for orchestrating role-playing, autonomous AI agents. By pointing CrewAI at Cencori, you can manage models centrally, enforce spend caps on expensive autonomous runs, and monitor every agent's thought process.

Quick Start

1. Deploy from the Marketplace

Navigate to Agents → Agent Marketplace in your project dashboard and click Deploy Agent on the CrewAI card.

2. Generate an API Key

Go to the agent's Configuration tab and click Generate Key. Copy the key.

3. Install CrewAI

pip install crewai

4. Configure Your Script

Set the standard OpenAI environment variables to point to Cencori before importing your CrewAI agents.

import os
 
# 1. Point CrewAI to Cencori
os.environ["OPENAI_API_BASE"] = "https://cencori.com/api/v1"
os.environ["OPENAI_API_KEY"] = "cake_YOUR_KEY_HERE"
 
# 2. Set the model name (must match a model in the Cencori dashboard)
os.environ["OPENAI_MODEL_NAME"] = "gpt-4o"
 
# 3. Create your crew as normal
from crewai import Agent, Task, Crew
 
researcher = Agent(
    role="Senior Research Analyst",
    goal="Uncover cutting-edge developments",
    backstory="You are an expert analyst at a leading tech think tank.",
    verbose=True,
    allow_delegation=False,
)
 
task = Task(
    description="Research the latest advancements in AI",
    expected_output="A comprehensive 3-paragraph report",
    agent=researcher
)
 
crew = Crew(
    agents=[researcher],
    tasks=[task],
    verbose=True
)
 
result = crew.kickoff()
print(result)

Using Specific Models per Agent

If you want different agents in your crew to use different models (e.g., a cheap gpt-4o-mini for basic research and a expensive claude-3-5-sonnet for writing), you can pass custom LLM instances to each agent.

See the CrewAI Custom LLM documentation for advanced configuration.

Features