|

Agent Frameworks

Power multi-agent systems like CrewAI and AutoGen with Cencori.

Agent frameworks often require powerful models (like GPT-4) but can quickly rack up costs and hit rate limits. Cencori solves this by providing caching, rate limiting, and failover for your agent swarms.

CrewAI

Configure CrewAI agents to route through Cencori by setting the OPENAI_API_BASE environment variable.

Configuration

import os
from crewai import Agent, Task, Crew
 
# 1. Point to Cencori
os.environ["OPENAI_API_BASE"] = "https://api.cencori.com/v1"
os.environ["OPENAI_API_KEY"] = "csk_..."
os.environ["OPENAI_MODEL_NAME"] = "gpt-4o"
 
# 2. Define Agents
researcher = Agent(
  role='Researcher',
  goal='Discover new AI trends',
  backstory="You are a senior analyst.",
  verbose=True
)
 
# 3. Create Crew
crew = Crew(
  agents=[researcher],
  tasks=[...],
  verbose=2
)

AutoGen

Microsoft AutoGen uses a config list to manage models. You can point this to Cencori.

OAI_CONFIG_LIST

[
    {
        "model": "gpt-4o",
        "api_key": "csk_...",
        "base_url": "https://api.cencori.com/v1"
    },
    {
        "model": "claude-sonnet-4.5",
        "api_key": "csk_...",
        "base_url": "https://api.cencori.com/v1"
    }
]

Python Setup

import autogen
 
config_list = autogen.config_list_from_json(
    "OAI_CONFIG_LIST"
)
 
assistant = autogen.AssistantAgent(
    name="assistant",
    llm_config={"config_list": config_list}
)
 
user_proxy = autogen.UserProxyAgent(
    name="user_proxy",
    human_input_mode="TERMINATE"
)
 
user_proxy.initiate_chat(
    assistant,
    message="Planning a trip to Tokyo."
)

[!NOTE] Unified Billing Even if your agents use different models (GPT-4 vs Claude), Cencori unifies the billing into a single invoice.

Did you like the content?