Skip to main content
Agno’s OpenAIChat model is OpenAI-compatible — point it at the Orbitrage gateway and every agent step is routed and traced.

​
Install

pip install -U orbitrage agno openai

​
Setup

import os, orbitrage
orbitrage.init(os.environ["ORBITRAGE_API_KEY"], user_id="customer_42")

from agno.agent import Agent
from agno.models.openai import OpenAIChat

agent = Agent(
    model=OpenAIChat(
        id="grok-4-fast",                            # direct model — or "claude-sonnet-4-6"; "auto" to route
        api_key=os.environ["ORBITRAGE_API_KEY"],
        base_url="https://api.orbitrage.ai/v1",
        default_headers={"x-orbitrage-end-user-id": "customer_42"},
    ),
    markdown=False,
)

print(agent.run("Summarize what an LLM router does, in one sentence.").content)

​
Managed tools (server-side, no keys)

Pass managed tools through Agno’s request_params — Orbitrage runs them server-side and loops the result back to the model. 
from agno.models.openai import OpenAIChat

model = OpenAIChat(
    id="grok-4-fast",
    api_key=os.environ["ORBITRAGE_API_KEY"],
    base_url="https://api.orbitrage.ai/v1",
    request_params={"tools": ["calculator_orbitrage", "tavily_orbitrage"]},
)
agent = Agent(model=model, markdown=False)
print(agent.run("Use the calculator to compute 1234*5678. Only the number.").content)  # 7006652
Pass your end-user id via default_headers={"x-orbitrage-end-user-id": ...} on the OpenAIChat model so per-user analytics line up. Agno’s own tools keep running on your side, right alongside managed tools.