Skip to content

Connect Smolagents to the AI Gateway

Send requests from your Smolagents agents through the Logfire AI Gateway.

Smolagents is Hugging Face’s Python library for building AI agents that call tools and run multi-step tasks. The example points OpenAIServerModel at an OpenAI-compatible gateway route.

Before you start

  • Complete the AI Gateway prerequisites, including setting LOGFIRE_GATEWAY_API_KEY in your terminal.
  • Use an existing Smolagents project. Install the OpenAI extra with pip install 'smolagents[openai]'.

Configure Smolagents

Set api_key to your gateway key and api_base to the OpenAI-compatible gateway route. Copy the route and a supported model name from the Gateway Connect tab.

smolagents-gateway.py
import os

from smolagents import OpenAIServerModel, ToolCallingAgent

model = OpenAIServerModel(
    model_id='gpt-5.4-mini',
    api_base='https://gateway-us.pydantic.dev/proxy/openai',
    api_key=os.environ['LOGFIRE_GATEWAY_API_KEY'],
)
agent = ToolCallingAgent(tools=[], model=model)

print(agent.run('What is the weather in London?'))

Verify it worked

Run the example from your terminal. It prints the agent’s response. That confirms the client reached the gateway. Organization admins can also open AI Engineering > Gateway > Spending to see usage for the key. If telemetry is enabled, open the selected project’s Live view to inspect the request trace.

Troubleshooting

  • Python raises KeyError: 'LOGFIRE_GATEWAY_API_KEY': set the environment variable in the same terminal where you run the example.
  • The request returns an authentication or model error: copy the URL and model name again from the Gateway Connect tab, and confirm that the selected route supports the OpenAI request format.