Skip to content

Connect LlamaIndex to the AI Gateway

Send requests from your LlamaIndex workflows through the Logfire AI Gateway.

LlamaIndex is a framework for building LLM applications over your data. This includes retrieval-augmented generation (RAG), where an application retrieves relevant data before asking a model to answer. The example points LlamaIndex’s OpenAI client 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 LlamaIndex project with the packages imported below installed.

Configure LlamaIndex

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.

llamaindex-gateway.py
import asyncio
import os

from llama_index.core.workflow import StartEvent, StopEvent, Workflow, step
from llama_index.llms.openai import OpenAI


class WeatherWorkflow(Workflow):
    llm = OpenAI(
        model='gpt-5.4-mini',
        api_key=os.environ['LOGFIRE_GATEWAY_API_KEY'],
        api_base='https://gateway-us.pydantic.dev/proxy/openai',
    )

    @step
    async def answer(self, ev: StartEvent) -> StopEvent:
        response = await self.llm.acomplete(f'What is the weather in {ev.city}?')
        return StopEvent(result=str(response))


async def main() -> None:
    result = await WeatherWorkflow(timeout=60).run(city='London')
    print(result)


if __name__ == '__main__':
    asyncio.run(main())

Verify it worked

Run the example from your terminal. It prints the workflow result. 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.