Basic agent - Lemma

Use this for the standard case: an agent function that calls an LLM and returns a single response.

TypeScript
Python

import { registerOTel, agent } from "@uselemma/tracing";

registerOTel();

const wrapped = agent("my-agent", async (input: { userMessage: string }) => {
  const response = await callLLM(input.userMessage);
  return response; // wrapper auto-captures output and closes the span
});

const { result, runId } = await wrapped({ userMessage });

from uselemma_tracing import register_otel, TraceContext, agent

register_otel()

async def run_agent(user_message: str, ctx: TraceContext) -> str:
    response = await call_llm(user_message)
    return response  # wrapper auto-captures output and closes the span

wrapped = agent("my-agent", run_agent)

result, run_id, _ = await wrapped(user_message)

To add per-call LLM visibility (prompt, response, tokens), register the matching OpenInference instrumentor for your provider at startup. Details: Adding provider instrumentation.

Recipes