Skip to main content
Get one complete, well-shaped trace into Lemma. The SDK sends JSON directly to the Lemma API.
Remember the one rule: one agent execution = one trace. Record model calls, tool calls, retrieval, and app logic inside that trace.

​
What you are building

support-agent              <- trace root (input, output)
|- draft-reply             <- generation (model, content)
`- search_docs             <- tool call (args, result)

​
Steps

1

Optional: install the Lemma tracing skill

If you want a coding agent to add tracing for you, install the Lemma skill from the public repo with the Skills CLI:
npx skills add uselemma/lemma --skill "lemma-tracing"
Then prompt your agent to use the Lemma tracing skill to add tracing to your application.
2

Install the SDK

npm install @uselemma/tracing
3

Set environment variables

Find these in Lemma project settings.
export LEMMA_API_KEY="lma_..."
export LEMMA_PROJECT_ID="proj_..."
4

Trace one agent execution

Create a client, open a trace, then record one-off generation and tool events inside the trace callback.
import { Lemma } from "@uselemma/tracing";

const lemma = new Lemma();

const answer = await lemma.trace(
  {
    name: "support-agent",
    input: userMessage,
    threadId: conversationId,
    userId: user.id,
  },
  async (trace) => {
    const docs = await searchDocs(userMessage);
    trace.recordTool({
      name: "search_docs",
      input: { query: userMessage },
      output: docs,
    });

    const response = await callModel(userMessage, docs);
    trace.recordGeneration({
      name: "draft-reply",
      input: response.messages,
      output: response.text,
      model: "gpt-4o",
    });

    return response.text;
  },
);
5

Verify in Lemma

Open the Lemma dashboard -> Traces. Within seconds you should see:
  • One trace for the whole execution.
  • The root trace with input and output.
  • A nested generation with model metadata.
  • A nested tool call with arguments and result.

​
What’s next

Trace contract

The exact shape Lemma reads.

Concepts

Traces, spans, generations, and threads.

Lemma tracing skill

Install the Lemma skill from the public repo.

Threads & context

Group conversations and attach users.

Instrument an agent

Build up a full trace with generations, tools, spans, and errors.

Vercel AI SDK

Record AI SDK model calls and tool executions.