Anthropic

wrapAnthropicSdk instruments an existing Anthropic client. Your call sites stay the same; each request is captured as an LLM event with model, prompts, output, token usage, duration, and status.

Quick setup

import Anthropic from "@anthropic-ai/sdk";
import { Carbon } from "@carbon-js/sdk";
import { wrapAnthropicSdk } from "@carbon-js/sdk/ai";

const carbon = new Carbon();
const anthropic = wrapAnthropicSdk(new Anthropic(), carbon);

const message = await anthropic.messages.create({
  model: "claude-haiku-4-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "What is the capital of France?" }],
});

wrapAnthropicSdk mutates the client you pass in and returns the same instance, typed to accept the optional carbon field on requests. Wrap the client once at startup and share the wrapped instance.

Deploying to a serverless platform — Next.js, Vercel, AWS Lambda? Flush buffered events before the function exits so none are lost. See Serverless.

What gets captured

Method	Captured as
`messages.create`	One LLM event
`messages.create` with `stream: true`	One LLM event, recorded after the stream completes
`messages.stream`	One LLM event, recorded after the stream completes

Attach metadata

Add a carbon field to any request to set a trace ID, context identifiers, or custom properties. The SDK strips the field before the request reaches Anthropic.

const traceId = carbon.createTraceId();

const message = await anthropic.messages.create({
  model: "claude-haiku-4-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Summarize this ticket." }],
  carbon: {
    traceId,
    context: { agentId: "support-agent", userId: "user-481" },
    additionalProperties: { release: "2026-06" },
  },
});

See Context for what each field powers in the dashboard.

Streaming

Streamed responses are captured after the stream finishes, so the stream must be fully consumed:

const stream = await anthropic.messages.create({
  model: "claude-haiku-4-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Write a haiku about ledgers." }],
  stream: true,
});

for await (const event of stream) {
  if (event.type === "content_block_delta" && event.delta.type === "text_delta") {
    process.stdout.write(event.delta.text);
  }
}

A stream that is abandoned before it completes produces no event.

Tool calls

The Anthropic Messages API does not execute tools. messages.create and messages.stream capture the LLM turn, including the tool_use blocks the model requests, but running those tools is your code’s job — so tool executions are always instrumented manually. Wrap each tool with wrapTool so its execution is captured. These methods don’t open a trace of their own, so create a trace ID with carbon.createTraceId() and pass it to both the model call and every tool call — that shared ID links the LLM event and its tool events into one trace:

const getWeather = carbon.wrapTool({
  name: "get_weather",
  run: async ({ city }: { city: string }) => fetchWeather(city),
});

const traceId = carbon.createTraceId();

const message = await anthropic.messages.create({
  model: "claude-haiku-4-5",
  max_tokens: 1024,
  messages,
  tools: [/* get_weather definition */],
  carbon: { traceId },
});

// Run each tool the model requested — each wrapped call is captured on the same trace.
for (const block of message.content) {
  if (block.type === "tool_use" && block.name === "get_weather") {
    await getWeather(block.input as { city: string }, { traceId });
  }
}

See Tool calls for the full wrapTool API.

Get started

Integrations

SDK

@carbon-js/connect

Concepts

Dashboard

Quick setup

What gets captured

Attach metadata

Streaming

Tool calls

​Quick setup

​What gets captured

​Attach metadata

​Streaming

​Tool calls

Quick setup

What gets captured

Attach metadata

Streaming

Tool calls