Skip to content

Model provider

Anthropic + Cloudmind

Trace Claude API calls including tool use and extended thinking, with the same wrapper pattern as every other Cloudmind integration.

How it works

The Anthropic integration covers the Messages API. Tool use blocks are traced as child spans, so a multi-turn tool conversation reads as a tree rather than a flat log.

Extended thinking blocks are captured separately from the visible response, which matters when you are debugging why a model reached a particular conclusion.

Setup

import Anthropic from "@anthropic-ai/sdk";
import { monitorAnthropic } from "@cloudmind-ai/js/anthropic";

const anthropic = monitorAnthropic(new Anthropic());

const result = await anthropic.messages.create({
  model: "claude-sonnet-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello" }],
});

Supported

  • Messages API

  • Streaming

  • Tool use

  • Extended thinking

  • Prompt caching metrics

  • Batch API

Stop guessing whether your AI feature got better

Free for 50,000 traces a month. No credit card, no sales call, five minutes to your first trace.

Start free