Anthropic

Import name: Sentry.anthropicAIIntegration

The anthropicAIIntegration adds instrumentation for the @anthropic-ai/sdk to capture spans by automatically wrapping Anthropic client calls and recording LLM interactions with configurable input/output recording.

import * as Sentry from "@sentry/cloudflare";
import Anthropic from "@anthropic-ai/sdk";

const anthropic = new Anthropic();
const client = Sentry.instrumentAnthropicClient(anthropic, {
  recordInputs: true,
  recordOutputs: true,
});

// Use the wrapped client instead of the original anthropic instance
const response = await client.messages.create({
  model: "claude-3-5-sonnet-20241022",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello!" }],
});

Options

recordInputs

Type: boolean

Records inputs to Anthropic API method calls (such as prompts and messages).

Defaults to true if sendDefaultPii is true.

Sentry.init({
  integrations: [Sentry.anthropicAIIntegration({ recordInputs: true })],
});

recordOutputs

Type: boolean

Records outputs from Anthropic API method calls (such as generated text and responses).

Defaults to true if sendDefaultPii is true.

Sentry.init({
  integrations: [Sentry.anthropicAIIntegration({ recordOutputs: true })],
});

Configuration

By default this integration adds tracing support to Anthropic API method calls including:

• messages.create() - Create messages with Claude models
• messages.stream() - Stream messages with Claude models
• messages.countTokens() - Count tokens for messages
• models.get() - Get model information
• completions.create() - Create completions (legacy)
• models.retrieve() - Retrieve model details
• beta.messages.create() - Beta messages API

The integration will automatically detect streaming vs non-streaming requests and handle them appropriately.

Supported Versions

• @anthropic-ai/sdk: >=0.19.2 <1.0.0