Skip to content
Pydantic Docs

Vercel AI SDK

The Vercel AI SDK can emit OpenTelemetry spans for model calls, tools, token usage, and streaming operations. Logfire can receive those spans through either @pydantic/logfire-node in Node.js scripts or @vercel/otel in Next.js applications.

Node.js Scripts

Terminal
npm install @pydantic/logfire-node ai @ai-sdk/openai

Replace @ai-sdk/openai with the provider package you use, such as @ai-sdk/anthropic or @ai-sdk/google.

Configure Logfire before importing the AI SDK:

instrumentation.ts
import * as logfire from '@pydantic/logfire-node'

logfire.configure({
  serviceName: 'ai-worker',
})

Enable telemetry on AI SDK calls:

import './instrumentation.ts'
import { openai } from '@ai-sdk/openai'
import { generateText } from 'ai'

const result = await generateText({
  model: openai('gpt-4.1-mini'),
  prompt: 'Write a short haiku about traces.',
  experimental_telemetry: { isEnabled: true },
})

console.log(result.text)

Next.js

In Next.js, configure @vercel/otel as shown in Next.js, then enable the same experimental_telemetry option on AI SDK calls. The instrumentation.ts file must live in the project root, or in src if your Next.js app uses src.

Terminal
npm install @vercel/otel @opentelemetry/api ai @ai-sdk/openai

Enabling Telemetry

The Vercel AI SDK emits OpenTelemetry spans when experimental_telemetry.isEnabled is set:

const result = await generateText({
  model,
  prompt: 'Write a short haiku about traces.',
  experimental_telemetry: { isEnabled: true },
})

This works with the AI SDK core functions that emit telemetry, including:

  • generateText and streamText
  • generateObject and streamObject
  • embed and embedMany

Example: Text Generation With Tools

import { openai } from '@ai-sdk/openai'
import { generateText, tool } from 'ai'
import { z } from 'zod'

const result = await generateText({
  model: openai('gpt-4.1-mini'),
  experimental_telemetry: { isEnabled: true },
  tools: {
    weather: tool({
      description: 'Get the weather in a location',
      inputSchema: z.object({
        location: z.string().describe('The location to get the weather for'),
      }),
      execute: async ({ location }) => ({
        location,
        temperature: 72,
      }),
    }),
  },
  prompt: 'What is the weather in San Francisco?',
})

console.log(result.text)

For Node.js scripts, import your Logfire instrumentation file before importing or calling the AI SDK.

What You Will See

When telemetry is enabled, Logfire captures a trace for each AI operation. Depending on the AI SDK provider and call type, traces can include:

  • parent spans such as ai.generateText
  • provider call spans for the model request
  • tool call spans
  • model and provider details
  • input and output token usage
  • timing information
  • tool call arguments and results
  • prompts and responses when the AI SDK emits them

Metadata

Use functionId and metadata to make traces easier to query:

await generateText({
  model,
  prompt,
  experimental_telemetry: {
    functionId: 'support-reply',
    isEnabled: true,
    metadata: {
      tenant: 'acme',
    },
  },
})

functionId appears in span names and helps distinguish use cases. metadata attaches custom key-value pairs to the emitted telemetry spans.