Vercel AI SDK + Logfire Demo

A minimal Next.js chat app instrumented with OpenTelemetry, sending traces to Pydantic Logfire.

Companion repo for the blog post: Vercel AI SDK and Logfire: what happens when everyone speaks OpenTelemetry

What this demonstrates

• @vercel/otel for OTel instrumentation in Next.js
• Vercel AI SDK's experimental_telemetry for LLM call tracing
• Streaming chat with streamText and useChat
• Standard OTLP export to Logfire (works with any OTLP-compatible backend)

Setup

1. Install dependencies:

npm install

2. Copy .env.local.example to .env.local and fill in your keys:

cp .env.local.example .env.local

You'll need:

• An OpenAI API key (OPENAI_API_KEY)
• A Logfire write token

3. Run the dev server:

npm run dev

Open http://localhost:3000 and send a message. Traces will appear in your Logfire project.

Project structure

src/
  instrumentation.ts      # OTel setup (3 lines)
  app/
    page.tsx              # Chat UI (useChat hook)
    api/chat/route.ts     # Streaming endpoint with experimental_telemetry
    layout.tsx            # Root layout