Skip to content
SideSeat

Vercel AI SDK

Vercel AI SDK is a TypeScript toolkit for building AI applications. SideSeat captures traces from generateText, streamText, and other AI SDK calls via OpenTelemetry.

Prerequisites

Section titled “Prerequisites”
  • SideSeat running locally (sideseat)
  • Node.js 18+
  • AWS credentials configured for Amazon Bedrock, or another supported provider

Quick Start

Section titled “Quick Start”

  1. Start SideSeat

    Terminal window
    npx sideseat

  2. Install dependencies

    Terminal window
    npm install @sideseat/sdk ai @ai-sdk/otel @ai-sdk/amazon-bedrock

  3. Add telemetry

    import { init, Frameworks } from '@sideseat/sdk';
    import { generateText, registerTelemetry } from 'ai';
    import { LegacyOpenTelemetry } from '@ai-sdk/otel';
    import { bedrock } from '@ai-sdk/amazon-bedrock';
    

    init({ framework: Frameworks.VercelAI });
    

    // AI SDK 7 delivers telemetry only to registered integrations - it no longer
    // creates OpenTelemetry spans itself. Register after init(): the integration
    // captures a tracer in its constructor.
    registerTelemetry(new LegacyOpenTelemetry());
    

    const { text } = await generateText({
      model: bedrock('us.anthropic.claude-sonnet-4-5-20250929-v1:0'),
      prompt: 'What is the capital of France?',
      experimental_telemetry: { isEnabled: true },
    });
    

    console.log(text);

  4. View runs

    Open http://localhost:5388 to see your runs.

Without SideSeat SDK

Section titled “Without SideSeat SDK”

  1. Start SideSeat

    Terminal window
    npx sideseat

  2. Set the endpoint

    Terminal window
    export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:5388/otel/default

  3. Install dependencies

    Terminal window
    npm install @opentelemetry/sdk-node @opentelemetry/exporter-trace-otlp-http ai @ai-sdk/otel @ai-sdk/amazon-bedrock

  4. Add telemetry

    import { NodeSDK } from '@opentelemetry/sdk-node';
    import { OTLPTraceExporter } from '@opentelemetry/exporter-trace-otlp-http';
    import { generateText, registerTelemetry } from 'ai';
    import { LegacyOpenTelemetry } from '@ai-sdk/otel';
    import { bedrock } from '@ai-sdk/amazon-bedrock';
    

    const sdk = new NodeSDK({ traceExporter: new OTLPTraceExporter() });
    sdk.start();
    

    // Required on AI SDK 7: the ai package emits no spans itself. Must come after
    // sdk.start(), since the integration captures a tracer in its constructor.
    registerTelemetry(new LegacyOpenTelemetry());
    

    const { text } = await generateText({
      model: bedrock('us.anthropic.claude-sonnet-4-5-20250929-v1:0'),
      prompt: 'What is the capital of France?',
      experimental_telemetry: { isEnabled: true },
    });
    

    console.log(text);

  5. View runs

    Open http://localhost:5388 to see your runs.

What You’ll See

Section titled “What You’ll See”

SideSeat shows each generateText/streamText call with the model name, token counts, prompt messages, and the full response. Tool calls appear as child spans.

Streaming

Section titled “Streaming”

These snippets assume registerTelemetry(new LegacyOpenTelemetry()) has already run once at startup, as in the Quick Start above. Without it isEnabled: true emits nothing.

import { init, Frameworks } from '@sideseat/sdk';
import { streamText } from 'ai';
import { bedrock } from '@ai-sdk/amazon-bedrock';


init({ framework: Frameworks.VercelAI });


const result = streamText({
  model: bedrock('us.anthropic.claude-sonnet-4-5-20250929-v1:0'),
  prompt: 'Tell me a story',
  experimental_telemetry: { isEnabled: true },
});


for await (const chunk of result.textStream) {
  process.stdout.write(chunk);
}

Tool Calls

Section titled “Tool Calls”

These snippets assume registerTelemetry(new LegacyOpenTelemetry()) has already run once at startup, as in the Quick Start above. Without it isEnabled: true emits nothing.

import { init, Frameworks } from '@sideseat/sdk';
import { generateText, tool } from 'ai';
import { bedrock } from '@ai-sdk/amazon-bedrock';
import { z } from 'zod';


init({ framework: Frameworks.VercelAI });


const { text } = await generateText({
  model: bedrock('us.anthropic.claude-sonnet-4-5-20250929-v1:0'),
  prompt: 'What is the weather in Paris?',
  tools: {
    getWeather: tool({
      description: 'Get weather for a location',
      inputSchema: z.object({ location: z.string() }),
      execute: async ({ location }) => `Sunny in ${location}`,
    }),
  },
  experimental_telemetry: { isEnabled: true },
});

Next Steps

Section titled “Next Steps”