Quickstart

Instrument an app, send a trace, and see it in bryel in under five minutes.

1. Get an API key

   Create a project in the bryel dashboard and copy its API key (bk_…). Set it as BRYEL_KEY in your environment.

   export BRYEL_KEY=bk_...
   
2. Install the SDK

   Pick the package for your stack.

   Vercel AI SDK

   bun add @bryel/vercel @vercel/otel
   
3. Register the processor

   Add the bryel span processor where you set up OpenTelemetry.

   Vercel AI SDK

   In instrumentation.ts:

   import { registerOTel } from "@vercel/otel";
   import { bryelVercelProcessor } from "@bryel/vercel";
   
   export function register() {
     registerOTel({
       serviceName: "my-app",
       spanProcessors: [bryelVercelProcessor({ apiKey: process.env.BRYEL_KEY! })],
     });
   }
   

   Any OTel app

   Wire the core processor into your tracer setup. See Any OpenTelemetry app for the full configuration.

4. Enable telemetry per call

   Turn on experimental_telemetry for each generation you want traced.

   const result = streamText({
     model: openai("gpt-5"),
     prompt,
     experimental_telemetry: {
       isEnabled: true,
       metadata: { sessionId, userId }, // groups a conversation in bryel
     },
   });
   
5. See it in bryel

   Run your app, then open the dashboard. You’ll see the run: streamed turns, tool calls, tokens, and cost.

   A streamed span finalizes only once the stream is fully consumed — always drain it.

Next steps

Instrumentation

Set up bryel for your framework — Vercel AI SDK, OpenTelemetry, or anything else.

Concepts

Understand traces, sessions, and turns, and how bryel models an agent run.