Send your first request through Clyvel.

Create or select a Clyvel API key

Use the key generated for the environment or workspace that will own this traffic.

Point your OpenAI-compatible client at the Clyvel Gateway

Your model-provider account remains separate. Clyvel becomes the operational route for the request.

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.CLYVEL_API_KEY,
  baseURL: "https://api.clyvel.com/v1",
});

const response = await client.chat.completions.create({
  model: "gpt-5-mini",
  messages: [
    { role: "user", content: "Hello from Clyvel" }
  ],
});

Confirm the request in Clyvel

Open Requests and verify model, provider, latency, tokens, cost and status.

Add operational controls

Once traffic is visible, configure budgets, alerts, routing and governance according to the workload.