Use Claude with the OpenAI SDK

1. Configure the OpenAI SDK

Use the Virouter OpenAI-compatible endpoint as your SDK base URL. Store the key server-side in VIROUTER_API_KEY.

import OpenAI from "openai";

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

2. Send a Claude chat completion

Keep the OpenAI chat completions method and pass a Claude model id. Virouter routes the request to the Anthropic lane behind the scenes.

const response = await client.chat.completions.create({
  model: "claude-opus-4-7",
  messages: [
    { role: "system", content: "You are concise and practical." },
    { role: "user", content: "Review this API design." },
  ],
});

console.log(response.choices[0]?.message?.content);

3. Switch between GPT and Claude

Your app can use the same client and wallet for both provider families. Change only the model field.

await client.chat.completions.create({
  model: "gpt-5.5",
  messages,
});

await client.chat.completions.create({
  model: "claude-sonnet-4-6",
  messages,
});

Claude model ids

Available models depend on active provider channels and your key allowlist. Use GET /v1/models for the current list.

Response cost and token headers

Claude calls through the OpenAI SDK still return Virouter gateway headers for cost and quota tracking.

Troubleshooting