From OpenRouter

If you're on OpenRouter, the wire format is familiar — both speak OpenAI. The differences are about positioning, not protocol.

What's different

Minimal diff

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.OPENROUTER_API_KEY!,
  baseURL: 'https://openrouter.ai/api/v1',
});

const res = await client.chat.completions.create({
  model: 'mistralai/mistral-large',
  messages: [{ role: 'user', content: 'Hi.' }],
});

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.CLEVERROUTER_API_KEY!,
  baseURL: 'https://api.cleverouter.eu/v1',
});

const res = await client.chat.completions.create({
  model: 'mistral/mistral-large-2',
  messages: [{ role: 'user', content: 'Hi.' }],
});

Two real changes: the baseURL / key, and the model id format (mistral/mistral-large-2 — no ai suffix in the family).

Routing differences

OpenRouter encodes routing in the model id (:floor, :auto, :nitro). CleverRouter uses HTTP headers — the model id stays clean:

defaultHeaders: {
  'X-CleverRouter-Strategy': 'cheapest', // or 'fastest' | 'best' | 'default'
  'X-CleverRouter-Provider': 'scaleway', // optional explicit pin
}

See Strategies.

Provider transparency

Every response carries the provider that ran the request:

X-CleverRouter-Provider: scaleway

You always know whether your call went to Scaleway, Tensorix or Bedrock — useful for compliance reviews.

What you can't bring over

• Models hosted only outside the EU (Anthropic via US, GPT-4 via US, etc.). CleverRouter has no US fallback by design.
• The :floor / :nitro model-id suffixes — use strategy: 'cheapest' or 'fastest' instead.

Related

• Strategies
• Provider pinning
• Why EU-native