Moving from OpenRouter

Change the base URL and the key. Your model names keep working.

Bring your OpenRouter key

On the Keys page, connect your company's OpenRouter key: your requests are then billed to your OpenRouter account, as today. You can also connect providers' own keys next to it; their models then go straight to them.

Change two settings

from openai import OpenAI

client = OpenAI(base_url="https://sov.solibra.ai/v1", api_key="sov_live_…")
reply = client.chat.completions.create(
    model="anthropic/claude-sonnet-5.5",   # OpenRouter names work as they are
    messages=[{"role": "user", "content": "Hello"}],
)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://sov.solibra.ai/v1", apiKey: "sov_live_…" });
const reply = await client.chat.completions.create({
  model: "anthropic/claude-sonnet-5.5", // OpenRouter names work as they are
  messages: [{ role: "user", content: "Hello" }],
});

Model names

OpenRouter's names, like anthropic/claude-sonnet-5.5 or openai/gpt-6.1-sol, mean the same model as Sovereign's own ids, like claude-sonnet-5-5. GET /v1/models lists every model your company can use.

What Sovereign does instead

• Model fallbacks (models) and routing (route): leave them out. Sovereign chooses the model, and if the chosen model fails, it answers on the model you asked for. These fields are ignored.
• Provider settings (provider): for OpenAI and Anthropic models Sovereign always uses the first-party provider.
• Your Anthropic-format clients also work: point them at https://sov.solibra.ai instead of OpenRouter's Anthropic endpoint.