One API key. 340+ models. Zero lock-in.

Most people sign up to Relay for the model access, and they stay for the infrastructure.

Automatic fallbacks.

If a provider goes down, Relay reroutes the request to another provider serving the same model, so your users see zero downtime.

Fallback docs

Drop-in SDK compatibility.

Swap the base URL and the API key and your existing code keeps working: any client that speaks the completions format.

Quickstart

Provider routing.

Pin a specific provider for compliance, or let Relay pick the fastest or cheapest option automatically.

Provider routing docs

One request, every model

curl relay.norvane.ai/v1/chat/completions \
  -H "Authorization: Bearer $NORVANE_API_KEY" \
  -d '{"model": "halcyon/tessera-3",
    "messages": [{"role": "user", "content": "Hi"}]}'

Change the model id and the same request reaches any of the 340+ models in the catalog.

Get up and running

Create an API key.
Make a first call, either in the playground or from the curl and SDK examples.
Add credits to unlock every model in the catalog.

Full documentation

New model drops, feature announcements, and help from the team land in the Norvane community.

The Norvane Relay Team.

Norvane Labs, 880 Kestrel Way, San Francisco, CA 94107

Privacy · Terms · Unsubscribe