Skip to main content
Corvex is an inference platform that exposes an OpenAI-compatible API in front of a self-hosted GPU cluster. You point your existing OpenAI client at our endpoint, swap your key, and you’re running.

​
What you get

  • OpenAI-compatible /v1/chat/completions and /v1/embeddings so your current SDKs work unchanged.
  • Governance built in — teams, virtual keys, and per-key rate limits.
  • A single source of truth for the API: the OpenAPI 3.1 spec that this site renders directly.

​
Where to start

Quickstart

Run your first inference request in under five minutes.

Authentication

Bearer tokens, virtual keys, and admin keys.

API Reference

Every endpoint, every parameter — generated from the spec.

Errors

HTTP status codes and what to do about them.
This documentation site is in Dogfooding. The Phase 1 surface (Get Started + API Reference) is live; the deeper Reference pages will fill in during Alpha.