Introduction

Fabrence serves open-weight models through an OpenAI-compatible API. Code that already works with OpenAI needs two changes: the base URL and the API key.

https://api.fabrence.com/v1

Billing is per token, with no minimums or contracts. We don’t store your prompts or outputs.

Where to start

• Quickstart: send your first request in cURL, Python or TypeScript.
• Authentication: create and use API keys.
• Chat completions: the full request and response reference.

What’s supported

• POST /v1/chat/completions, including streaming, tool calling and JSON output
• GET /v1/models
• Automatic prompt caching, billed at the lower cached input price

Any other endpoint returns 404. If you need one, email [email protected].