Documentation

Faucet is an OpenAI-compatible gateway in front of every model we carry. One key, one base URL, no SDK to install.

Faucet puts one API key and one base URL in front of every model in the catalog. The endpoint speaks the OpenAI wire format, so the client library you already use already speaks to Faucet — switching models is a string change rather than a second integration.

Where to start

Getting started

Get an organization and a key, then make your first metered request. About five minutes.

Guides

Task-oriented walkthroughs: SDKs, streaming, chat, embeddings, images, transcription, errors and limits.

Models

Every model we carry, with its provider, capabilities, context window and reference rates.

Reference

Look-up material for what each endpoint accepts and returns.

Two answers this site cannot give

Both are worth knowing before you read anything else here, because both are ways a static page could mislead you.

GET /v1/models is authoritative for your key. It applies that key's allowlist and skips providers holding no usable credential. This site is compiled ahead of time and can see neither.
The console is authoritative for prices. Rates published here are catalog reference figures; context-tiered, deployment-tiered and surcharged models can bill above them.

Getting access

Faucet is in closed alpha, so organizations are created by hand rather than signed up for. Email keunwoopark.524@gmail.com with a line about what you are building.