Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

API overview

Note

To be written. The per-endpoint detail on this page is deliberately thin: it will be generated from the control plane’s OpenAPI contract rather than written by hand, so that it cannot drift. See Endpoint reference below.

Thatch’s customer-facing API is OpenAI-compatible. If you already have code talking to an OpenAI-shaped endpoint, you point it at Thatch by changing the base URL and the key.

Base URL

https://api.thatch.cloud/v1

Authentication

A bearer API key, issued from the Portal and prefixed thatch_sk_:

curl https://api.thatch.cloud/v1/models \
  -H "Authorization: Bearer $THATCH_API_KEY"

A missing or invalid credential is a 401; a valid credential that is not permitted on the route is a 403. See Authentication and Errors.

The surface

EndpointPurpose
POST /v1/chat/completionsSubmit a completion request and get a response
GET /v1/modelsThe model catalogue available to your key
GET /v1/usageYour own usage, scoped to the calling identity

All three endpoints are live.

Usage figures here are scoped to the caller. Account-wide billing lives in the Portal — see Usage & metering.

Endpoint reference

The per-endpoint reference is generated from the control plane’s own OpenAPI contract, filtered to the operations that are public, and committed to this repository so it cannot silently disagree with what the API serves. The tooling is in place; it is waiting on the control plane to emit a contract for the serving surface. Until then, this page is the surface.

Versioning