API Reference API
API Overview
Public endpoints, auth, and maintenance behavior for the LM0 gateway.
LM0 exposes a unified public surface under /v1/*, plus a health check.
Base URL
https://api.17.wtf
Public endpoints
| Endpoint | Compatibility | Auth | Maintenance pause |
|---|---|---|---|
GET /healthz |
LM0 native | No | No |
GET /v1/models |
OpenAI or Anthropic shaped | Yes | No |
GET /v1/models/* |
OpenAI or Anthropic shaped | Yes | No |
POST /v1/chat/completions |
OpenAI Chat Completions | Yes | Yes |
POST /v1/responses |
OpenAI Responses (stateless subset) | Yes | Yes |
POST /v1/messages |
Anthropic Messages | Yes | Yes |
POST /v1/messages/count_tokens |
Anthropic count_tokens | Yes | No |
Authentication
Either header style works on /v1/*:
- OpenAI:
Authorization: Bearer <api-key> - Anthropic:
x-api-key: <api-key>(and typicallyanthropic-version: 2023-06-01)
See Authentication for details.
Maintenance pause
When API maintenance mode is enabled, generation endpoints are paused:
POST /v1/chat/completionsPOST /v1/responsesPOST /v1/messages
Model listing, token counting, dashboard, and admin APIs are not blocked by maintenance pause.
How requests flow
- Authenticate the key
- Translate inbound wire format → canonical request
- Authorize model / pool access
- Route to upstream (retry / fallback as configured)
- Translate canonical response → client wire format
- Record usage (DB-backed production mode)
Read more in Architecture.