API Reference

FreeLLM exposes an OpenAI-compatible API. Use any OpenAI SDK by setting base_url to your gateway address. All endpoints are available at both /v1/... (direct) and /api/v1/... (proxied via the dashboard).

Endpoints

Method	Endpoint	Description
POST	/v1/chat/completions	Chat completion (streaming and non-streaming)
GET	/v1/models	List all available models + meta-models
GET	/v1/status	Gateway health, provider states, per-key state, token usage, recent requests
POST	/v1/status/providers/{id}/reset	Force-reset a provider’s circuit breaker
PATCH	/v1/status/routing	Switch between round_robin and random routing
GET	/healthz	Simple health check (used by load balancers and Docker)

Chat completion

POST /v1/chat/completions
Content-Type: application/json

{
  "model": "free-fast",
  "messages": [
    {"role": "user", "content": "Hello!"}
  ],
  "stream": false,
  "temperature": 0.7,
  "max_tokens": 1024
}

Response shape:

{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "created": 1700000000,
  "model": "llama-3.3-70b-versatile",
  "choices": [
    {
      "index": 0,
      "message": { "role": "assistant", "content": "Hi there!" },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 5,
    "total_tokens": 15
  },
  "x_freellm_provider": "groq"
}

The x_freellm_provider field is FreeLLM-specific and tells you which upstream provider handled the request.

Status endpoint

GET /v1/status returns the full gateway state including provider health, per-key rate-limit windows, token usage totals, and recent requests. See Multi-Key Rotation and Token Usage Tracking for examples of the response shape.