Skip to main content
Start here before you send chat, embeddings, or audio requests. These endpoints help you verify the gateway is healthy and discover the aliases you should pass as model.

​
GET /health

Checks gateway health and the status of its dependencies. Auth: none

​
Response

The gateway returns 200 when healthy and 503 when one or more required services are unavailable. 
{
  "ok": true,
  "services": {
    "database": true,
    "litellm": true,
    "deepgram": true
  }
}
If ok is false, expect HTTP 503.

​
GET /v1/models

Returns the model aliases you should use in model fields across the public API. Auth: none

​
Response

{
  "object": "list",
  "data": [
    {
      "id": "basics-chat-smart",
      "object": "model",
      "owned_by": "basics",
      "capabilities": ["chat", "streaming"]
    }
  ]
}

​
How to use model aliases

Use the id values returned here as the model in other requests, for example:
  • basics-chat-smart
  • basics-embed-small
  • basics-stt
  • basics-tts

​
Tips

  • Call this endpoint to discover active aliases instead of hard-coding assumptions
  • If you send an unknown alias to another endpoint, you may get 404 with code: "model_not_found"