Chat completions

POST /v1/chat/completions is OpenAI-compatible. Send a model and a messages array. Receive a choices payload with usage.

Request

POST https://www.velacompute.com/v1/chat/completions
{
  "model": "vela-4",
  "messages": [
    { "role": "system", "content": "You are concise." },
    { "role": "user", "content": "What is a meridian?" }
  ],
  "temperature": 0.7,
  "max_tokens": 400
}

Models

vela-4, vela-fast, and vela-mini accept this endpoint. Streaming is supported via "stream": true (SSE).

Response

{
  "id": "chatcmpl_...",
  "object": "chat.completion",
  "model": "vela-4",
  "choices": [
    {
      "index": 0,
      "message": { "role": "assistant", "content": "..." },
      "finish_reason": "stop"
    }
  ],
  "usage": { "prompt_tokens": 24, "completion_tokens": 80, "total_tokens": 104 }
}