Inference v2

Synchronous inference endpoint (OpenAI-compatible path).

Submits an inference request and polls for results up to 30 seconds. If stream=true, returns SSE text/event-stream with token deltas.

post/v1/inference

Request

  • Base URL
    https://xcelsior.ca
  • URL
    https://xcelsior.ca/v1/inference
  • Auth
    HTTP bearer

Request body

modelstring required

Model name or HuggingFace repo

max_tokensinteger
temperaturenumber
streamboolean

Stream response via SSE

Response

Successful Response

unknown required

Changes