Skip to main content

Rate an assistant message

POST 
/api/v1/llm/gateway/rate-message

Records the caller's rating on an assistant-generated message, addressed by its per-conversation message_sequence. Two rating styles are available via kind: a binary thumb (FEEDBACK_KIND_THUMB + thumb_up) or a 1–10 scale (FEEDBACK_KIND_SCALE + rating).

The rater is derived server-side from the verified request headers — there is deliberately no "who is rating" parameter. One rating is kept per rater per message, so re-rating the same message replaces the caller's previous entry and returns is_update: true.

The response carries the full message with the complete (all-rater) feedback list applied.

Request​

Responses​

Rating recorded