Run brain insights against an evaluation transcript
POST/v1/brains/:brain_id/simulations/runs/:run_id/evaluations/:evaluation_id/insights
Permissions - brains.write. API keys - manage. Runs every insight
configured on the brain against the evaluation's transcript via the
Dialog service. The transcript is loaded from analytics (same path used
by PUT /brains/{brain_id}/insights/test). Per-insight gRPC failures
surface in the row's error field rather than aborting the batch
(Promise.allSettled). Result is not persisted; the client decides what
to keep. Returns [] when the brain has no insights configured.
Request
Responses
- 200
- 400
- 401
- 404
- 500
One result row per configured insight
Evaluation has no session yet (no-session) or brain is not trained (brain_not_trained).
Unauthorized
Response Headers
Unique identifier for the API request, useful for debugging and tracking.
abc123xyz789Evaluation not found, or analytics transcript missing for the session.
Internal Server Error
Response Headers
Unique identifier for the API request, useful for debugging and tracking.
abc123xyz789