List feedback for an agent
Every score the agent has earned, across all its objectives, in one flat stream.
score on [-1.0, 1.0], an optional comment, and info.submittedBy, the profile that left it.
It is flat: no per-variation attribution
This is the one thing to know. An entry tells you the score, the note, and who left it, but not which objective or variation earned it. There is noobjectiveId or variationId on the row. So this endpoint answers “how is this agent doing overall?”, not “which variation is winning?”.
For attribution, reach for a different read:
- Per variation: List variations returns each variation’s aggregate
scoreandfeedbackCount, which is what drives weighted selection. - Per objective: List objective feedback returns the scores on one specific run.
Search the comments
Thequery param does a free-text search over the comment text, so you can pull the feedback that mentions a theme:
limit and cursor page the results the usual way.
Related
Submit feedback
List variations
score this flat stream cannot give you.List objective feedback
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Path Parameters
"workspace_01HXKD2E5NQM3T9AYWCF133E3Q"
The ID of the agent. Supports "external_id:" prefix for external IDs.
"agent_01HXKD2E5NQM3T9AYWCFMGWT9Y"
Query Parameters
Maximum number of results to return.
Pagination cursor from previous response.
Free-text search applied to the feedback comment. Case-insensitive substring match.
Filter by sentiment. UNSPECIFIED returns feedback regardless of score.
FEEDBACK_SENTIMENT_UNSPECIFIED, FEEDBACK_SENTIMENT_POSITIVE, FEEDBACK_SENTIMENT_NEGATIVE Optional filter to limit results to feedback on objectives run by a single agent variation. Supports "external_id:" prefix for external IDs.
Inclusive lower bound on feedback creation time.
Exclusive upper bound on feedback creation time.
Filters by metadata labels. Comma-separated key=value pairs, e.g. "env=prod,team=ai". A resource matches only if every pair matches exactly (AND semantics).
When set to true you may use more of your alloted API rate-limit