Skip to main content

Record a dataset run

POST 
/api/v1/evals/record-dataset-run

Records the result of executing a dataset — the motivating case for this API. A harness runs a dataset, scores each item, and posts the whole run here in one call.

profileRevisionHash ties the run to the exact agent-profile revision under test, so two runs can be compared meaningfully.

Request​

Responses​

OK