Skip to main content
These operations measure the quality of agents and their output: machine judges with their rubrics, verdicts and drift alerts, and evaluation sets you run and compare. Every path here is under /api/v3; see authentication and permissions for the credential each operation takes.

Domains

Judges

Rubrics, verdicts, drift alerts, overrides and re-judging.

Evals

Evaluation sets, metered runs and result comparison.

Common tasks