List eval runs
Returns eval runs visible to the authenticated team context. suite_version is a per-suite ordinal and must be paired with suite_id; supplying it alone returns 400.
Authorizations
Gateway auth: send Authorization: Bearer <mka1-api-key>. For multi-user server-side integrations, you can also send X-On-Behalf-Of: <external-user-id>.
Headers
Optional external end-user identifier forwarded by the API gateway.
Query Parameters
Row cursor from a previous page's last_id. Page with a fixed filter set: the cursor row is resolved by id regardless of the other filters, so replaying one taken under different filters skips a row instead of resuming cleanly.
1 <= x <= 100Narrow to one suite's runs. Omit for every suite you can see; an empty value is rejected rather than treated as omitted, so a client that interpolates an undefined variable fails instead of silently widening the read.
1Narrow to runs recorded against one version of the suite. Version is a per-suite ordinal, so this requires suite_id on /evals/runs; on /evals/schedules/{schedule_id}/runs the schedule already pins the suite. Scores are only comparable within a version. A version the suite does not have returns an empty list, not 404, so it is indistinguishable from a version with no runs yet.
1 <= x <= 2147483647queued, in_progress, finalizing, completed, failed, cancelling, cancelled