Start an Autoimprove job
Commission Roark on an objective. It clones the agent into a staging shadow (or uses your designated staging agent), authors a validation suite of simulated callers, measures a baseline, changes the staging configuration, and re-tests until the objective metric passes its target. Production is never touched by the loop; verified changes wait for promotion.
Requires an active provider integration (Vapi or Retell) with agent config writes enabled. One live job per agent: starting a second returns a conflict.
The job runs asynchronously; poll GET /v1/autoimprove/job/ or watch it in the dashboard. When its status is NEEDS_INPUT, answer via the answer endpoint; when AWAITING_PROMOTE, promote or dismiss.
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Body
The production agent to improve. It is never modified until you promote.
"b3b0c8e2-4c1d-4f6a-9e2b-1a2b3c4d5e6f"
The metric that defines success: a pass/fail metric, or a threshold variant of a scale metric (for example "PII Handling >= 4"). Roark measures the pass rate of this metric across simulated calls.
"f2f0c8e2-4c1d-4f6a-9e2b-1a2b3c4d5e6f"
Human-readable label for the objective, shown everywhere the job appears.
1 - 200"Consent collection should pass"
The pass-rate percentage that counts as fixed. Defaults to 90.
1 <= x <= 10090
An existing agent to stage changes on instead of the default shadow clone. Must be a different agent from agentId, on the same provider.
The provider integration whose credentials Roark uses. Omit to use the project's active integration for the agent's provider. The integration must have agent config writes enabled.
An existing simulation run plan to validate with. Omit to let Roark author its own suite.
Cap on decision turns. Defaults to 10.
1 <= x <= 50Cap on simulated calls dialed. Defaults to 200.
1 <= x <= 2000Response
The job was created and its loop started.
One Autoimprove job: Roark autonomously improving one agent toward one objective metric. Roark only ever changes the staging agent (a shadow clone by default); production changes exactly once, when verified changes are promoted.