Create Select Run
Start a select-mode simulation run: test calls placed against library scenarios picked directly by id, rather than generated by an LLM (see Create Run).
Scenarios are written synchronously (no LLM generation phase, same shape as Create
Replay Run) and test calls are queued immediately, so generated_count reflects the
true selected count on the 202 response, not zero. Poll List Runs for progress.
Each picked scenario is scored against every PASS_FAIL eval currently configured on
the agent, same as a generated functional/adversarial scenario — there is no
run-level eval/intent narrowing in select mode, since each library scenario already
carries its own scoring config (target_ai_evaluation).
Atomic: every selected scenario id is validated before anything is created. If any single id does not resolve to an active scenario belonging to this agent, the entire request is rejected with a 404 and nothing is created.
Authorizations
Use Tuner API key (tr_api_...) or user session token. Find your API key in Workspace Settings > API Keys.
Path Parameters
Tuner's internal numeric ID for the agent.
Workspace ID. Find this in Workspace > General Settings.
Body
Request to launch a select-mode run (ENG-1541): test calls placed against library scenarios picked directly by id, rather than generated by an LLM.
Deliberately has no functional_count/adversarial_count/selected_eval_ids/
selected_intent_ids -- select mode has no "mix" to balance (each picked scenario is
already exactly what it is) and no run-level eval/intent narrowing (each library
scenario carries its own scoring config via target_ai_evaluation).
Library scenario.id values to run, in the order they should be queued.
1 - 20 elements1 <= x <= 12x > 0x > 0Allowed gains: 0.5, 1.0, or 1.5
Response
Successful Response
One selectable background environment.