Skip to main content
POST
Start a run

Authorizations

Authorization
string
header
required

An API key, sent as Authorization: Bearer stfy_live_…. See Authentication for how to get one and which scopes it needs.

Body

application/json
case_set_id
string
required

The case set, from POST /v1/evaluations/case_sets.

changes
WorkerChanges · object | null

Run with these settings instead of the worker's saved ones, as PATCH /v1/digital_workers/{worker_id} takes them (without status). Like config, never written to the worker.

config
Config · object | null

Run with this configuration INSTEAD of the worker's saved one. It is recorded against the run and never written to the worker, so an experiment cannot change live behaviour. Omit it to run the worker as it stands.

label
string | null

Shown against the run, e.g. 'warmer opening'.

Maximum string length: 200
version_id
string | null

Run one of the worker's saved versions, by id from GET /v1/digital_workers/{worker_id}/versions.

Maximum string length: 64

Response

201 - application/json

Successful Response

case_set_id
string
required
id
string
required
status
string
required

queued, running, done or failed.

worker_id
string
required
config_snapshot
Config Snapshot · object | null
error
string | null

Why the run as a whole failed. Per-case failures sit on the result.

finished_at
string | null
judged_by
string | null
label
string | null
started_at
string | null
started_by
string | null
totals
Totals · object | null

Counts over the run's results: cases, how many are judged, how many passed, how many could not run at all, and the mean score. They follow the verdicts, so a run that nobody has scored yet reports nothing judged.

untested
string[] | null

On the response that starts a run with changes: the parts of them a run never reads (channels, autonomy, accounts, assignment). Not stored: null on a later read.