Skip to main content
A runId identifies one hosted workflow execution. Preserve it exactly from an official-template, private-template, or Market submission response.

Create an official-template run from rows

The CLI separates preparation from execution:
run execute is the current command for new integrations. The older combined run submit flow is hidden.

List and inspect

run list supports --status, --page-size, --page-token, and ordering by createdAt or updatedAt in ascending or descending order.

Watch progress

Without --max-wait, watch continues until a terminal status. Recognized terminal statuses are completed, failed, partially_failed, and cancelled/canceled.

Read results

result-rows joins persisted input snapshots with row results and supports page tokens. result-workbook downloads the server-generated workbook containing original inputs and results.
Prefer result-workbook for workbook workflows. It is generated from the server-side input snapshot and result data.
Do not construct a console URL from a runId; no stable run-detail URL format is published. Use the CLI unless the service explicitly returns a URL.

Retry safely

After an ambiguous execution failure, query the run or related usage record before retrying. Reuse the original --client-request-id only for the identical confirmed payload. A changed input or new confirmation requires a new ID.