Runs

Trigger and inspect executions via the API.

The execution object#

{
  "id": "uuid",
  "workflow_id": "uuid",
  "status": "success | running | failed | cancelled",
  "trigger_type": "manual | schedule | webhook | ...",
  "input_data": { ... },
  "output_data": { ... },
  "started_at": "...",
  "finished_at": "...",
  "logs": [ { node, status, output, duration_ms }, ... ]
}

List runs#

# Runs for one workflow
GET /api/v1/executions/?workflow_id={uuid}
 
# Runs across the workspace (paginated + filterable)
GET /api/v1/executions/all?limit=50&offset=0&status=failed
# → { executions, total, limit, offset }

Run detail#

Fetch one execution including its full logs array — the same data the run inspector renders.

GET /api/v1/executions/{execution_id}

Rerun, cancel, resume#

POST /api/v1/executions/{id}/rerun    # replay with the original input
POST /api/v1/executions/{id}/cancel   # stop a running execution
POST /api/v1/executions/{id}/resume   # continue a paused run
{ "token": "<resume-token>", "input": { ... } }

Resume is used by human-in-the-loop steps: a paused run emits a one-time token; posting it back (no other auth required — the token is the secret) continues the run with the supplied input.