List agent tasks
Tasks this brand has run, newest first. A `build_page` task reports the state of the page build it wraps, so its `status` is always what the opportunity itself says. A task that has been running for more than two hours with no word from the agent is reported as `failed`.
Authorization
tenantApiKeyAuth In: header
Path Parameters
Query Parameters
1-50, default 20.
Rows to skip.
Only tasks in this state.
Value in
- "failed"
- "running"
- "succeeded"
Response Body
application/json
application/json
application/json
application/json
application/json
application/json
application/json
application/json
curl -X GET "https://example.com/api/v1/public/brands/string/tasks/"{ "total": 0, "offset": 0, "limit": 0, "note": "string", "items": [ { "taskId": "e6e9d88a-9b63-468a-aec3-b7a11de27af8", "command": "prompt", "status": "running", "createdAt": "2019-08-24T14:15:22Z", "finishedAt": "2019-08-24T14:15:22Z", "resultingRoute": "string", "chatUrl": "string", "error": "string" } ]}Last updated on
Get a page (or brand-wide) report
How a page is performing, projected to the `metrics` you ask for. Each one is a separate upstream read, so ask for what you need: - `score`: quality and citability grades plus the ranked priority fixes - `performance`: Search Console clicks/impressions/CTR/position and AI-assistant referral clicks by provider - `traffic`: visitor sessions and conversions from the signals mart - `conversions`: per-definition conversion counts - `anatomy`: which components get clicked and how far visitors scroll Every performance block carries a `state`: only `ok` means the numbers are real. `no_data`, `unconfigured` and `error` return null metrics. Report the state, never substitute zero. `traffic`, `conversions` and `anatomy` are sampling-weighted estimates and honour `window`. `performance` is snapshot-shaped: it reports the window the capture covered, which `window` does not change. Omit `urlPath` for a brand-wide roll-up.
Start an agent task
Run one autonomous agent turn for this brand. `command: "prompt"` takes a free-form `prompt`, the instruction the agent carries out ("Add an About page with a team section"). `command: "build_page"` takes an `opportunityRef` from the page-opportunities list and builds that page. A task takes about five minutes. Poll the task detail route every 30-60 seconds, or watch the chat thread the response links to. **One task at a time per brand.** A second create while one is running answers 409: the agent holds a single workspace per brand, so it runs tasks one after another rather than queuing them. Wait for the running task to settle, then submit the next one.