Your own agents can do more of the loop over API v2 and MCP:
- Move actions through their lifecycle. Accept a suggestion, mark a fix live to open its before/after window, block or archive it, and reverse each step. The action page has the same controls, including Move back to suggestions and Restore.
- List tracked prompts and topics, with the IDs other tools take.
- Trace a content gap to the tracked prompts and cited pages behind it.
- Filter top cited pages and domains by topic, platform and prompt.
- Read the guard rails for one URL, such as page locks, without paging through the full list.
- Read 404 demand over 7 days as well as 30.
- Read your retained AI response history over 30 or 90 days, including failed and unfinished runs.
- Manage webhook subscriptions from MCP: list, create, delete and rotate signing secrets.
- See the connected project and the access the connection holds.
Every call uses your existing API access and scopes.