Skip to main content
Peel keeps a record of every call to your APIs and to the marketplace: the endpoint, whether it succeeded, the site’s status code, how long it took, what it was charged, and the requests it sent to the site. Open Runs in the app, or read them over REST and MCP.

List runs

This run searched Go packages from the marketplace on an unlimited plan, so it was charged 0. actor says who made the call: a signed-in person (session, with user_id and email), a key (api_key, with key_id) or an MCP client (bearer). It is null on runs made before November 17, 2026. Runs come newest first. Pass api_id to see one API’s runs, limit for 1 to 100 per page (25 by default), and next_cursor as cursor to get the next page. next_cursor is null on the last page.

One run and its requests

GET /v1/runs/{run_id} returns the run with requests: the method, origin, path and status code of each request the run sent, in order, up to 32. Static files such as images, stylesheets, scripts and fonts are left out. Query strings are dropped, and any value you passed in the call’s input is replaced with … in the path, so the log does not keep what you searched for or who you looked up. A run that is not yours, or does not exist, is 404 run_not_found.

From an MCP client

The Peel MCP lists list_runs and get_run. Both need mcp:read; an API key needs apis:read.