Crawls

A crawl test visits pages by following links from a start path and reports problems on each page, without any steps.

{
  "schemaVersion": 1,
  "kind": "crawl",
  "startPath": "/",
  "maxPages": 20,
  "maxDepth": 3,
  "include": [
    "/docs"
  ],
  "exclude": [
    "/logout",
    "/signout",
    "/sign-out"
  ],
  "checks": {
    "brokenLinks": true,
    "pageErrors": true,
    "consoleErrors": true,
    "failedRequests": true,
    "missingTitle": true,
    "accessibility": false
  }
}

• Only links on the same origin as the start URL are followed. Query strings are part of a page's identity; fragments are not.
• Each page reports its status, title, uncaught errors, console errors and failed requests. Failing pages get a screenshot (up to 8 per run).
• Finally, every internal link found, crawled or not, is checked and broken ones are listed with the pages that link to them.
• The plan limits pages per crawl: 10 on Free, 50 on Pro, 100 on Team.

Exclude sign out links and anything with side effects. Crawls only follow links; they do not submit forms or click buttons.