Checking links across an entire site
Most of the time, checking the page you're looking at is enough. When it isn't, "Scan site" extends the same checking engine across multiple pages of one site, with real limits, stated plainly.
How it works
In the popup, click Scan site. Link Alpha opens its own Site Audit tab, with live progress and the results as they arrive, and a background tab (never switched to; up to three for larger Pro crawls) that it navigates through same-site pages it discovers, reading each page's links the same way it reads the page you're already looking at. Background tabs are always closed when the scan finishes, is cancelled, or fails. Chrome only asks for that site's access if you have limited Link Alpha's site access in its settings.
Scope and limits
- Crawl size: up to 25 pages on Free; on Pro you choose 25, 100, 250, 500, 1,000 pages (default 100). You see the size before the scan starts. Same site only: links to other sites are checked, never crawled.
- Site scans per day: 10 on Free and 10 on Pro. This is an operational limit, separate from crawl size.
- Pages are found by following links. On Pro, the site's own sitemap (and sitemap index, including
.xml.gz) is also read, same site only and within fixed size limits; if no sitemap can be read, or reading stopped at a limit, the report says so. The site'srobots.txtDisallow rules decide which pages are read; those pages' links are still checked. - Endless URL patterns (calendars, filter combinations, repeating paths) are skipped as likely crawl traps, and redirects to a page already read are not read twice. A crawl that reaches its page limit, is cancelled, or skips pages is labelled partial, never as the whole site, and its results are kept.
- If Chrome stops the extension mid-scan (or the browser restarts), the pages read so far are kept and the Site Audit tab offers Resume scan.
- Pro can compare two saved audits of the same site: newly broken, now working, still broken, and not rechecked. An issue is never marked fixed just because the later audit didn't reach its page or couldn't check it.
- A page that's slow or broken is skipped so the rest of the crawl can continue. It doesn't hang or abort the whole scan.
- Each ordinary page scan still allows up to 5,000 links, a shared technical safety cap, not something full-site scanning changes.
Reviewing results that span multiple pages
Results found on more than one crawled page show their full occurrence count, same as duplicate links within a single-page scan. Locate works normally for a result found on the page you started from; for a result found only on a different crawled page, Link Alpha tells you to open that original page rather than highlighting the wrong one.
What this isn't
This is not an unlimited, scheduled, or cloud website crawler, and not continuous monitoring. Every scan runs in your own Chrome when you start it, and Link Alpha doesn't change your site: the AI feature prepares a repair brief for you to review and send yourself.