How scanning works

Which pages

Up to 25 pages per site per scan: the homepage first, then pages from your sitemap with the shallowest paths first, then pages found by following links on the same site. Assets (PDFs, images), pages blocked by robots.txt and pages behind a login are skipped. Pages that return an error are listed under "Pages that could not be scanned" and their previous issues are marked unverified rather than fixed.

How politely

Our crawler identifies itself as AccessDiffBot/1.0 (+https://accessdiff.com/bot), follows robots.txt (including Crawl-delay), loads two pages at a time with pauses between them, never signs in, keeps no cookies and stores only page URLs, titles, element selectors and short element snippets. Nothing about your visitors is collected.

Which checks

Automated checks against WCAG 2.2 level A and AA using the axe-core rule set, plus a few labelled best-practice checks such as heading order. Each issue names the rule, the WCAG success criterion, the element on the page and how to fix it, with a link to the rule documentation. Automated checks find a portion of accessibility issues; a review by a person is needed to find the rest.

The weekly diff

Each issue gets a stable identity from its rule, its place on the page and the element itself. Next week we compare: issues that disappeared are fixed, issues not seen before are new, and the rest are unchanged. If a page could not be scanned this week, its previous issues are neither fixed nor unchanged; the report says so.

The score

A 0 to 100 health score for the site, averaged over the pages scanned. Each page starts at 100 and loses points per failing rule by impact (critical 10, serious 5, moderate 2, minor 1), with a smaller penalty for each additional element failing the same rule, capped per rule. The formula is printed at the bottom of every report. Use it to watch the trend, not as a pass mark.

When

Each site has a fixed weekly slot (a weekday and an hour chosen when the site is added, so scans spread across the week). The first scan runs as soon as a site is added. Scan now on the dashboard runs an extra scan; a scan that fails is retried daily until it succeeds or three attempts are used up.

Browser and viewport

Headless Chromium at 1280 by 800 pixels, with a short pause for client-side rendering. Mobile viewports are on the roadmap.