Scanner internals

How our scanner works

This is the pipeline behind the Instant Automated Scan — not a sales slogan. It is also not a substitute for keyboard and screen reader testing.

  1. 1

    URL discovery

    The crawler reads the sitemap and follows same-site links from rendered pages. Trap URLs and off-site redirects are dropped. Cap: 40 pages per scan.

  2. 2

    Browser rendering

    Each page is opened in a real Chromium browser so JavaScript-rendered markup is what the engines see — not the empty shell of a static fetch.

  3. 3

    Automated engines

    axe-core runs first. IBM Equal Access then evaluates the same page against a WCAG 2.2 policy. IBM potential and manual items can surface as Needs Review rather than hard fails.

  4. 4

    WCAG normalization

    Engine rule IDs are mapped onto WCAG 2.2 A/AA criteria and a canonical ADA-* issue id so the report is citable, not a dump of vendor-specific names.

  5. 5

    Deduplication

    Shared component issues are grouped by rule, not by CSS selector. One header contrast defect with 100 occurrences is one finding.

  6. 6

    Human-review boundary

    Needs Review holds incomplete and unconfirmed checks. A full expert audit is a separate service. The scan never claims the site meets WCAG.

A real grouping example

From a recent full-site run: sitemap plus in-page links discovered 18 URLs, all 18 were scanned in a rendered browser, both engines ran, then canonical grouping collapsed the raw hit list.

18

pages discovered

18 pages scanned

214

raw occurrences

9

unique accessibility issues

  • 4 Serious
  • 5 Moderate
  • 1 Needs Review

One header contrast defect repeated 100 times is one systemic finding with 100 occurrences — not 100 unrelated problems.

Scan a website →

Run the scanner, then decide if you need a person

The Instant Automated Scan is free. Expert review and the full audit stay separate on purpose.

Scan a websiteRequest a full audit