6 min read

How PDF Accessibility Checkers Actually Work

An automated accessibility checker is the fastest way to find out where a PDF stands. But to use one well, it helps to understand what it is actually doing — and where it stops and human judgment has to take over.

What a checker tests

A checker inspects the PDF structure against the machine-testable requirements of PDF/UA (ISO 14289) and WCAG. The PDF Association codified these as the Matterhorn Protocol — 31 checkpoints and 136 failure conditions — which is why checker reports read like a list of specific, granular failures.

What it can catch automatically

  • Missing or incomplete tags.
  • Images without alternative text.
  • A missing document title or language.
  • Tables without header cells.
  • Fonts that are not embedded.
  • Missing PDF/UA conformance flags.

What it cannot judge

A checker can tell you alt text is present, but not whether it is meaningful. It can confirm tags exist, but not whether the reading order makes sense. Those require a human review.

The right way to use a checker

  1. 1Run a check first to get a baseline and clear the automatable failures.
  2. 2Fix structure, tags, alt text, and tables.
  3. 3Re-check to confirm those issues are gone.
  4. 4Do a focused manual review — read the alt text, follow the reading order with a screen reader.

You can run a free, instant check on this site to get that baseline report against PDF/UA and WCAG 2.1 AA.

Check your PDF for free

See exactly which PDF/UA and WCAG 2.1 AA issues your document has — instant, no signup.

Frequently asked questions

What can an automated accessibility checker not detect?
It cannot judge meaning. It can confirm alt text is present but not whether it is accurate, and confirm tags exist but not whether the reading order makes sense. Those judgment-based checks require a human, ideally with a screen reader.
Can I rely on a checker alone for compliance?
No. A checker clears the machine-testable failures, which is most of the work, but genuine compliance also needs a short manual review of what automation cannot assess — meaningful alt text, sensible reading order, and contrast on complex graphics.

Related how-to fixes