Glossary · Testing & tools

Automated accessibility testing

Definition: Automated accessibility testing means software checking web pages against fixed rules – for missing alt text, low contrast or unlabelled fields, for example. It is fast and repeatable, but only a minority of WCAG success criteria can be fully decided by automation. The rest needs human judgement.

German: Automatisierte Prüfung · Last reviewed: 28 September 2026

How we test, and where the limits are, is set out in our methodology.

What software can decide – and what it cannot

A rule can tell whether an image has alt text. It cannot tell whether the text describes the image. It measures the contrast of text on a plain background, but not on a photo. It sees whether a link has a name, but not whether "Learn more" makes sense in context.

In Reviseberg's own catalogue of the 55 WCAG 2.2 Level A and AA criteria, it looks like this:

What software can decide – and what it cannot
Automated checkCriteriaExamples
decides fully5page title (2.4.2), page language (3.1.1), contrast (1.4.3)
decides partly19alt text (1.1.1), name, role, value (4.1.2)
cannot reach31captions (1.2.2), error identification (3.3.1), status messages (4.1.3)

Every one of these classifications is on the criterion's own page in our WCAG overview.

Why it is still indispensable

Automated checks are fast, thorough and repeatable. They cover hundreds of pages where a person manages ten, and they find the same errors on every run. That makes them right for the common, mechanical errors and for checking that a fixed error stays fixed. People's time can then go where no rule reaches.

The usual mistake

A scan with no findings means the rules found nothing. It does not mean the page is accessible. That leap is behind accessibility statements that claim full conformance, and behind overlays that promise accessibility by script.

How Reviseberg handles it

Reviseberg runs every axe-core rule on every page and sends a keyboard agent through your most important journey. Anything a rule cannot decide appears as a potential issue, never as a pass. Criteria no automated check reaches stay "untested" until a person records a result – see manual accessibility testing.

Manual accessibility testing · axe-core · Accessibility overlay · WCAG · WCAG conformance levels

Further reading

Frequently asked questions

How much does an automated test find?

It fully decides only a minority of WCAG criteria. In our catalogue that is 5 of the 55 Level A and AA criteria; on 19 more it finds some of the failures.

Is an automated test enough for the European Accessibility Act?

No. It is a good start and shows many common errors. A reliable statement also needs manual testing.

How often should you test automatically?

After every significant change and on a schedule, weekly for example. New content brings new errors.

Sources

  1. W3C WAI, Selecting Web Accessibility Evaluation Tools – https://www.w3.org/WAI/test-evaluate/tools/selecting/
  2. W3C, WCAG 2.2 – https://www.w3.org/TR/WCAG22/
  3. Deque, axe-core – https://github.com/dequelabs/axe-core

See what a scan finds on your site

One page in about thirty seconds, no email needed. The full report covers up to 100 pages and a keyboard journey.