Skip to content
sitebrunch - go to homepage

What the automated check tests

Which WCAG success criteria the automated check covers, sorted by the four principles, and what it looks at for each one.

Last updated: October 13, 2026

The automated check in WCAG monitoring tests your pages against 19 WCAG success criteria. They are spread across all four principles: perceivable, operable, understandable and robust.

This page shows which criteria those are and what the check looks at for each one. The conformance level of the criterion, A or AA, is given in brackets. On top of that come best-practice checks that do not belong to any WCAG criterion.

Perceivable

Content has to be offered in a way everyone can perceive, for example without seeing or hearing it.

  • 1.1.1 Non-text Content (A): Images, image buttons, embedded objects and elements with the role img, such as SVG graphics, have a text alternative. Progress bars and meters have a name.

  • 1.2.2 Captions (Prerecorded) (A): Videos have a captions track.

  • 1.3.1 Info and Relationships (A): Lists, definition lists and tables are built correctly in the code. Header cells of a table are linked to their data cells, for example. ARIA roles that require certain child or parent roles have them.

  • 1.3.5 Identify Input Purpose (AA): The autocomplete attribute of form fields has a valid value.

  • 1.4.1 Use of Color (A): Links inside running text differ from the text around them by more than color.

  • 1.4.2 Audio Control (A): Audio and video that start on their own do not play sound for more than three seconds without controls to stop them.

  • 1.4.3 Contrast (Minimum) (AA): Text has enough contrast against its background.

  • 1.4.4 Resize text (AA): The viewport setting of the page does not block or limit zooming.

  • 1.4.12 Text Spacing (AA): Line, word and letter spacing are not fixed by inline styles with !important, so they can be adjusted.

Operable

Everyone has to be able to operate the page, including with the keyboard alone and at their own pace.

  • 2.1.1 Keyboard (A): Scrollable regions and content inside frames can be reached with the keyboard. Server-side image maps are not used.

  • 2.2.1 Timing Adjustable (A): The page does not reload or redirect on a timer through meta refresh.

  • 2.2.2 Pause, Stop, Hide (A): The outdated elements blink and marquee are not used.

  • 2.4.1 Bypass Blocks (A): Repeated blocks can be skipped, for example with a skip link, headings or landmarks.

  • 2.4.2 Page Titled (A): The page has a title.

  • 2.4.4 Link Purpose (In Context) (A): Links and clickable areas of image maps have a discernible name.

Understandable

Content and operation have to be understandable. That includes assistive technology knowing the language of the page.

  • 3.1.1 Language of Page (A): The html element has a lang attribute with a valid language code. If xml:lang is set as well, it agrees with it.

  • 3.1.2 Language of Parts (AA): lang attributes inside the page have a valid language code.

  • 3.3.2 Labels or Instructions (A): No form field has several labels at the same time.

Robust

The code has to be clean enough for browsers and assistive technology such as screen readers to interpret it reliably.

  • 4.1.2 Name, Role, Value (A): Buttons, form fields, frames and ARIA controls have an accessible name. ARIA roles and ARIA attributes are valid and allowed where they are used. IDs used in ARIA references are unique, elements hidden with aria-hidden are not focusable, and controls are not nested inside each other. More individual checks run for this criterion than for any other.

Best Practice

Some checks do not belong to any WCAG criterion but still catch common stumbling blocks. In the report they are listed in the "Best Practice" section. A finding there is not a WCAG failure, but a good pointer to what you can improve.

The check looks, for example, at whether the page has a level 1 heading, whether heading levels follow each other in the right order, whether there is exactly one main region, whether all content sits inside landmarks and whether headings or table headers are empty.

AI Readiness

With "AI Readiness" sitebrunch also checks how well AI systems and language models can read and understand your page. The check is switched off at first. You switch it on in the settings of your project. It has three parts:

  • No-JavaScript Readability: Text, images, headings and landmarks are already in the source HTML and do not only appear after JavaScript has run.

  • Structured Data for AI: The JSON-LD of the page is valid. Schemas such as Organization, WebSite and BreadcrumbList are present and their required fields are filled.

  • Alt Text Quality: An AI model judges, for a limited number of images per page, whether the text alternative describes the purpose of the image.

The results are recommendations. They do not change the score of your page.

What the automated check does not cover

Info
A passed check does not mean the criterion is met. Automation can only ever test part of a criterion.

An example: the check detects whether an image has a text alternative. Whether that text describes the image well is something a person has to judge. The same goes for the edge cases of the criteria above and for every criterion not listed here. How to test those manually with sitebrunch is covered in the Audits area.