What WCAG 2.2 is
The W3C published WCAG 2.2 as a Recommendation in October 2023. It keeps everything from WCAG 2.1, removes the old criterion 4.1.1 Parsing, and adds nine new success criteria. Most of the new ones help people with low vision, motor impairments or cognitive disabilities.
Each criterion has a level: A (the minimum), AA (the usual legal target) and AAA (the highest, rarely required for whole sites). If you meet AA, you also meet A.
The WCAG 2.2 AA checklist
These are the criteria that most often fail on real websites. The last column shows what an automated audit like Lensux can check and what still needs a person.
| Success criterion | Level | What to check | Lensux |
|---|---|---|---|
| 1.1.1 Non-text content | A | Informative images have alt text; decorative images use an empty alt="". | Automated: images without an alt attribute |
| 1.2.2 Captions (prerecorded) | A | Videos with sound have captions. | Partly: looks for caption tracks on videos |
| 1.3.1 Info and relationships | A | Headings, lists, tables and form labels are real markup, not just styling. | Partly: heading order and semantic landmarks |
| 1.3.5 Identify input purpose | AA | Fields for personal data use the right autocomplete value. | Manual |
| 1.4.1 Use of colour | A | Colour is never the only way to show meaning (errors, links, charts). | Manual (flagged in the AI review) |
| 1.4.2 Audio control | A | Sound that plays on its own for more than 3 seconds can be paused or muted. | Automated: autoplay with sound |
| 1.4.3 Contrast (minimum) | AA | Text has 4.5:1 contrast; large text 3:1. | Automated: every text element on the rendered page |
| 1.4.4 Resize text | AA | Text can be zoomed to 200% without losing content. | Partly: font sizes in rem/em and the viewport tag |
| 1.4.10 Reflow | AA | Content works at 320 CSS px wide without horizontal scrolling. | Partly: horizontal overflow on mobile |
| 1.4.11 Non-text contrast | AA | Buttons, inputs and focus indicators have 3:1 contrast. | Manual |
| 2.1.1 Keyboard | A | Everything works with the keyboard alone. | Manual |
| 2.4.1 Bypass blocks | A | A skip link or landmarks let people jump past repeated navigation. | Automated |
| 2.4.2 Page titled | A | Each page has a unique, descriptive title. | Manual |
| 2.4.3 Focus order | A | Tab order follows the visual and logical order. | Manual |
| 2.4.4 Link purpose | A | Link text says where the link goes (no lone “click here”). | Manual |
| 2.4.7 Focus visible | AA | The keyboard focus is always visible. | Partly: checks that focus styles exist |
| 2.4.11 Focus not obscured (minimum) — new | AA | Sticky headers, banners and chat buttons never fully hide the focused element. | Manual |
| 2.5.7 Dragging movements — new | AA | Anything you can drag can also be done with a single click or tap. | Manual |
| 2.5.8 Target size (minimum) — new | AA | Click and tap targets are at least 24 × 24 CSS px, or have enough space around them. | Automated: flags targets under 44 px on desktop and mobile (stricter than WCAG) |
| 3.1.1 Language of page | A | The html element declares the page language. | Automated |
| 3.2.6 Consistent help — new | A | Help and contact options appear in the same place on every page. | Partly: checks that a help or contact link exists |
| 3.3.1 Error identification / 3.3.2 Labels | A | Every field has a visible label; errors are described in text. | Partly: form validation attributes |
| 3.3.7 Redundant entry — new | A | People don't have to type the same information twice in one process. | Manual |
| 3.3.8 Accessible authentication (minimum) — new | AA | Login needs no memory or puzzle test; paste and password managers work. | Manual |
| 4.1.2 Name, role, value | A | Custom controls expose a name, role and state to assistive technology. | Partly: ARIA misuse and duplicate IDs |
How to test a website, step by step
Run an automated scan
Start with the measurable checks: contrast, alt text, page language, target size and heading order. They are fast to fix and often the most common failures.
Use only the keyboard
Tab through the page. Can you reach every link and button, see where the focus is and open and close every menu and dialog?
Zoom to 200% and narrow the window
Text should grow without overlapping, and the page should work at 320 px wide without sideways scrolling.
Fill in every form
Check visible labels, clear error messages, autocomplete on personal data and that nothing has to be typed twice.
Try a screen reader on key pages
VoiceOver (Mac, iPhone) and NVDA (Windows) are free. Listen to the home page, a product or service page and your main form.
WCAG and the law in the EU
The European Accessibility Act (Directive 2019/882) applies since 28 June 2025 to many consumer services, including e-commerce and online banking. In Germany it is implemented by the Barrierefreiheitsstärkungsgesetz (BFSG). The harmonised standard EN 301 549 currently refers to WCAG 2.1 level AA; meeting WCAG 2.2 AA covers it as well. Very small businesses providing services are exempt, but accessible sites still reach more customers.
This page is practical guidance, not legal advice.
Lensux tests the automated part on every audit: contrast on the rendered page, alt text, page language, skip links, heading order, ARIA misuse, duplicate IDs, focus styles, autoplay and tap-target size on desktop and mobile. Monitoring re-checks these on a schedule so new problems don't slip in.
Frequently asked questions
Is WCAG 2.2 legally required?
It depends on the country and sector. In the EU, the European Accessibility Act and public-sector rules point to EN 301 549, which currently uses WCAG 2.1 AA. WCAG 2.2 AA includes everything in 2.1 AA, so meeting 2.2 is the safer target.
Can an automated tool test all of WCAG?
No. Tools can measure things like contrast, missing alt attributes and target size, but many criteria need human judgement — for example whether alt text is meaningful or whether the focus order makes sense.
What is new in WCAG 2.2?
Nine criteria, including Focus Not Obscured, Dragging Movements, Target Size (Minimum), Consistent Help, Redundant Entry and Accessible Authentication. The old criterion 4.1.1 Parsing was removed.