The WCAG 2.2 AA checklist, in plain English.
All 55 Level A and AA success criteria, grouped by principle, with one line on what to check and whether a scan can help. Work through it page by page.
Last reviewed November 3, 2026. General information, not legal advice.
At a glance
- Standard
- WCAG 2.2, October 2023
- Criteria at AA
- 55 in 2.2, 50 in 2.1
- New in 2.2 at A or AA
- 6 criteria
- Removed in 2.2
- 4.1.1 Parsing
- Scan can flag
- About 20 of 55, partly
- WCAG 2.2 Level AA has 55 success criteria: 31 at Level A and 24 at Level AA. WCAG 2.1 AA has 50.
- Meeting 2.2 AA also meets 2.1 AA, so one checklist covers both. The six criteria marked New in 2.2 are the difference, plus 4.1.1 Parsing, which 2.2 removed.
- An automated scan can flag some cases of about 20 criteria. The rest, and the judgment calls in those 20, need a person with a keyboard and a screen reader.
This is every WCAG 2.2 success criterion at Level A and Level AA, the level most laws and contracts ask for. Each row gives the criterion number and name as the W3C writes it, its level, one line on what to check, and whether an automated scan can flag problems with it. For the background on versions, levels and the law, read the WCAG compliance guide first.
How to use this checklist
- Pick your pages. List each page template (home, category, product or article, cart, checkout, contact form) and any process that spans pages. Check one of each, then a sample.
- Scan first. Run the free accessibility checker or an automated audit on those pages. Fix what it flags. That clears many of the rows marked “Some cases”.
- Test by hand. Go through the rest with a keyboard, a screen reader (VoiceOver on Mac, NVDA on Windows) and the browser zoomed to 200 and 400 percent.
- Write down what is left. Known issues and dates go in your accessibility statement.
“Some cases” means a scan catches the clear-cut version of the problem, such as an image with no alt attribute. It cannot tell whether alt text that exists is accurate. A row marked “No” always needs a person.
1. Perceivable
People can take in the content with at least one sense. 20 criteria: 9 at Level A, 11 at Level AA.
| Criterion | Level | What to check | Scan can flag it |
|---|---|---|---|
| 1.1.1 Non-text Content | A | Every meaningful image, icon and image button has alt text that says what it shows or does. Decorative images have empty alt. | Some cases |
| 1.2.1 Audio-only and Video-only (Prerecorded) | A | A podcast has a transcript. A silent video has a text or audio description. | No, test by hand |
| 1.2.2 Captions (Prerecorded) | A | Recorded videos with sound have accurate captions, not just auto-generated ones left unchecked. | Some cases |
| 1.2.3 Audio Description or Media Alternative (Prerecorded) | A | Visual information in a video that the soundtrack does not explain is described in audio or in a text alternative. | No, test by hand |
| 1.2.4 Captions (Live) | AA | Live video with sound, such as a webinar, has live captions. | No, test by hand |
| 1.2.5 Audio Description (Prerecorded) | AA | Recorded videos have audio description for visual information the soundtrack does not cover. | No, test by hand |
| 1.3.1 Info and Relationships | A | Headings, lists, tables and form labels are marked up as such in the code, not just styled to look that way. | Some cases |
| 1.3.2 Meaningful Sequence | A | The reading order in the code matches the order that makes sense on screen. | No, test by hand |
| 1.3.3 Sensory Characteristics | A | Instructions do not rely only on shape, position or color, as in “click the round button on the right”. | No, test by hand |
| 1.3.4 Orientation | AA | The site works in both portrait and landscape on a phone. | No, test by hand |
| 1.3.5 Identify Input Purpose | AA | Name, email, address and phone fields carry the right autocomplete value so browsers can fill them in. | Some cases |
| 1.4.1 Use of Color | A | Color is never the only signal. Links in body text are underlined or otherwise marked, and errors say what is wrong in text. | Some cases |
| 1.4.2 Audio Control | A | Audio that plays on its own for more than 3 seconds can be paused or muted. | Some cases |
| 1.4.3 Contrast (Minimum) | AA | Text has a contrast ratio of at least 4.5:1, or 3:1 for large text. | Some cases |
| 1.4.4 Resize Text | AA | At 200 percent browser zoom, text is readable and nothing is cut off. Pinch zoom is not disabled. | Some cases |
| 1.4.5 Images of Text | AA | Text is real text, not part of an image, except for logos. | No, test by hand |
| 1.4.10 Reflow | AA | At 320 CSS pixels wide (400 percent zoom on a laptop), content fits without sideways scrolling. | No, test by hand |
| 1.4.11 Non-text Contrast | AA | Input borders, icons, focus rings and chart lines have at least 3:1 contrast against what is next to them. | No, test by hand |
| 1.4.12 Text Spacing | AA | Nothing breaks when a reader increases line height, letter spacing or paragraph spacing. | No, test by hand |
| 1.4.13 Content on Hover or Focus | AA | Tooltips and hover menus can be dismissed with Escape, stay open while the pointer is over them, and do not vanish on their own. | No, test by hand |
2. Operable
People can use every control, whatever input they use. 20 criteria: 14 at Level A, 6 at Level AA. Three are new in 2.2.
| Criterion | Level | What to check | Scan can flag it |
|---|---|---|---|
| 2.1.1 Keyboard | A | Every link, button, menu, slider and form control works with the keyboard alone. | Some cases |
| 2.1.2 No Keyboard Trap | A | Focus never gets stuck. You can Tab into and out of every widget, embed and modal. | No, test by hand |
| 2.1.4 Character Key Shortcuts | A | Single-letter shortcuts can be turned off or remapped, or only work when their component has focus. | No, test by hand |
| 2.2.1 Timing Adjustable | A | Time limits, such as a session timeout or a cart hold, can be turned off, adjusted or extended. | Some cases |
| 2.2.2 Pause, Stop, Hide | A | Carousels, tickers and animations that run longer than 5 seconds can be paused. | Some cases |
| 2.3.1 Three Flashes or Below Threshold | A | Nothing flashes more than three times a second. | No, test by hand |
| 2.4.1 Bypass Blocks | A | A skip link or proper landmarks let keyboard users jump past the header and menu. | Some cases |
| 2.4.2 Page Titled | A | Each page has a unique title that says what it is. | Some cases |
| 2.4.3 Focus Order | A | Tabbing moves through the page in an order that makes sense. Opening a modal moves focus into it. | No, test by hand |
| 2.4.4 Link Purpose (In Context) | A | Each link says where it goes, alone or with its surrounding sentence. Icon links have a text name. | Some cases |
| 2.4.5 Multiple Ways | AA | There is more than one way to find a page: menu plus search, a sitemap or related links. | No, test by hand |
| 2.4.6 Headings and Labels | AA | Headings and form labels describe what follows them. | No, test by hand |
| 2.4.7 Focus Visible | AA | You can always see which element has keyboard focus. | No, test by hand |
| 2.4.11 Focus Not Obscured (Minimum) (new in 2.2) | AA | A focused element is never completely hidden by a sticky header, cookie banner or chat bubble. | No, test by hand |
| 2.5.1 Pointer Gestures | A | Anything done with a pinch, swipe or multi-finger gesture also works with a single tap or click. | No, test by hand |
| 2.5.2 Pointer Cancellation | A | Actions fire when the pointer is released, so people can slide off a button they pressed by mistake. | No, test by hand |
| 2.5.3 Label in Name | A | A button's accessible name contains its visible text, so voice control users can say what they see. | No, test by hand |
| 2.5.4 Motion Actuation | A | Anything triggered by shaking or tilting the device also has a normal control, and motion can be turned off. | No, test by hand |
| 2.5.7 Dragging Movements (new in 2.2) | AA | Anything you can drag, such as a slider or sortable list, also works with clicks or taps. | No, test by hand |
| 2.5.8 Target Size (Minimum) (new in 2.2) | AA | Buttons and links are at least 24 by 24 CSS pixels, or have enough space around them. | Some cases |
3. Understandable
Content and controls behave in ways people can predict. 13 criteria: 7 at Level A, 6 at Level AA. Three are new in 2.2.
| Criterion | Level | What to check | Scan can flag it |
|---|---|---|---|
| 3.1.1 Language of Page | A | The html element has the right lang attribute, so screen readers use the right voice. | Some cases |
| 3.1.2 Language of Parts | AA | Passages in another language are marked with their own lang attribute. | Some cases |
| 3.2.1 On Focus | A | Moving focus onto a control does not open a new page, submit a form or move focus somewhere else. | No, test by hand |
| 3.2.2 On Input | A | Choosing an option or typing in a field does not change the page unexpectedly, unless the user was told first. | No, test by hand |
| 3.2.3 Consistent Navigation | AA | Menus that repeat across pages stay in the same order. | No, test by hand |
| 3.2.4 Consistent Identification | AA | Things that do the same job have the same name everywhere, such as one search icon and one label for it. | No, test by hand |
| 3.2.6 Consistent Help (new in 2.2) | A | Help such as contact details, a chat link or an FAQ link sits in the same place on each page that has it. | No, test by hand |
| 3.3.1 Error Identification | A | Form errors are described in text and point to the field that needs fixing. | No, test by hand |
| 3.3.2 Labels or Instructions | A | Every field has a visible label. Required fields and formats are stated before the user types. | Some cases |
| 3.3.3 Error Suggestion | AA | When the site knows how to fix an error, it says how, as in “Enter the date as MM/DD/YYYY”. | No, test by hand |
| 3.3.4 Error Prevention (Legal, Financial, Data) | AA | Orders, payments and legal submissions can be reviewed, corrected or reversed before they are final. | No, test by hand |
| 3.3.7 Redundant Entry (new in 2.2) | A | Information already entered in the same process, such as a shipping address, is filled in or offered again instead of retyped. | No, test by hand |
| 3.3.8 Accessible Authentication (Minimum) (new in 2.2) | AA | Logging in does not require a memory or puzzle test. Password managers and paste work, and there is an alternative to cognitive CAPTCHAs. | No, test by hand |
4. Robust
Content works with browsers and assistive technology. 2 criteria: 1 at Level A, 1 at Level AA.
| Criterion | Level | What to check | Scan can flag it |
|---|---|---|---|
| 4.1.2 Name, Role, Value | A | Custom buttons, toggles, tabs and menus expose their name, role and state to assistive tech. | Some cases |
| 4.1.3 Status Messages | AA | Messages like “Added to cart” or “3 results” are announced by screen readers without moving focus. | No, test by hand |
WCAG 2.1 vs 2.2
WCAG 2.2 became a W3C Recommendation on October 5, 2023. Against 2.1, it adds six criteria at Level A or AA (marked “new in 2.2” above) and three at AAA, and it removes 4.1.1 Parsing, which browsers made redundant. So:
- WCAG 2.1 AA: 50 criteria. Use this list, skip the six new ones, and count 4.1.1 Parsing as met.
- WCAG 2.2 AA: 55 criteria, all of the above.
The ADA Title II rule and the current EN 301 549 name 2.1 AA. Meeting 2.2 AA meets 2.1 AA as well, so if you are starting now, work to 2.2. Our post WCAG 2.2 in plain English explains each new criterion with examples.
After the checklist
A checklist is a snapshot. New products, plugins and campaigns change the page, so re-check on a schedule. Accessora’s audits can scan your pages daily, weekly or monthly and list each issue with its WCAG criterion and the elements affected. Automated fixes repair some common issues, such as a missing label or accessible name, in the browser (Enterprise plan). The accessibility widget gives visitors adjustments like larger text and high contrast. None of that replaces the rows marked “No, test by hand”.
When you are done, publish what you found. The accessibility statement generator turns a short form into a statement that names your target, the date you checked and any known issues.
Passing this checklist on the pages you tested is evidence of conformance for those pages on that date. It is not a certificate. Date your results and keep them with your statement.
Is this also a WCAG 2.1 AA checklist?
Yes. WCAG 2.1 AA is this list minus the six criteria marked New in 2.2, plus 4.1.1 Parsing, which 2.2 removed because browsers now handle it. That gives 50 criteria. If a law or contract names 2.1 AA, skip the six and you are working to the right list.
How many WCAG 2.2 AA criteria are there?
55: 31 at Level A and 24 at Level AA. Conforming at AA means meeting every one of them on every page in scope. Level AAA adds 31 more, and the W3C does not recommend AAA as a target for a whole site.
Can an automated tool check all of this?
No. A scan can flag some cases of about 20 criteria, such as missing alt text, low contrast or a form field with no label. It cannot judge whether alt text is accurate, whether the reading order makes sense, or whether a keyboard user can finish your checkout. Use a scan for the first pass, then test by hand.
Which criteria fail most often?
On most sites: 1.1.1 Non-text Content (missing or useless alt text), 1.4.3 Contrast (Minimum), 2.4.4 Link Purpose, 3.3.2 Labels or Instructions and 4.1.2 Name, Role, Value. Start there if you have little time.
Do I need to check every page?
Check every page type, then a sample of each. Most sites have a handful of templates: home, listing, product or article, cart, checkout, forms. Fix a template and every page built on it improves. Full processes such as checkout count as one unit, so test them end to end.
Let the scan do the first pass.
Add your site and Accessora lists each issue its audit finds, with the WCAG criterion and the elements affected, so you spend your manual time on the rest. 7-day free trial on every plan.