Accessibility Certified – Home
Free check Menu

Our test method, in the open

A seal is only worth the test behind it. So here is how we test, what a machine can contribute and where it stops.

Free checkRequest an audit
A man with headphones types on a laptop, with a refreshable braille display placed just behind the keyboard.

In short

We test against WCAG 2.2 level AA: 55 success criteria. Pages are sampled with WCAG-EM, the W3C evaluation methodology; automated tests with axe-core and our own measurements provide findings and screenshots. Then a person decides each criterion and tests with a screen reader. No criterion passes just because a machine found no error.

Standard

We test against the Web Content Accessibility Guidelines (WCAG) 2.2, conformance level AA: 55 success criteria at levels A and AA. Criterion 4.1.1 (Parsing) is obsolete in WCAG 2.2 and not tested.

The European standard EN 301 549 (version 3.2.1) refers to WCAG 2.1 AA for the web. WCAG 2.2 AA includes those criteria plus six more; they are marked below.

Page sample

We follow the W3C evaluation methodology (WCAG-EM): home page, one page per template, the pages for contact, forms and legal information, and the user journeys that matter for your service, such as checkout or booking. The tested URLs are listed in the report and on the certificate.

Automatic checks

  • axe-core on every page (rules for WCAG 2.0 to 2.2, A and AA)
  • Keyboard walk: order, visible focus, obscured focus, keyboard traps
  • Reflow at 320 CSS pixels and 200% zoom, with screenshots
  • Text spacing per WCAG 1.4.12: is anything cut off?
  • Forms: labels, autocomplete, browser error messages, without submitting
  • Media and motion: video, audio, autoplay, carousels, animations
  • Portrait and landscape, language, page titles, link texts, headings and landmarks

Human review

A person decides every criterion: passed, failed or not applicable. The scanner prepares what it finds as a suggestion with evidence. Whether an alt text fits, whether the order makes sense or whether an error message helps can only be judged by a person. No criterion is ever set to “passed” automatically.

Every decision is recorded in the report with its reasoning and evidence, so you can follow it.

Screen reader test

Every audit includes a walkthrough of two or three key user journeys with a screen reader (NVDA on Windows or VoiceOver on macOS and iOS). Anything not announced, announced wrongly or unreachable is recorded.

When a certificate is issued

  • All 55 criteria have been decided by a person.
  • No criterion has failed.
  • The screen reader test is complete.

The certificate states scope, standard, test date and expiry date and has a public verification page. It is valid for twelve months from the test date. It can be withdrawn, for example when monitoring finds a serious regression that is not fixed.

Limits

The certificate is a snapshot: it describes the tested version on the test date. It does not cover changes made afterwards, even while it is still valid. Only monitoring keeps testing, with scheduled scans and scans after changes to the website. The certificate is not legal advice and no promise that you meet every legal obligation.

Evidence, not claims

These are the screenshots our scanner takes of every page. Captured on this website.

Home page of this website with dashed blue boxes round every focusable element, numbered 1 to 17 in tab order, from the toolbar through the menu to the buttons in the hero.
Tab order Every focus stop is numbered, so you can see whether the order makes sense (2.4.3).
The same home page at 320 pixels wide: heading, text and buttons stack below each other and nothing is cut off.
Reflow at 320 px The page at 320 pixels wide: nothing may be cut off (1.4.10).
Home page with increased line, word and letter spacing. All text stays fully readable.
Text spacing Larger line, word and letter spacing must not hide anything (1.4.12).

Real output of our scanner, captured on accessibilitycertified.eu.

All 55 criteria

For each criterion: what the scanner contributes and the question a person answers.

axe-core: automatic rules find violations. Evidence: the scanner collects screenshots, lists and measurements. People only: nothing useful can be measured.

Four principles, 55 criteria

  • 1. Perceivable 20 criteria
  • 2. Operable 20 criteria
  • 3. Understandable 13 criteria
  • 4. Robust 2 criteria

1. Perceivable

  • 1.1.1 A

    Non-text Content

    axe-core evidence

    Does every image have an alt text that says what the image means here? Decorative images need alt="".

  • 1.2.1 A

    Audio-only and Video-only (Prerecorded)

    evidence

    Pre-recorded audio needs a transcript; silent video needs a text or audio alternative.

  • 1.2.2 A

    Captions (Prerecorded)

    axe-core evidence

    Do videos with sound have correct, synchronised captions?

  • 1.2.3 A

    Audio Description or Media Alternative (Prerecorded)

    evidence

    Is visual information in videos available as audio description or full text alternative?

  • 1.2.4 AA

    Captions (Live)

    evidence

    Live audio content (streams) needs live captions.

  • 1.2.5 AA

    Audio Description (Prerecorded)

    evidence

    Do pre-recorded videos have audio description where the picture carries information?

  • 1.3.1 A

    Info and Relationships

    axe-core evidence

    Are headings, lists, tables, form groups and landmarks marked up as such (not only styled)?

  • 1.3.2 A

    Meaningful Sequence

    evidence

    Does the reading order (DOM order, see list) make sense when the layout is removed?

  • 1.3.3 A

    Sensory Characteristics

    people only

    Do instructions rely only on shape, colour, size, position or sound ("click the round button on the right")?

  • 1.3.4 AA

    Orientation

    evidence

    Does the content work in both portrait and landscape?

  • 1.3.5 AA

    Identify Input Purpose

    axe-core evidence

    Do fields for personal data (name, email, phone, address) have the right autocomplete value?

  • 1.4.1 A

    Use of Color

    axe-core

    Is colour ever the only way to show information (links in text, required fields, errors, charts)?

  • 1.4.2 A

    Audio Control

    axe-core evidence

    Audio that plays automatically for more than 3 seconds needs pause/stop or volume control.

  • 1.4.3 AA

    Contrast (Minimum)

    axe-core

    Text contrast at least 4.5:1 (large text 3:1). Check also text on images and hover/focus states.

  • 1.4.4 AA

    Resize Text

    axe-core evidence

    At 200 % zoom, is all content and function still there without loss (see screenshot)?

  • 1.4.5 AA

    Images of Text

    evidence

    Is text shown as an image where real text would work (logos excepted)?

  • 1.4.10 AA

    Reflow

    evidence

    At 320 CSS px width: no horizontal scrolling, nothing cut off (see screenshot and overflow measurement).

  • 1.4.11 AA

    Non-text Contrast

    evidence

    Borders of inputs, icons, focus indicators and chart parts at least 3:1 against their background.

  • 1.4.12 AA

    Text Spacing

    evidence

    With increased line, paragraph, letter and word spacing, is any text cut off or overlapping?

  • 1.4.13 AA

    Content on Hover or Focus

    people only

    Tooltips and menus that appear on hover/focus: dismissable (Esc), hoverable, and they stay until dismissed.

2. Operable

  • 2.1.1 A

    Keyboard

    axe-core evidence

    Can every function be reached and used with the keyboard alone (menus, sliders, dialogs, forms)?

  • 2.1.2 A

    No Keyboard Trap

    evidence

    Can focus always leave every component again with the keyboard (watch embedded players and dialogs)?

  • 2.1.4 A

    Character Key Shortcuts

    people only

    Single-key shortcuts (letters, numbers) must be switchable off, remappable or only active on focus.

  • 2.2.1 A

    Timing Adjustable

    axe-core

    Time limits (session timeouts, auto-redirects) can be turned off, adjusted or extended.

  • 2.2.2 A

    Pause, Stop, Hide

    axe-core evidence

    Moving, blinking or auto-updating content (carousels, tickers, animations) can be paused or stopped.

  • 2.3.1 A

    Three Flashes or Below Threshold

    people only

    Nothing flashes more than three times per second.

  • 2.4.1 A

    Bypass Blocks

    axe-core evidence

    Is there a skip link or are landmarks/headings there to jump past repeated navigation?

  • 2.4.2 A

    Page Titled

    axe-core evidence

    Does every page title describe its topic and differ between pages (see list)?

  • 2.4.3 A

    Focus Order

    evidence

    Does the tab order (numbered screenshot) follow a logical sequence?

  • 2.4.4 A

    Link Purpose (In Context)

    axe-core evidence

    Is every link purpose clear from its text or context? Check the flagged generic link texts.

  • 2.4.5 AA

    Multiple Ways

    evidence

    Is there more than one way to find pages (navigation plus search, sitemap or index)?

  • 2.4.6 AA

    Headings and Labels

    evidence

    Do headings and labels describe their topic or purpose?

  • 2.4.7 AA

    Focus Visible

    evidence

    Is the keyboard focus always clearly visible? Elements without a detected focus style are flagged.

  • 2.4.11 AA

    Focus Not Obscured (Minimum)

    evidence new in 2.2

    Is a focused element ever completely hidden (e.g. by a sticky header or cookie banner)? Flagged cases listed.

  • 2.5.1 A

    Pointer Gestures

    people only

    Multi-point or path gestures (pinch, swipe) have a single-pointer alternative.

  • 2.5.2 A

    Pointer Cancellation

    people only

    Actions trigger on release (up-event), not on press, or can be aborted.

  • 2.5.3 A

    Label in Name

    axe-core

    The accessible name contains the visible label text (important for voice control).

  • 2.5.4 A

    Motion Actuation

    people only

    Functions triggered by shaking or tilting have a UI alternative and can be switched off.

  • 2.5.7 AA

    Dragging Movements

    evidence new in 2.2

    Everything done by dragging (sliders, sortable lists, maps) also works with single clicks.

  • 2.5.8 AA

    Target Size (Minimum)

    axe-core new in 2.2

    Click targets at least 24×24 CSS px or with enough spacing.

3. Understandable

  • 3.1.1 A

    Language of Page

    axe-core evidence

    Is the page language set correctly in <html lang>?

  • 3.1.2 AA

    Language of Parts

    axe-core

    Are passages in another language marked with lang (e.g. English quotes on a German page)?

  • 3.2.1 A

    On Focus

    people only

    Receiving focus never changes context on its own (no auto-submit, no new window on focus).

  • 3.2.2 A

    On Input

    people only

    Changing a setting (select, checkbox) does not change context unless announced beforehand.

  • 3.2.3 AA

    Consistent Navigation

    evidence

    Is repeated navigation in the same order on every page? Differences found are listed.

  • 3.2.4 AA

    Consistent Identification

    people only

    Components with the same function are labelled the same way across pages.

  • 3.2.6 A

    Consistent Help

    evidence new in 2.2

    Help (contact, phone, chat, FAQ) appears in the same relative place on every page where it exists.

  • 3.3.1 A

    Error Identification

    evidence

    Input errors are identified and described in text (validation messages captured without submitting).

  • 3.3.2 A

    Labels or Instructions

    axe-core evidence

    Every input has a visible label or instruction, including required-field and format hints.

  • 3.3.3 AA

    Error Suggestion

    evidence

    Where a correction is known, the error message suggests it (e.g. "Please enter a date as DD.MM.YYYY").

  • 3.3.4 AA

    Error Prevention (Legal, Financial, Data)

    evidence

    Orders, contracts and data changes can be reviewed, corrected or reversed before final submission.

  • 3.3.7 A

    Redundant Entry

    people only new in 2.2

    Information already entered in the same process is filled in again or selectable, not asked twice.

  • 3.3.8 AA

    Accessible Authentication (Minimum)

    evidence new in 2.2

    Login works without a cognitive test (no puzzle CAPTCHA, paste allowed, password managers work).

4. Robust

  • 4.1.2 A

    Name, Role, Value

    axe-core

    Custom controls (menus, tabs, accordions, toggles) expose name, role and state to assistive technology. Check with a screen reader.

  • 4.1.3 AA

    Status Messages

    people only

    Status messages (added to cart, form sent, search results) are announced by screen readers without focus moving.