How we test every WCAG criterion.

Three ways of testing.

Automatic, partly automatic or human only: this is the breakdown across all 55 criteria.

This count is fixed in our test code and changes whenever that code changes. We test against WCAG 2.2, the version EN 301 549 V4.1.1 refers to. That is six more success criteria than in 2.1. Five of them are hard for a machine to judge: dragging movements, consistent help, redundant entry, accessible authentication, and a focus that does not disappear behind a bar. Our machine-tested share therefore gets smaller, not bigger: 31 of 55 criteria, where it was 30 of 50.

Every finding carries a certainty label.

AI-suggested findings can never fail a criterion or change the criteria matrix, and likely findings are clearly separated, so you decide where judgement is needed instead of chasing noise.

Established

Measured automatically and reliably. This criterion fails, with the evidence attached.

Likely: please verify

The machine flags it. Check it yourself in about two minutes before you act on it.

AI indication: confirmation needed

An AI suggestion. Never counts in the criteria matrix until a person confirms it.

The three labels are the same as on every Wexlo report.

Principle 1: Perceivable (20 criteria).

From alt text to contrast: information and interface must be perceivable.

• 1.1.1 · non-text contentLevel AHow we test: Partly automatic
• 1.2.1 · audio-only and video-only (prerecorded)Level AHow we test: Partly automatic
• 1.2.2 · captions (prerecorded)Level AHow we test: Partly automatic
• 1.2.3 · audio description or media alternative (prerecorded)Level AHow we test: Manual
• 1.2.4 · captions (live)Level AAHow we test: Manual
• 1.2.5 · audio description (prerecorded)Level AAHow we test: Manual
• 1.3.1 · info and relationshipsLevel AHow we test: Partly automatic
• 1.3.2 · meaningful sequenceLevel AHow we test: Manual
• 1.3.3 · sensory characteristicsLevel AHow we test: Partly automatic
• 1.3.4 · orientationLevel AAHow we test: Partly automatic
• 1.3.5 · identify input purposeLevel AAHow we test: Partly automatic
• 1.4.1 · use of colorLevel AHow we test: Partly automatic
• 1.4.2 · audio controlLevel AHow we test: Partly automatic
• 1.4.3 · contrast (minimum)Level AAHow we test: Partly automatic
• 1.4.4 · resize textLevel AAHow we test: Partly automatic
• 1.4.5 · images of textLevel AAHow we test: Manual
• 1.4.10 · reflowLevel AAHow we test: Partly automatic
• 1.4.11 · non-text contrastLevel AAHow we test: Manual
• 1.4.12 · text spacingLevel AAHow we test: Partly automatic
• 1.4.13 · content on hover or focusLevel AAHow we test: Manual

Principle 2: Operable (20 criteria).

Keyboard, timing and navigation: every function must be operable, even without a mouse.

• 2.1.1 · keyboardLevel AHow we test: Partly automatic
• 2.1.2 · no keyboard trapLevel AHow we test: Partly automatic
• 2.1.4 · character key shortcutsLevel AHow we test: Manual
• 2.2.1 · timing adjustableLevel AHow we test: Partly automatic
• 2.2.2 · pause, stop, hideLevel AHow we test: Partly automatic
• 2.3.1 · three flashes or below thresholdLevel AHow we test: Manual
• 2.4.1 · bypass blocksLevel AHow we test: Partly automatic
• 2.4.2 · page titledLevel AHow we test: Partly automatic
• 2.4.3 · focus orderLevel AHow we test: Manual
• 2.4.4 · link purpose (in context)Level AHow we test: Partly automatic
• 2.4.5 · multiple waysLevel AAHow we test: Partly automatic
• 2.4.6 · headings and labelsLevel AAHow we test: Partly automatic
• 2.4.7 · focus visibleLevel AAHow we test: Partly automatic
• 2.4.11 · focus not obscured (minimum)Level AAHow we test: Manual
• 2.5.1 · pointer gesturesLevel AHow we test: Manual
• 2.5.2 · pointer cancellationLevel AHow we test: Manual
• 2.5.3 · label in nameLevel AHow we test: Partly automatic
• 2.5.4 · motion actuationLevel AHow we test: Manual
• 2.5.7 · dragging movementsLevel AAHow we test: Manual
• 2.5.8 · target size (minimum)Level AAHow we test: Partly automatic

Principle 3: understandable (13 criteria).

Language, behaviour and error messages that are predictable and understandable.

• 3.1.1 · language of pageLevel AHow we test: Automatic
• 3.1.2 · language of partsLevel AAHow we test: Partly automatic
• 3.2.1 · on focusLevel AHow we test: Manual
• 3.2.2 · on inputLevel AHow we test: Manual
• 3.2.3 · consistent navigationLevel AAHow we test: Partly automatic
• 3.2.4 · consistent identificationLevel AAHow we test: Manual
• 3.2.6 · consistent helpLevel AHow we test: Manual
• 3.3.1 · error identificationLevel AHow we test: Partly automatic
• 3.3.2 · labels or instructionsLevel AHow we test: Partly automatic
• 3.3.3 · error suggestionLevel AAHow we test: Manual
• 3.3.4 · error prevention (legal, financial, data)Level AAHow we test: Manual
• 3.3.7 · redundant entryLevel AHow we test: Manual
• 3.3.8 · accessible authentication (minimum)Level AAHow we test: Manual

Principle 4: Robust (2 criteria).

Finally: content must work reliably with browsers and assistive technology.

• 4.1.2 · name, role, valueLevel AHow we test: Automatic
• 4.1.3 · status messagesLevel AAHow we test: Manual

All 55 criterion IDs, levels, names and coverage values above come verbatim from our own criteria map; we have not added or removed anything.