Liste de contrôle interactive WCAG 2.1 AA

All non-text content (images, icons, controls) has a text alternative that serves the equivalent purpose.

Provide a transcript for prerecorded audio-only, and a transcript or audio track for prerecorded video-only.

Synchronized captions are provided for all prerecorded audio in video content.

Provide an audio description or full text alternative for prerecorded video content.

Synchronized captions are provided for all live audio content in synchronized media.

An audio description is provided for all prerecorded video content in synchronized media.

Structure and relationships (headings, lists, tables, form labels) are conveyed in markup, not by visual styling alone.

When sequence affects meaning, the correct reading order can be programmatically determined.

Instructions don't rely solely on shape, size, position, or sound (e.g. "click the round button on the right").

Content does not restrict its view to a single display orientation (portrait or landscape) unless essential.

The purpose of common input fields (name, email, etc.) is programmatically identified, e.g. via autocomplete.

Color is not the only visual means of conveying information, indicating an action, or distinguishing an element.

Audio that plays automatically for more than 3 seconds can be paused, stopped, or muted.

Text and images of text have a contrast ratio of at least 4.5:1 (3:1 for large text).

Text can be resized up to 200% without loss of content or functionality and without assistive technology.

Real text is used instead of images of text, except where the presentation is essential (e.g. logos).

Content reflows to a single column at 320 CSS pixels wide without loss of information or two-dimensional scrolling.

UI components and meaningful graphics have a contrast ratio of at least 3:1 against adjacent colors.

No loss of content when users override line height, paragraph, letter, and word spacing within set limits.

Content that appears on hover or focus is dismissable, hoverable, and persistent.

All functionality is operable through a keyboard, without requiring specific timings for keystrokes.

Keyboard focus can be moved away from any component using only the keyboard.

Single-character key shortcuts can be turned off, remapped, or are only active on focus.

Users can turn off, adjust, or extend any time limit, with limited exceptions.

Moving, blinking, scrolling, or auto-updating content can be paused, stopped, or hidden.

Nothing flashes more than three times per second, or flashes are below the general and red thresholds.

A mechanism (skip link, landmarks, headings) lets users bypass blocks of content repeated on pages.

Pages have titles that describe their topic or purpose.

Focusable components receive focus in an order that preserves meaning and operability.

The purpose of each link can be determined from the link text alone or with its surrounding context.

More than one way is available to locate a page within a set (e.g. search and a sitemap or menu).

Headings and labels describe the topic or purpose of the content they introduce.

A visible focus indicator is shown for any keyboard-operable interface.

Functionality using multipoint or path-based gestures has a single-pointer alternative.

Actions triggered by a single pointer can be aborted or undone (no down-event activation).

For components with visible text labels, the accessible name contains that visible text.

Functionality triggered by device or user motion can also be operated by UI components and can be disabled.

The default human language of each page is programmatically set (e.g. the html lang attribute).

The language of phrases or passages that differ from the page language is programmatically identified.

Receiving focus does not trigger an unexpected change of context.

Changing a setting does not automatically cause an unexpected change of context unless the user is warned.

Navigation repeated across pages appears in the same relative order each time.

Components with the same functionality are identified consistently across the set of pages.

Input errors are detected and described to the user in text.

Labels or instructions are provided when content requires user input.

When an input error is detected and a fix is known, suggestions are provided to the user.

For legal, financial, or data submissions, actions are reversible, checked, or confirmed.

Markup has complete tags, no duplicate IDs, and proper nesting. (Obsolete in WCAG 2.2, but required for 2.1.)

For all UI components the name, role, state, and value are programmatically determinable (correct ARIA/HTML).

Status messages can be presented to assistive technology without receiving focus (e.g. aria-live regions).