How does it work?

This library is pretty simple, if you're curious about how it works you can read the following information and look at the internals.

@react-three/a11y purpose is to manage the accessibility part of what is inside your canvas by syncing semantic DOM absolutely positioned over what's currently visible in your page.

Basically when you add the A11y component with a role, react-three-a11y will append the corresponding HTML tag to your document.

• role="link" => a
• role="button" => button
• role="content" => p
• role="togglebutton" => button ( + aria-pressed )

The position is synced by a minimalist fork of the Drei Html component

Inside an A11y component, you can access the hover, focused and pressed state through the useA11y() hook. This hook returns the context of the A11y component.

The A11yAnnouncer is used to communicate with screen readers through a div only visible to screen readers. It uses a zustand store to update the div with each new message. The div is roughly like this.

<div aria-atomic="true" aria-live="polite">{message}</div>

The A11ySection component appends an HTML section in which a p element describe the content of the section. Wrapped around some A11y components, it will cause the HTML from those component to be inside the section.

You would then have a generated DOM that could look something like this.

<section aria-label="label">
  <p>description</p>
  <button></button>
  <p></p>
  <button></button>
</section>

For the A11yUserPreferences component, it simply exposes through a context the state of prefers-color-scheme and prefers-reduced-motion media queries. It watches for change through

window.matchMedia('(prefers-reduced-motion: reduce)')
window.matchMedia('(prefers-color-scheme: dark)')