Accessibility

What the library provides

• ARIA roles are declared on interactive widgets, and widgets are keyboard-reachable (tabindex="0" on button, toggle, radio, sliders, input).
• Keyboard operation — see the full key table in Interaction.
• Accessible names stay in sync: changing a widget's value updates its aria-label along with the rendered text.
• Typing safety — global shortcuts never fire while the user is typing in a text field, during IME composition, or on auto-repeated keydowns.
• State reflection — aria-checked and aria-valuenow follow the visual state of toggles, radios and sliders.

Native text editing (opt-in)

By default an a-sxr-input is a canvas-rendered field updated programmatically. Adding native-editing="true" makes activation focus a visually hidden, labelled native <input> kept in sync with the widget:

• typing, arrow-key caret/selection movement, paste and IME composition all work with the real keyboard
• input events fire as the text changes
• change fires on commit (Enter or losing focus), once per changed value

<a-sxr-input width="2.8" height="0.5"
             native-editing="true"
             value="Type here"></a-sxr-input>

Known limitations

Screen readers

Widget ARIA roles and labels live on 3D entities in the WebGL scene; screen readers do not announce them. Keyboard operation is provided for sighted keyboard users. For a screen-reader-accessible experience, provide a parallel DOM interface alongside the scene.

• Immersive (VR) text entry is application-provided. The release does not bundle a Quest virtual keyboard. On desktop, enable opt-in native editing to type with a real keyboard.
• Focus indicators are visual only (widget border/focus colors); there is no roving-tabindex container for flex layouts, so Tab order follows DOM order rather than visual layout order.
• Focus is not persisted across DOM focus loss in a way assistive tech can observe — focus state is expressed visually.