# Accessibility Tree

The accessibility tree is the browser's semantic representation of relevant interface objects for platform accessibility APIs and assistive technologies.

- Type: term
- Domain: Experience & interface
- Level: Intermediate
- Last reviewed: 2026-09-23
- Original: [Soroush’s Product Wiki](<https://wiki.soroush.uk/topic/accessibility-tree/>)

## What it means

The accessibility tree contains accessible objects derived from host-language semantics, rendered state, names, roles, properties, relationships, and ARIA. Browsers expose these objects through platform accessibility APIs. It is related to the DOM but is not a copy: purely presentational or hidden nodes may be omitted, while computed accessible information may differ from markup.

## Why it matters

Screen readers and other assistive technologies depend on the roles, names, states, and relationships browsers expose. Visual appearance or DOM presence alone cannot show whether a control is represented correctly.

## When to use it

Inspect the accessibility tree when building or debugging controls, labels, live regions, hidden content, or dynamic interfaces. Start with native semantic HTML, then add ARIA only where required and test actual tasks with assistive technology.

## How it works

1. The browser parses markup and computes element semantics from HTML and supported ARIA.
2. Rendering and hidden-state rules determine which relevant nodes are exposed.
3. Accessible names, roles, states, properties, and relationships are computed.
4. The browser maps accessible objects to the operating system's accessibility API.
5. Assistive technologies query changes and present navigation or interaction to users.

## A practical example

A custom disclosure looks correct but appears as unnamed text in browser accessibility tools. Replacing its clickable div with a button and synchronizing aria-expanded exposes the expected name, role, state, and keyboard behavior.

## Where it breaks

Developer-tool trees are approximations of platform exposure, mappings vary, and correct semantics do not guarantee usable interaction. Assistive technologies add their own modes and interpretations, so inspection cannot replace user-centered testing.

## Common mistakes

- Assuming every DOM node appears in the accessibility tree.
- Using ARIA to repair a custom control while omitting keyboard behavior.
- Hiding focusable descendants with aria-hidden.
- Treating a clean tree inspection as proof of an accessible product.

## Visualization

[View the visualization on the website](<https://wiki.soroush.uk/topic/accessibility-tree/#visualization-accessibility-tree>)

## Sources

1. [Core Accessibility API Mappings 1.2](<https://www.w3.org/TR/core-aam-1.2/>) — W3C · W3C recommendation
2. [Accessibility Tree](<https://developer.mozilla.org/en-US/docs/Glossary/Accessibility_tree>) — MDN Web Docs · Official technical documentation

## Related topics

- [Document Object Model \(DOM\)](<https://wiki.soroush.uk/topic/document-object-model/>)
- [Semantic HTML](<https://wiki.soroush.uk/topic/semantic-html/>)
- [Screen Reader Design](<https://wiki.soroush.uk/topic/screen-reader-design/>)
