Vision
- Read aloud (focused element, selection, or whole page)
- Describe images on demand (Claude vision)
- High-contrast presets, larger text, dyslexia fonts
- Smart magnifier, focus highlighter
Assistive Agent
An AI agent that reads, describes, captions, and operates the page.
Visitors reach it as a floating landmark and pick what they need — read-aloud, image description, live captions, voice control. It coexists with JAWS, NVDA, and VoiceOver; it never overrides them, and it never claims conformance on its own.
Capabilities
One agent. Every modality.
Floating landmark on every page. Visitors choose what they need.
Motor
- Voice commands (“click checkout”, “scroll down”)
- Click-target enlargement, dwell click
- Switch input + eye-tracking adapters
Hearing
- Live captions for any audio/video element
- On-demand transcripts
- Visual sound alerts
Cognitive
- Simplify text at any reading level
- Summarize page → overview + bullets
- Distraction remover, define-on-hover, translate
Try it now
Click Read this page below.
A live demo running the agent against a sample article. Read it aloud, hide clutter with focus mode, or overlay live captions. Nothing to install.
The demo runs in your browser. No request leaves the page until you trigger a capability.
Two ways to ship it
Embed it, or install it.
Site owners add one script tag and every visitor gets the agent. Individuals install the browser extension (Chrome / Edge, Firefox via build) and carry it to any site they visit.
The widget is MIT-licensed and built to AAA itself — an accessibility tool should never be the worst accessibility on the page.
One-line install
<script
src="https://cdn.wholisphere.ai/embed.js"
data-key="org_yourkey"
data-position="bottom-right"
data-wholisphere
></script>