Skip to content

Latest commit

 

History

History
144 lines (99 loc) · 5.53 KB

README.md

File metadata and controls

144 lines (99 loc) · 5.53 KB

Accessible 🧠🌍

Accessible AI Preview

Making the web inclusive and accessible with multimodal GenAI

Accessible AI🧠🌍 is a multimodal Chrome extension that uses Generative AI powered by Chrome’s built-in Gemini Nano model to make the web more inclusive and accessible for people with visual impairments, dyslexia, and ADHD.

It combines multiple interaction modes — text, voice, and (soon) facial gestures — to create a seamless and inclusive experience where users can listen, speak, or express themselves to understand and navigate the web.

🌟 Features

For Visually Impaired Users

  1. Page Summary (Alt+Shift+H / Cmd+Shift+H on Mac)

    • Automatically generates a structured summary of the current page
    • Reads the summary aloud automatically
    • Perfect for quickly understanding page content without visual reading

  2. Voice Question (Alt+Shift+Q / Cmd+Shift+Q on Mac)

    • Ask questions about the page content using voice
    • Displays the transcribed question for confirmation
    • Automatically sends the question and reads the AI response aloud
    • Full hands-free interaction

For Users with Dyslexia/ADHD

  1. Question Interface (Alt+Shift+A / Cmd+Shift+A on Mac)
    • Interactive interface to ask questions about page content
    • Supports both text input and voice input
    • Clear, readable interface with soft blue theme
    • Get answers based only on the current page content

🎨 Design

  • Soft blue gradient theme - Gentle on the eyes, promoting focus
  • Smooth blur effect - Background blur when interfaces are open
  • Clean typography - Simple, readable Arial font
  • Accessible UI - High contrast, clear buttons, keyboard navigation

🚀 Installation

Prerequisites

  • Chrome 138+ (required for Language Model API)
  • Desktop version (Mobile Chrome doesn't support the Language Model API)
  • Sufficient RAM and disk space for the AI model

Steps

  1. Clone or download this repository
  2. Open Chrome and navigate to chrome://extensions/
  3. Enable Developer mode (toggle in the top right)
  4. Click Load unpacked
  5. Select the folder containing the extension files
  6. The extension is now installed and ready to use!

⌨️ Keyboard Shortcuts

Shortcut Description
Alt+Shift+H (Windows/Linux)
Cmd+Shift+H (Mac)		 Summarize the current page (for visually impaired users)
Alt+Shift+Q (Windows/Linux)
Cmd+Shift+Q (Mac)		 Ask a voice question about the page (for visually impaired users)
Alt+Shift+A (Windows/Linux)
Cmd+Shift+A (Mac)		 Open question interface (for users with dyslexia/ADHD)
Esc Close any open interface

🎯 How It Works

  1. Content Extraction: The extension extracts text from the current page (or selected text if available)
  2. AI Processing: Uses Chrome's built-in Language Model API to analyze and respond
  3. Accessibility Features:
    • Text-to-speech for responses
    • Speech-to-text for voice questions
    • Visual confirmation of transcribed text
    • Automatic audio feedback

📋 Requirements

  • Chrome 138+ - Required for Language Model API support
  • Desktop Chrome - Mobile versions don't support this API
  • Internet connection - For initial model download (if needed)
  • Permissions:
    • activeTab - To read page content
    • scripting - To inject content scripts
    • notifications - For error messages

🔒 Privacy & Security

  • All processing happens locally using Chrome's built-in Language Model API
  • No data is sent to external servers
  • Page content is processed only in your browser
  • The extension only accesses pages you explicitly interact with

🐛 Troubleshooting

Extension doesn't work?

  1. Check Chrome version: Ensure you have Chrome 138 or higher

    • Go to chrome://version/ to check

  2. Verify installation: Check that the extension is enabled in chrome://extensions/

  3. Check page compatibility: The extension doesn't work on Chrome system pages (chrome://, about:, etc.)

    • Navigate to a regular webpage (http:// or https://)

  4. Model availability: If you see "Model unavailable", try:

    • Restarting Chrome
    • Ensuring sufficient disk space and RAM
    • Waiting a moment and trying again (model may be downloading)

Speech recognition not working?

  • Ensure your microphone permissions are granted
  • Check that your browser supports Web Speech API
  • Try refreshing the page and trying again

Text-to-speech not working?

  • Check your browser's speech synthesis settings
  • Ensure system volume is up
  • Try using the "Read" button instead of automatic reading

📝 Notes

  • The Language Model API is currently experimental and may vary in availability
  • First use may require downloading the model (progress shown in console)
  • The extension works best with text-heavy pages
  • Selected text takes priority over full page content

🤝 Contributing

This extension is designed with accessibility in mind. If you have suggestions for improvements, especially regarding accessibility features, please feel free to contribute!

📄 License

This project is created for accessibility purposes. Use and modify as needed to help make the web more accessible.

🙏 Credits

Built with Chrome's Language Model API, Web Speech API, and a focus on inclusive design.

Accessible 🧠🌍 - Making the web inclusive and accessible with AI agents