PresentationCaptureForReview

A completely frontend JavaScript application that records your spoken presentation and converts it to text using Whisper AI - all running in your browser with no server required!

Features

• 🎤 Browser-based audio recording - Uses your microphone to capture presentations
• 🤖 Whisper Small model - English-only transcription (~300MB model)
• 💾 Automatic caching - Model downloads once and caches in your browser
• 📋 Easy copying - One-click copy to clipboard for pasting into LLMs
• 🔒 Privacy-focused - All processing happens locally in your browser
• 🌐 GitHub Pages ready - Deploy as a static site with zero server costs

Live Demo

Visit the live demo (replace with your actual GitHub Pages URL)

How to Use

1. Open the application in a modern browser (Chrome, Edge, or Safari recommended)
2. Wait for the Whisper model to download (only happens on first visit, ~300MB)
3. Click "Start Recording" and grant microphone permissions
4. Speak your presentation
5. Click "Stop Recording" when done
6. Wait a few seconds for transcription
7. Copy the text and paste it anywhere you need!

Deployment to GitHub Pages

1. Enable GitHub Pages:

   • Go to your repository settings
   • Navigate to "Pages" section
   • Select "main" branch as source
   • Save changes

2. Access your site:

   • Your site will be available at: https://your-username.github.io/PresentationCaptureForReview/
   • First load will download the model (~300MB) - subsequent visits use cached version

Technical Details

• Frontend Framework: Vanilla JavaScript (ES6 modules)
• Speech Recognition: Transformers.js with Whisper Small English model
• Audio Recording: MediaRecorder API with WebM format
• Model Size: ~300MB (downloads and caches automatically)
• Browser Support: Chrome, Edge, Safari (requires modern browser with Web Audio API)

Browser Requirements

• Modern browser with ES6 module support
• Microphone access permissions
• Sufficient storage for model caching (~300MB)
• Web Audio API support
• MediaRecorder API support

Privacy & Security

All audio processing happens entirely in your browser. No data is sent to any server. The Whisper model is downloaded from Hugging Face CDN and cached locally.

License

MIT License - Feel free to use and modify as needed!