Media Vizard is a lightweight, privacy-first image redaction tool that runs entirely in the browser—no server required. It’s built to help users quickly and securely redact faces and sensitive metadata from images, especially in high-risk or vulnerable situations such as protests, immigration enforcement zones, and activist documentation.
🌐 Live Demo: https://mediavizard.netlify.app
🎥 Demo Video For Hackathon Submission: Watch on YouTube
😈 Hackathon project page : https://devpost.com/software/media-vizard
- ✅ 100% client-side image redaction (no uploads, no server processing)
- 🧠 Multi-model AI face detection (MediaPipe, BlazeFace, FaceDetection TF.js)
- 🔍 Toggle between low, medium, and high sensitivity
- 🖼 Manual face tagging for crowds or missed detections
- 👁 Multiple redaction modes: blur, pixelate, blackout
- 🧾 EXIF metadata removal (strips GPS, timestamps, device info)
- 📱 Fully responsive and mobile-friendly
- ⚡ Works offline via PWA caching
- React (Vite)
- TensorFlow.js &
@tensorflow-models/face-detection - MediaPipe + BlazeFace
- TypeScript
- Netlify (for hosting)
- HTML Canvas API
- EXIF.js (for metadata scrubbing)
- Vite + Tailwind CSS
- Clone this repo
git clone https://github.com/jeredhiggins/hackathon_project_mediavizard.git
cd media-vizard