1. 📁 Upload photo
Choose an image that may contain sensitive information.
Obscure identities instantly with AI-powered face detection. All processing runs locally in your browser.
or click to browse
Processed locally
On-device face blur: tries MediaPipe, then the browser Face Detector when available. Only faces are detected — not license plates or text.
In an era of increasing privacy regulation (GDPR, CCPA, LGPD), sharing photos containing identifiable faces carries real legal and ethical risks. Whether you’re a journalist, real estate photographer, event organizer, or just someone posting crowd photos online, you often need to blur faces before publishing.
Manually blurring faces in Photoshop or a phone app is tedious — especially when a photo contains dozens of faces. This tool automates the entire process using AI.
Our face blur pipeline uses MediaPipe Face Detection — Google’s lightweight AI model running entirely within your browser:
The AI model itself (~2 MB) is loaded once and runs locally via WebAssembly. No cloud API calls, no third-party AI services, no privacy paradox.
Most AI-powered face blur tools send your photos to a cloud server for processing. This creates an ironic situation: you’re sending private photos to a third party in order to make them more private.
Our approach eliminates this contradiction:
Blur bystanders’ faces before posting urban photography to social media or your portfolio.
Property photos often capture neighbors, delivery drivers, or passersby. Auto-blur protects their identities before listing publication.
Conference, concert, and party photos can contain hundreds of identifiable faces. Essential for GDPR compliance in the EU.
Report on public events while protecting the identities of visible individuals — particularly important for protests and sensitive gatherings.
1. 📁 Upload photo
Choose an image that may contain sensitive information.
2. 🫥 Detect sensitive regions
AI identifies faces, plates, screens, or private text zones.
3. ⬇️ Review and export
Confirm blur masks and download the protected image.
Planned detection includes faces, license plates, screens, and selected text regions.
Yes, the workflow is designed to allow review and manual corrections.