What is FaceHush?
FaceHush is a free web tool that automatically detects and blurs, pixelates or covers faces in photos. It runs entirely in your web browser: photos are never uploaded, there is no account, and saved images have their location metadata removed.
How it works
- Decoding: your browser opens the image locally, applying the correct rotation.
- Detection: Google’s open-source MediaPipe Face Detector (BlazeFace) runs on your device via WebAssembly. Big photos are scanned in overlapping tiles to catch small faces, and duplicate hits are merged.
- Hiding: each face region is blurred (multi-pass down-sampling), pixelated or filled, inside an oval or rectangle mask.
- Saving: the image is re-encoded from the canvas, so no EXIF metadata (GPS, camera, time) is copied into the new file. Photos above 16 megapixels are scaled to 16 MP, the largest canvas iPhone Safari allows.
What gets downloaded from the internet
Only the detection code and model: the MediaPipe library from jsDelivr and the model file from Google’s model storage. These are the same for everyone and contain none of your data. See the privacy page.
Contact
Questions, bug reports or feature requests: send a message through the feedback chat (a real person reads every one), or reach the developer at karakabakov.com.
Credits
The example photo at the top of the tool pages is by Land O’Lakes, Inc. on Unsplash.
Who builds it
FaceHush is built and maintained by Gjorge Karakabakov, an independent developer who makes small, focused apps and web tools. Feedback and feature requests are welcome: use the feedback chat.
Tools
Last updated