Detects and blurs every face in a photo or your live webcam, using an on-device AI model. The image never leaves your browser — no upload, no server, not even for the live camera feed.
Loading face-detection model (~230KB, one time)…
How it works
A MediaPipe BlazeFace model runs locally in your browser (via WebAssembly/GPU) to find each face's bounding box, then every detected region — expanded by the padding percentage you set — is shrunk onto a tiny offscreen canvas and drawn back at full size with image smoothing disabled, producing a pixelated block instead of the original detail.
FAQ
Is it actually blurring the image, or something else?
It's pixelating, not applying a gaussian blur: each detected face region is shrunk down onto a tiny low-resolution canvas and then drawn back up at full size with image smoothing turned off, which produces the blocky mosaic look rather than a soft blur.
Does this work on live webcam video, or only uploaded photos?
Both — switching to webcam mode reconfigures the same on-device model into video mode and reruns detection on every animation frame, redrawing pixelated boxes in real time. The camera feed is processed entirely on your device and is never uploaded anywhere.
Free, unlimited, no account. If this saved you a paid tool, a small thanks is appreciated.