Photo Scrubber
simonwillison.net- Category
- Cybersecurity
- Rank
- No. 214Apps board
Previous survey · No. 197 ·
- Type
- TOOL
- Date
About
Photo Scrubber is an experimental browser tool by Simon Willison that detects faces in a photo and blurs them automatically, with metadata removal. It uses Google's MediaPipe compiled to WebAssembly along with the BlazeFace detection model, so processing happens locally. It was built to avoid sharing identifiable faces of strangers.
At a glance
Photo Scrubber makes a new copy of a photo with faces covered and embedded metadata stripped. It is worth trying for routine sharing when you can inspect the whole image yourself.
Good fit
People preparing an ordinary JPEG, PNG or WebP for sharing who want face suggestions and a manual solid cover in the same tool.
Consider something else when
Use an audited workflow if publication could expose someone to harm. A face detector can miss people, and clothing, signs and backgrounds can still give them away.
Start here
Start with a disposable test image. Check the suggested face boxes, draw solid covers over missed faces or visible identifiers, then use Prepare & verify image before saving the clean copy.
Cost and requirements
The public tool lists no account or payment step. It accepts JPEG, PNG and WebP up to 75 MB. The author says image processing runs in the browser, and automatic detection loads JavaScript, WebAssembly and model files from external providers.
Limits to know
The documented export rebuilds the pixels and removes metadata containers. It leaves identifiers visible in the scene unless you cover them. HEIC, RAW, SVG and animated inputs are unsupported. Detection and export were reviewed from source, with no image run through the tool.
Sources checked 2026-11-05. Source-based assessment.
Tags
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.