Detect faces with an on-device model, choose exactly which ones to blur, adjust the blur strength, and draw boxes over licence plates, screens, or anything else the detector misses.
Your images are processed on your device and are never uploaded.
This tool is compute-intensive and works best on a desktop browser.
Drop your image here
or drag and drop
Accepts JPG, PNG, WEBP · Up to 512 MB per file
Add a photo containing faces.
Detect faces, then click each one you want blurred.
Draw extra boxes over anything else, process, and download.
Uploading a photo to anonymise it hands the unredacted original to a third party first, which defeats the purpose entirely. Detection and blurring both happen here on your device — the identifiable version never exists anywhere but your own machine.
Face detectors do not find licence plates, house numbers, screens, or badges. Draw a box over any region and it is blurred with the same strength, so a single pass can redact a whole scene.
A strong Gaussian blur destroys the underlying pixel information and is not reversible from the output image. Light blurring or pixelation at a large block size can sometimes be attacked — use a generous radius when anonymity actually matters.
Detection weakens on profiles, partial occlusion, heavy shadow, and very small or distant faces. Always review the result and add manual boxes rather than trusting the automatic pass alone.
Re-encoding the image discards EXIF data, including GPS coordinates. That matters — a blurred face in a photo tagged with a home address is still identifying.