About Face Blur
Face Blur anonymizes photos in seconds. It finds faces and license plates automatically and covers them, and you can drag over anything else you want to hide, such as screens, documents, house numbers or name badges.
How it works
When you add a photo, two small AI models scan it. YuNet, a fast face detector from the OpenCV project, locates faces at different sizes and angles. A second model looks for vehicle license plates. Every detection becomes an editable area that you can remove with a click, and you can add your own areas by dragging a box anywhere on the image.
You then choose how each area is hidden: a soft blur, a blocky pixelate effect or a solid fill, in a rectangle or an ellipse. The effect is painted into the pixels of the exported image, so the downloaded file contains no hidden copy of the original.
What you can use it for
- Social media and blogs: share event, street or travel photos without exposing bystanders.
- Privacy compliance (US state privacy laws, the GDPR in Europe) before publishing photos of employees, customers or the public.
- Car listings: hide the license plate before posting a vehicle for sale.
- Journalism, real estate and support tickets, where screens, documents or faces appear by accident.
Tips for the best results
- Always look over the result. Very small, turned or partly hidden faces can be missed; drag a box to cover them.
- For truly sensitive images, use Solid or a strong Pixelate. A light blur can sometimes be partly reversed.
- Make areas a little larger than the face itself so hair and ears, which can also identify someone, are covered too.
- The ellipse shape looks more natural on faces; the rectangle suits plates, screens and documents.
Why on-device matters here: the whole point of blurring is protecting privacy. Uploading unblurred photos to a server would defeat that, so this tool never sends your image anywhere.