Which faces are found, and which are missed
The tool looks for faces with a small open-source detector. Faces that look towards the camera at a reasonable size, as in a group photo, a selfie or a class picture, are the easy case.
Two kinds get missed: very small faces, such as people far off in the background, and faces turned away from the camera. Go over the photo once before saving and drag a box across any face still showing. A box also works for things that are not faces at all: a name badge, a number plate, a tattoo.
Profile pictures in a chat screenshot are a different problem, since many of them are not faces. Finding those, together with the names next to them, is what the Blur ID iPhone app is for.
Blur, pixelate, solid block or emoji
Four looks are available, and whichever you pick, only the face is covered. Hair, clothes, height, a distinctive bag and the place itself remain, and people who know someone can recognise them from those. When a person's safety depends on not being identified, use the solid block and consider cropping as well: a blur keeps a softened version of the face, a block does not.
- Blur: soft and quiet. It keeps the colours and the mood of the photo, which suits bystanders in the background of a street or event picture.
- Pixelate: a mosaic of squares. It tells the viewer plainly that a face was hidden on purpose.
- Solid block: a flat shape in a colour you choose. Nothing of the face is left under it.
- Emoji: a friendlier cover for a social post. The page for putting an emoji on each face explains that style.
Why the photo never leaves your device
Nothing on this page sends your photo to a server. The detector is downloaded to your browser from blurid.app's CDN, cached, and run on your own hardware; the blurred copy is produced in the same tab and saved by you.
Because the face model is small, there is no long wait: it is a fraction of a megabyte, and the browser keeps it for next time.
What this page does not do
- No clean export: each download has a small "via Blur ID" mark in the bottom-right corner. If you searched for a face blur without a watermark, this is not quite it.
- No output above 4,000 pixels on the long side; larger photos are scaled down to that.
- No video. A clip cannot be opened, and faces are not tracked from frame to frame.
- No batch. Photos are opened one at a time.
- No promise of completeness: check every face yourself before posting.
When the iPhone app or another tool is the better choice
In its current version the Blur ID iPhone app covers faces with an emoji or a solid block, so for the blurred look this page is the place to do it, on an iPhone as well. What the app adds is everything around that: it opens pictures from the Photos library, it also finds phone numbers and addresses in screenshots, and its Premium removes the mark. The guide to hiding faces in photos shows how it works, and the Blur ID home page covers the rest.
For a document or an ID photo, use a solid block instead of a blur; the guide to covering an ID card explains what to hide. For video you need a video editor: this page works on one still picture at a time.