Is the blur reversible?

For text — plates, addresses, card numbers — a solid box is the default precisely because a mild blur on a known font can sometimes be read back. For faces, blur or pixelation at the strength used here replaces the pixels; the original is not in the file. The edit is done to the picture itself, not as a removable layer.

Does it also remove the metadata?

Yes. The downloaded copy is rendered fresh and carries no EXIF, GPS or camera fields.

Can a photo ever be fully anonymous?

We will not promise that. Covering faces, plates and metadata removes the obvious identifiers; context — places, clothing, reflections — can still carry identity, and judging that needs you, not a detector. The tool shows you what it found and gives you the brush for the rest.

Available now

Anonymize a photo

Anonymizing a photo means dealing with both halves of what identifies it: the visible half — faces, license plates, addresses, anything legible — and the invisible half, the metadata that records where and when it was taken and on what device. This page does both in one pass, in your browser, and the photo is never uploaded.

Works today, in your browser. Photos open in the editor and Word documents are checked for hidden data; either way the file is read on your device and never uploaded.

Anonymize a photo now How we can promise that

What gets covered automatically

Open a photo and the identifying things are found and covered before you touch anything: faces (with background faces as their own group, so you can keep the subject and cover the crowd), license plates, street addresses and other legible text. Each category is a switch; anything missed is a drag of the mouse.

The metadata goes with it

The copy you download is drawn fresh from the covered picture, so it carries no EXIF, no GPS, no camera details — covering the face and leaving the coordinates in the file would be an odd place to stop.

What anonymization cannot promise

Honesty requires saying this plainly: detection is best-effort, and identity does not live only in faces and plates. A distinctive jacket, a reflection, a landmark, a tattoo — context can identify a person with every face covered. That is what the manual box is for, and why the result of a check is "no issues found", never a promise that nothing identifying remains. You know what is identifying in your situation; the tool knows the pixels.

How to do it

1Open the photoFaces, plates and addresses are found and covered automatically.
2AdjustKeep who should stay visible; drag a box over anything missed.
3DownloadSolid boxes or blur burned into the pixels, no metadata in the file.

Questions

For text — plates, addresses, card numbers — a solid box is the default precisely because a mild blur on a known font can sometimes be read back. For faces, blur or pixelation at the strength used here replaces the pixels; the original is not in the file. The edit is done to the picture itself, not as a removable layer.

We report no issues found, never “safe”. Absence of detections is not proof of absence, and detection is best-effort.