The whole privacy kit on one page
A photograph that is about to be published can give away more than a face: a name on a lanyard, a number on a form, an address on an envelope, and — in the file itself — the place and time it was taken. This page puts every tool for that job in one workspace, opened with the solid fill selected.
- Faces: press Find the faces and each one is proposed as a head-sized ellipse you can move, resize or delete.
- Text and numbers: drag a rectangle over each line, or trace it with the lasso if it sits at an angle.
- Coverings: solid fill, mosaic or blur per region. The fill is the default; the other two carry a note saying what they give up.
- Verdicts: every region in the list says what its covering is worth, and Before you save sums them up.
- Receipt: What the file carries lists the metadata in the photo you opened; after saving, the new file is read back and the result printed under Export.
A finder that found nothing has not proved there is nothing. Look at the saved copy before you post it.
Measured, not assumed
Saving decodes the file it just wrote and measures, inside each covered region only, how much detail is left within a cell — the whole region for a fill, one block for a mosaic, an eighth of the region for a blur. The verdict line then carries the figure before and after. A solid fill reads as zero apart from the encoder’s own grain; a blur reads as a percentage of what was there.
When another mode is better
For a group photo where only faces matter, blur face in a photo opens on ellipses. For a screenshot, paste it into redact a screenshot and copy the result straight back out. For a folder of photos, blur faces in many photos covers every face to one floor and measures each file.
Limits, stated plainly
- Text is not found automatically yet; draw over it.
- The face finder misses small, turned or half-hidden faces. Every miss is yours to draw.
- The site cannot tell you whether a redacted picture will be accepted by a bank, an employer or a court; ask them.
- An animated GIF or WebP is saved as its first frame.
Questions about redacting a photo
- What makes a redaction hold up?
- Three things: a covering that leaves nothing underneath (a solid fill, not a blur), a covering baked into the pixels rather than layered on top, and a file that no longer carries the metadata the camera wrote. This page does all three, and reports on each instead of asking you to take it on trust.
- What does the verdict on each region mean?
- Destroyed means the pixels under the region were replaced by one colour. Weak means the covering can be worked on — a blur can sometimes be run backwards, and a mosaic keeps large shapes. Before you save, the verdict is worked out from the covering's arithmetic; after you save, or press Measure on a test copy, it is read again out of the encoded file and prints the detail left inside each cell.
- Does it find text for me?
- Not yet. Faces are proposed by an on-device finder; text, numbers and names are covered by drawing over them, which takes one drag each. A finder that proposes text lines is planned, and when it ships it will propose rather than decide, like the face finder.
- Is the location removed from the photo?
- Yes. The saved copy is drawn from scratch from the pixels, so it carries no GPS block, capture time, camera or software name. After saving, the page reads the new file back with the same reader that listed what the original carried, and says what it found.
- Is the photo uploaded?
- No. Everything runs in this tab. The face finder is a 224 KB model downloaded once after you press its button; it comes to your device and the photo does not go anywhere.