Three ways to hide part of a photo, and when to use each
| Method | Look | Best for |
|---|---|---|
| Pixelate | Blocky, unmistakably "this was redacted" | Published photos where the redaction should be obvious |
| Blur | Soft, less jarring | Casual photos, background faces in a group shot |
| Solid colour | A flat block, nothing recoverable | Genuinely sensitive content - document numbers, anything that must not be reconstructable |
Pixelation and blur are not perfectly irreversible
This is worth being upfront about. Heavy pixelation with large blocks is difficult to reverse in any practical sense, but averaging pixels together isn't the same as deleting them - research on de-pixelation exists, and it works best against small block sizes over simple, low-detail content. Blur has a similar limitation. If what's underneath genuinely cannot be allowed to leak - a face that must stay anonymous, an account number - use the solid colour option instead. A flat fill discards the original pixels entirely; there's nothing to average back.
Working with multiple areas
For a single area - the common case, one face or one licence plate - just drag the box into place and run the tool; there's no extra step required. For more than one, position the box over the first spot and click "Add this area" before moving on to the next. Each added area appears in a short list below the controls, with its own remove button if you placed one wrong. Everything in that list gets censored the same way, with the same method and strength, when you run the tool.
Choosing a block size or blur radius
Both the pixelate and blur strength values work the same way: larger numbers hide more detail but also cover more of the surrounding area with visible effect. For a face at typical photo resolution, something in the 12-24px range usually removes recognisable detail without needing an enormous box. Very small, high-detail subjects - text in a screenshot, a small distant face - may need a larger block size relative to the box size to actually obscure the content, since a block size close to the box's own dimensions barely changes anything.
This works on more than faces
The box has no idea what's underneath it - it censors whatever pixels fall inside the rectangle you drew. That makes it equally useful for a licence plate, a street address on an envelope, a signature, an account number in a screenshot, or a name tag in a group photo you're about to share publicly.
Nothing is uploaded to do this
The whole process - drawing the box, applying the pixelation, blur or fill - happens entirely inside your browser using the Canvas API. Open your network tab, or disconnect from the internet after the page loads - the tool keeps working, because the photo never leaves your device. That matters particularly here: a tool that censors sensitive content for privacy reasons shouldn't be sending that same content to a server in the process.