When is blur better than pixelation?
Blur is better when you want the hidden area to feel softer and less visually harsh while still making the sensitive detail unreadable.
Does this upload images anywhere?
No. The image stays in the browser while you blur the selected area and export the result.
Can I blur only one part of the image?
Yes. The tool uses rectangular selection, so you can blur exactly the part you want and leave the rest untouched.
Can blurred text be recovered?
Sometimes, yes. A Gaussian blur is a known, reversible transform. If an attacker knows the font, the approximate size and the blur radius, they can generate candidate strings, apply the same blur, and compare. It has been demonstrated repeatedly on redacted screenshots. For text, use an opaque box.
When is blur the right choice?
When you want to de-emphasise rather than hide: a busy background, a bystander's face in a crowd photo, a competitor's logo. If the consequence of someone recovering the content is serious, blur is the wrong tool.
Does the exported file still contain the original pixels?
No. The image is drawn into a canvas, the blur is painted over those pixels, and a new file is encoded from the result. There is no layer stack and no hidden original underneath, which is the difference between this and a black rectangle drawn in a PDF or a slide deck.
What blur radius should I use?
It defaults to 18 pixels. For obscuring a face at a distance that is plenty. For anything with structure, the radius needs to be large relative to the feature you are hiding, several times the height of a character if it is text.
Does the image get uploaded?
No. The image is decoded and processed on a canvas in your browser and downloaded directly. Nothing is sent to a server, which is the only sensible way to handle a file you are redacting.