Blur Image

// Most people blurring an image are hiding something specific

The search says blur an image; the job is almost always narrower than that. A face in a photo you want to post. A number plate on a car for sale. A home address on a delivery screenshot, an account number on a statement, a colleague's name in a bug report. What you want is not a softer photo, it is one particular rectangle gone and everything else untouched.

So that is what this does. Drag across the part you want hidden and it is hidden immediately; drag again for a second area. Each area carries a small cross in its corner to take it back again, which also means a redaction can be undone without a mouse. There is no mode to switch into and no brush size to set first. The picture you are dragging on is the finished image, not a preview of one, so what you see is exactly what downloads.

// Blur, pixelate, or a solid box — and only one of them is final

Blur and pixelation are the familiar looks, and they share a weakness worth knowing before you rely on either. Neither deletes anything: both are reversible transforms applied to pixels that are still, mathematically, related to what was underneath. Researchers have recovered legible text from pixelated screenshots and recognisable faces from blurred ones, and the narrower the range of possible originals — a six-digit code, a licence plate, a name from a known list — the easier that gets.

A solid box is different in kind. It paints over the region, so the pixels that were there are replaced rather than rearranged, and there is nothing left to invert. If what you are covering genuinely must stay secret, that is the one to use. The page states this next to the style you have actually got rather than burying it, because a tool that lets you feel safe while being recoverable is worse than no tool.

Strength scales with the size of the area rather than being a fixed number of pixels. A blur that hides a face in a small photo would barely mark the same face in a 48 megapixel one, so the radius is derived from the region you drew.

// It never leaves your device, which is the entire point here

Every other tool in this category can afford to be relaxed about uploads. This one cannot: the images people bring to a redaction tool are passports, bank statements, medical letters, screenshots of private conversations, photos of their children. Sending those to a stranger's server to have a rectangle drawn on them is a worse outcome than the one you were trying to avoid.

The image is decoded, redacted, and re-encoded by your own browser. There is no upload step and no copy on any server, which is also why the result appears the instant you finish a drag rather than after a round trip. Once the page has loaded it keeps working with the network off, with one exception: an iPhone HEIC fetches its decoder the first time you drop one.

Saving re-encodes the picture through a canvas, and that drops the EXIF metadata along the way — the GPS coordinates, the timestamp, the camera serial number. Worth knowing, because a carefully blurred photo that still carries the coordinates of where it was taken has not really been anonymised.

// fair questions

How do I blur just one part of an image?

Drag across it. The area is hidden as soon as you release, and you can drag as many separate areas as you need. Each one gets a small cross in its corner that removes it again, so there is nothing to undo through a menu.

Can a blurred or pixelated image be un-blurred?

Sometimes, yes. Blur and pixelation transform the pixels rather than deleting them, and both have been reversed in published work — most easily when the hidden content is predictable, such as a plate number or a short code. If the information must stay secret, use the solid box option, which replaces the pixels outright and leaves nothing to recover.

Is this good enough for redacting a document?

With the solid box style, the pixels underneath are genuinely gone from the saved image. Two cautions: check the result at full size before sharing it, and remember that this covers what is visible in the picture — it cannot help with text stored elsewhere in a PDF or with a copy of the original you still have.

Does my photo get uploaded?

No. It is decoded and redacted by your browser on your own device, with no upload and no server copy. That matters more here than on most tools, because the photos people redact are exactly the ones they cannot afford to hand over.

Does saving remove location data from my photo?

Yes. The image is re-encoded through a canvas when you save, and that discards the EXIF block, including GPS coordinates, the capture timestamp, and camera details. A blurred photo that still carries its coordinates is not really anonymous, so this happens whichever style you pick.

Can I blur the whole image instead of one area?

Yes. Before you have marked anything, the picture carries a cover the whole image button that does it in one click. Dragging a box across the entire picture is the same thing by hand, and both accept the style and strength you have chosen.

What image formats can I use?

Anything your browser decodes — JPG, PNG, WebP, GIF, AVIF, BMP — plus HEIC and HEIF from an iPhone, via a decoder fetched only when one is dropped. You can save as JPG, PNG, or WebP. An animated GIF is read as a still, so you get its first frame.