Blur an Image or a Region

No. Blurring happens entirely in your browser using a box blur on the raw pixels. The image never leaves your device, which matters when you are hiding private information.

Yes. Switch to Region mode and drag a box over the area you want to soften, such as a face or an address. Only that rectangle is blurred, and edge pixels are clamped so no detail from outside the box leaks in.

For truly hiding a face or text, push the strength high, toward 40 pixels, so no readable detail survives. A light blur can still leave recognizable shapes and legible letters, so err on the heavier side.

No. A box blur averages each pixel with its neighbors and the original values are discarded in the downloaded file. There is no way to sharpen the exact detail back, which is what makes it safe for redaction.

It is the blur radius in pixels. A radius of 12 averages each pixel over a 25 pixel wide square, since the window is two times the radius plus one. A bigger radius means a wider average and a heavier blur.

The preview blurs a copy scaled to fit about 1000 pixels on its longest edge, with the radius scaled the same way, so it closely matches. The download applies the full radius to the full-resolution image.

A heavy blur is strong, but for the most sensitive text a solid fill or a coarse pixelate leaves even less to reconstruct. If security is critical, combine a high strength with a tight region over the exact area.

PNG (lossless), JPEG, and WebP. PNG keeps the blur clean without adding compression artifacts, which is a good default for redaction. The tool exports at high quality for the lossy formats.

Each download uses one region box. To hide several spots, blur one area and download, then load that result back in and blur the next area, repeating until every spot is covered.