Blur & Pixelate Image

This tool blurs, pixelates or blacks out parts of a PNG, JPG, WebP, GIF, BMP or AVIF image in your browser. It is for hiding faces, licence plates, names, account numbers or addresses before you share a photo or screenshot, or for blurring a whole picture as a background — and the image is never uploaded, which matters most when the thing you are hiding is sensitive.

Drag on the picture to draw a box over each area to hide (up to 50 boxes), drag a box to move it or its handles to resize it, or tick "Whole image". Finding faces is manual: there is no automatic face detection, so you draw a box over each face yourself. The preview shows the effect live at screen size; the download is processed at full resolution.

There are three effects. Blur averages each pixel with its neighbours: three passes of a box blur of the strength you set (1–100 px), which approximates a Gaussian blur — at strength 12 the spread is about 12.5 px. Pixelate divides each box into squares of the block size (2–200 px), starting at the box's top-left corner, and paints every square with its average colour. Solid box replaces every pixel in the box with one opaque colour (black by default). Each effect reads only the pixels inside its own box, so nothing from inside leaks out past the edge.

Which one to use for sensitive information: Solid box. Blurring and pixelation keep a summary of the original pixels, and that summary can sometimes be reversed or read — a light blur of large text often stays legible, and pixelated text has been reconstructed by trying candidate text against the blocks (Depix is a published proof of concept that recovers text from pixelated screenshots). A solid box keeps nothing of what was underneath. Use blur and pixelation for faces in a crowd, backgrounds and anything where "hard to read" is enough, and make the strength or block size large compared with the detail you are hiding.

How it works: your browser decodes the image (phone photos are turned upright using their EXIF orientation), draws it onto a canvas at full size, and the effect is computed in JavaScript on the raw pixel values — not with the canvas filter feature, so it behaves the same in every browser. The result is encoded as a new file with the original pixels replaced; there are no layers to undo, and EXIF metadata — including GPS location — is not copied, which is a privacy gain in itself. PNG output is lossless; JPG and WebP are re-compressed at the quality you set (92 by default). GIF, BMP and AVIF are saved as PNG for "Same as input", and an animated GIF keeps only its first frame.

Limits: one image at a time, up to 50 MB and 100 megapixels (16,384 px per side). A single blur box can cover up to 40 megapixels, because the blur needs about 9 bytes of working memory per pixel; pixelate and solid box have no such limit, and a large blur shows its progress on the button. iPhones and iPads allow canvases of only about 16.7 megapixels, so very large photos may fail there with a message saying so.

Your image stays on your device. It is processed in the page's memory and offered back as a download; the Network tab in your browser's developer tools shows no request carrying it.

1. The file's size is checked (50 MB) before it is read and its first bytes confirm it is PNG, JPG, WebP, GIF, BMP or AVIF; the browser decodes it upright. 2. Each box is snapped to whole pixels and kept inside the image. 3. The image is drawn at full size and its RGBA pixels are read. 4. Blur (strength r): colours are premultiplied by alpha, then a running-average box blur of width 2r + 1 is applied horizontally then vertically, three times, with the edges clamped to the box; the result is converted back. The spread (standard deviation) is √(r(r + 1)) px. 5. Pixelate (block b): the box is split into b × b squares from its top-left corner (the last row and column may be narrower); each square gets its alpha-weighted average colour and its average opacity. 6. Solid box: every pixel in the box becomes the chosen colour, fully opaque. 7. The pixels are written back and the canvas is encoded, named like photo-blurred.jpg, photo-pixelated.jpg or photo-redacted.jpg.