Resize Images with Precision

The image resizer changes how many pixels a picture contains. That sounds obvious, but it is the one operation people most often confuse with compression: resizing alters the width and height, while quality settings alter how those pixels are stored. This tool is firmly about dimensions — give it a photo and it will produce the same picture at a size you specify.

There are two honest ways to express a new size. Pixel entry is absolute: when an upload form demands a 400 by 400 avatar or a print house wants 3000 pixels on the long edge, you type the number and get exactly that. Percentage scaling is relative: picking 50% halves both dimensions whatever the original was, which is the fastest way to make a stack of camera photos uniformly smaller without doing arithmetic. The aspect-ratio lock ties the two fields together so a width change recalculates the height automatically, preserving the proportions that stop faces and logos from looking squashed.

Direction matters more than people expect. Downscaling throws pixels away and almost always looks good, because the resampler has more information than it needs and can average neighbouring pixels into clean, smooth edges. Upscaling is the hard direction: there is no hidden detail to recover, so the tool interpolates new pixels from the ones around them. The high-quality smoothing used here keeps that interpolation as gentle as possible, but the realistic expectation for enlarging is a softer image, not a sharper one — which is why starting from the largest original you have always beats blowing up a small copy.

Common targets fall into a few buckets. Avatars and thumbnails want small, square or near-square dimensions; social platforms publish recommended sizes you can type in directly; print wants a high pixel count on the long edge so the dots-per-inch stay high at physical size. Because the whole operation runs locally, you can experiment with sizes freely and only keep the export you are happy with.