Resize image
Set pixel dimensions, resize in batches, and lock the original ratio.
How it works
Change dimensions while keeping the ratio.
- 01
Choose the images to resize.
- 02
Enter the dimensions; locked ratio keeps each image proportional.
- 03
Check the output dimensions and download the new files.
FAQ
What is the difference between resizing and compression?
Resizing changes pixel dimensions. Compression mainly changes file size. Use compression afterward when an upload has a KB limit.
How do I avoid distortion?
Keep Lock original ratio enabled. Each image will calculate its own height from the chosen width.
Does enlarging improve sharpness?
No. Browser resizing cannot restore missing detail. Enlarged images may look softer.
Dimensions, ratio, and distortion
Resizing means changing the pixel width and height. The easiest mistake is distortion: force a 4:3 photo into a 1:1 square and the people and objects in it get squashed or stretched. To avoid that, keep Lock original ratio enabled — then you set only a width, the height is computed from the original ratio, and nothing warps.
Turn locking off only when you genuinely need a fixed width and height and can accept the distortion. A more reliable approach is to resize close to your target first, then use the crop tool to cut the exact aspect ratio — that way you get a definite size without warping.
How each image is handled in a batch
When you select several images at once, their original sizes usually differ. The tool uses your chosen width as a common baseline and computes each image's height from its own ratio. So a mix of landscape and portrait images processed together comes out at a uniform width, none of them distorted — handy for tidying an album or standardising asset widths. If you need every image at identical width and height, turn locking off and set both manually, but expect images with a different ratio to distort.
Enlarging will not add sharpness
Many people assume blowing up a small image yields a high-resolution one. It does not. The browser's scaling algorithm can only interpolate new pixels from existing ones; it cannot invent detail the original never recorded. Enlarge 400×400 to 2000×2000 and you get a blurry, bigger version of the same image. True lossless upscaling needs an AI super-resolution model, which is a different class of tool. This tool is for shrinking, or modest adjustments near the original size; when you need something bigger and sharper, go back to the source for a higher-resolution original.