Why resize images locally instead of through a website upload?

Typing in pixel dimensions or a percentage and getting a resized file back shouldn't require sending the original anywhere. NearIMG's resize runs on your device's own processor, so the scaling happens the moment you set the numbers — no queue, no upload, no copy left on a server.

How it works

1. Add your images and choose how to specify the size: exact pixels, or a percentage of the original.
2. Enter a width and/or height. With "keep aspect ratio" on, one side is derived from the other, and giving both sides fits the image inside that box without distorting it.
3. Click Resize — the scaling runs on your device with high-quality smoothing, and the result downloads immediately.

Downscaling is the safe direction: the browser's high-quality resampling merges pixels and the result stays sharp. Enlarging is different — a plain resize cannot invent detail, so upscaled photos look soft. That's what the optional AI upscale mode is for: it runs a Real-ESRGAN super-resolution model in your browser via WebAssembly (a one-time ~5 MB model download from this site, never a third-party server) and predicts plausible detail instead of stretching pixels. It's slower than a normal resize and, being a prediction, works best on small photos rather than on text.

Will resizing stretch my image?

Keep "aspect ratio" on and it scales proportionally. Enter both width and height only when you need an exact size.

Can I enlarge an image?

You can scale up, but a browser cannot invent detail — enlarging a small image looks softer than the original.

What does the "AI upscale" option actually do?

It runs a super-resolution neural network (Real-ESRGAN) locally in your browser to add predicted detail at 200% or 400%. The model file is fetched once from this site and inference happens on your device — the image still never leaves it. The result is saved as PNG.