Upscale an image with AI
Make a small or blurry picture bigger without the soft, smeared look of ordinary resizing. A neural network trained on real photos rebuilds edges and texture as it enlarges — and it runs on your own device.
Your files never leave your browser
How to use
- 1
Choose how much to enlarge: ×2 or ×4.
- 2
Drop your JPG, PNG or WebP into the box. The first run downloads a small AI model once.
- 3
Download the enlarged image, or process several and grab a ZIP.
FAQ
How is this different from just resizing?
Ordinary resizing stretches the pixels you already have, so edges go soft and blurry. The AI model was trained on pairs of low- and high-resolution photos and predicts the missing detail, so lines stay crisp and textures look natural.
Is my image uploaded?
No. The model runs in your browser using WebAssembly; the image never leaves your device. The model itself (about 5 MB) is downloaded once and cached.
How large can the result be?
Up to 16 megapixels — for example a 1000×1000 image enlarged ×4, or a 2000×2000 image enlarged ×2. Larger images would exhaust the memory a browser tab is allowed to use.
How long does it take?
From a few seconds for a small icon to a minute or two for a large photo on a laptop. Phones are slower. A progress bar shows how far along it is.
Can it restore faces or remove compression artefacts?
It cleans up mild noise and JPEG blockiness as a side effect, but it is a general upscaler, not a face restorer. It cannot invent detail that was never in the picture, such as an unreadable licence plate.
Is transparency kept?
Yes. Transparent PNG and WebP images keep their transparency, enlarged to match.