What AI model does this use?

ESRGAN (Enhanced Super-Resolution Generative Adversarial Network), a state-of-the-art image super-resolution model. It runs locally via TensorFlow.js and WebGL acceleration.

Is my image uploaded to a server?

No. The ESRGAN model downloads once (~5 MB) and runs entirely in your browser using WebGL. Your images never leave your device.

How long does upscaling take?

After the model loads (first-time only), upscaling takes 5–30 seconds depending on your image size and GPU. Smaller images (under 500px) process in a few seconds.

What is the maximum input size?

The tool accepts images up to 10 MB. For very large images (4K+), processing may take longer or may hit browser memory limits — resize to 1080px first for faster results.

Which scale factor should I use?

2× is faster and sufficient for most uses — doubling a 500px image to 1000px. 4× is best for very small images (thumbnails, avatars, old photos) where you need maximum enlargement.

What output format should I choose?

PNG preserves every detail losslessly. JPEG is smaller if you're going to print or share. WebP gives the best file size for web use.

AI Image Upscaler

Enlarge images 2× or 4× using ESRGAN — a state-of-the-art AI super-resolution model that runs entirely in your browser.

Drop images here, or click to browse

JPG, PNG, WEBP · max 10 MB · single file only

How it works

1Upload an image — JPEG, PNG, or WebP, up to 10 MB.
2Choose 2× or 4× upscale factor. The AI model runs on your GPU via WebGL.
3On first use, the ESRGAN model (~5 MB) downloads once and is cached in your browser.
4Click Upscale — processing takes 5–30 seconds depending on image size and your device.
5Download the upscaled image as PNG, JPEG, or WebP.
6Everything runs in your browser. Your image is never uploaded to any server.