What the AI image upscaler does
Enlarging a photo the ordinary way stretches the pixels you already have, so the result looks soft and blocky. This tool takes a different approach: it runs a super-resolution AI model that predicts the missing detail, so your image comes out larger and sharper rather than just bigger. It all happens inside your browser, so nothing is uploaded to a server.
How it works
- On-device AI: An ESRGAN super-resolution model reconstructs edges and texture as it scales your image up.
- 2x or 4x: Choose to double or quadruple each side. A 500 by 500 photo becomes 1000 by 1000 or 2000 by 2000.
- Private by design: Your photo is processed locally and never leaves your device.
How to use it
- Choose an image in JPG, PNG or WebP format.
- Pick a scale of 2x or 4x depending on how much larger you need it.
- Press Enhance and wait while the model loads and processes the image in tiles.
- Compare before and after, then download the result as a PNG.
Tips for the best results
- Start from the sharpest source you have; the model enhances detail but cannot invent what was never captured.
- Use 2x for large photos and 4x for small images or crops that need the most rescue.
- The first run downloads the AI model, so it takes longer; later upscales are quicker because the model is cached.
- Very large images are processed in small tiles, which keeps memory use low but takes more time.
Common uses
- Rescuing small or low-resolution photos for printing or presentations.
- Enlarging product shots, avatars and thumbnails without the blur.
- Bringing detail back to old scans and screenshots.