How to Enlarge an Image Without Losing Quality (AI Upscaling)

Why Enlarging an Image Usually Makes It Blurry

When you enlarge an image in a classic editor, the software has to invent pixels that don't exist. It does this by interpolation — averaging the neighboring pixels. The result: an image that's bigger, but blurry, soft and sometimes pixelated. No detail is added; the existing ones are just stretched.

That's a real problem when you need to:

print an old photo or a picture found on a phone,
use a small logo on a banner or a poster,
put a low-resolution product image on a shop,
reuse a screenshot or an image from a video in a document.

How AI Upscaling Works

An AI upscaler doesn't average pixels: it uses a neural network trained on millions of images to recognize textures, edges and patterns, and to reconstruct plausible details. Edges stay sharp, textures (hair, foliage, fabric) look natural, and compression artifacts are reduced.

Plausible, not real

AI reconstructs details that are consistent with the image — it doesn't recover information that was never captured. A tiny, unreadable license plate won't become readable: the model will produce something plausible, not the real numbers.

Enlarge an Image with AI, in Your Browser

The AI image upscaler uses Real-ESRGAN, a reference model for image super-resolution. It runs entirely on your device, with WebGPU acceleration when your browser supports it: your images are never uploaded.

AI Image Upscaler

Enlarge your photos 2x or 4x with AI, in your browser.

Try it for free

Import your image — JPG, PNG or WebP.
Choose the model: Photo for real-world photos and portraits, Illustration for drawings, anime, cartoons and digital art.
Choose the factor: 2x or 4x.
Download as JPG (adjustable quality) or PNG (lossless).

The model (about 5 MB) is downloaded on your first visit and cached: the next times, the tool starts instantly — even offline.

2x or 4x?

Factor	Pixels	When to use it
2x	Width × 2, height × 2 (4 times more pixels)	Most cases: web, documents, prints up to A4
4x	Width × 4, height × 4 (16 times more pixels)	Very small images: icons, thumbnails, old photos for large prints

The input image is limited to 2000 px on its longest side (a larger image is reduced first, and the tool tells you). Beyond that, the memory needed on your device would explode — and an image that big rarely needs upscaling anyway.

How Many Pixels for Printing?

For a sharp print, aim for 300 pixels per inch (about 118 per cm):

Print size	Pixels needed (300 DPI)
10 × 15 cm (4 × 6 in)	1200 × 1800
A4 (21 × 29.7 cm)	2480 × 3508
A3 (29.7 × 42 cm)	3508 × 4961

A 900 × 1200 photo upscaled 2x (1800 × 2400) is enough for a nice 15 × 20 cm print.

Tips for the Best Results

Start from the best available version of the image — the original rather than a screenshot of it.
Choose the right model: the Illustration model on a photo gives a "painted" look; the Photo model on a drawing can add unwanted texture.
Crop first with the crop tool: upscaling only the useful area is faster and goes further within the 2000 px limit.
Then resize precisely with the image resizer if you need exact dimensions.

Conclusion

AI upscaling is the best way to enlarge an image while keeping it sharp: edges are clean, textures natural, and artifacts reduced. Choose the right model, prefer 2x unless the image is tiny, and keep your images private — everything runs in your browser.