Free tool

AI image upscaler, with no upload.

Make a small or soft picture 2, 3 or 4 times larger with sharper edges and text than a plain resize. An open AI model, Real-ESRGAN, runs on your own graphics card in this tab, so your screenshots and photos never leave your computer. No sign-up, no watermark.

This runs in your browser. Your pictures are never uploaded.

How it works

Three steps, nothing to install.

01 / Drop

Choose a picture

Drag in a screenshot, a photo or a logo. The page shows how big each scale would be.

02 / Upscale

Pick 2x, 3x or 4x

The first time, the page loads the AI model, under 3 MB. Then your graphics card does the work, part by part.

03 / Compare

Check it and download

Drag the line to compare with a plain enlargement, change the strength if you like, then save.

Questions

Good to know.

Are my pictures uploaded?

No. Many online upscalers process your picture on their server. This one downloads the AI model to your browser instead and runs it there, so the picture never leaves your computer.

Which AI model does it use?

Real-ESRGAN General x4v3, an open model by Xintao Wang and others, released under the BSD licence. It was trained to turn small, blurry or compressed pictures into sharp ones four times larger. For 2x and 3x the page shrinks its 4x result smoothly.

Can it make text I cannot read readable?

No. An upscaler guesses detail that is not in the picture. Text that is soft but readable usually comes out crisp. Text too small to read can come out as made-up shapes, so do not trust it for numbers or names. It works best on pictures that are only a little too small.

How large can the result be?

Up to 16.7 million pixels, such as 4096 x 4096, which is the most every current browser can save. A 1920 x 1080 screenshot can go to 2x, 3840 x 2160. The page shows which scales fit your picture.

Does it keep transparency?

Yes. The see-through parts of a PNG or WebP are enlarged separately and put back, so a logo stays transparent.

What does AI strength do?

At 100% you get the AI picture alone. Lower values mix it with a plain smooth enlargement, which softens edges that look too hard or too bold.

Which browsers work?

It needs WebGL, which current Chrome, Edge, Safari and Firefox have. Without it, the page runs the model on the processor, which is much slower. Safari cannot save WebP files, so it saves those as PNG.