Upscale image to 2× or 4×

Why upscale with OptImg?

Enlarging an image normally just stretches its pixels into a blur. OptImg redraws edges and texture as it enlarges, so small photos, logos and thumbnails come out crisp at 2× or 4×. Online upscalers upload your photo to their servers and often keep the big result behind a paywall; OptImg does the same job on your device — nothing is uploaded, and the full-resolution result is free.

Frequently asked questions

Are my photos uploaded to a server?

No. The enlargement runs entirely in your browser: your image never leaves your device, and there's no queue and no account.

How is this better than normal resizing?

Normal resizing spreads the same pixels over a larger area, which blurs. This tool redraws edges and fine texture at the new size — the before/after slider shows the difference on your own image.

Why is the first image slower?

The upscaler itself is downloaded once (~9 MB) and kept by your browser. After that it starts instantly; the enlargement takes seconds on most computers and longer on slower devices.

Is there a size limit?

A practical one: inputs are capped at 8 megapixels for 2× and 2 megapixels for 4× on current Chrome, Edge and Safari, and 1 megapixel on browsers without the fast graphics path. Very large photos rarely need upscaling — resize them down first if you hit the cap.

What format is the result?

A high-quality JPG for photos, or a PNG whenever the original has transparency — always at the enlarged resolution. Run it through Compress if you want a smaller file.

Will this fix a blurry photo?

It depends on why the photo is blurry, and the distinction matters. If it is blurry because it is small — a thumbnail stretched too far, an old low-resolution file — then yes, this is exactly the case it was built for. If it is blurry because the camera missed focus or moved during the shot, then no: the detail was never captured, and enlarging a soft image produces a larger soft image. A quick test is to view the original at 100%: if it looks sharp at its own size and only falls apart when enlarged, upscaling will help.

What can it not recover?

Anything the original does not contain. Text that has degraded into a few grey smudges will come back as confident-looking but wrong letterforms, which is worse than leaving it blurry when the words matter. Faces at very small sizes may be reconstructed plausibly rather than accurately — fine for a background figure, not for identification. Heavy compression artefacts are sometimes amplified rather than removed. Treat the result as a better-looking image, not as new evidence about what was there.

Should I use 2× or 4×?

Start with 2×, and only go further if the result is still too small for its purpose. Each step of enlargement is more invention and less observation, so 4× looks noticeably more synthetic on the same source. If you need a specific final size, pick the setting that overshoots it least, then resize down — shrinking a slightly-too-large result is lossless in the way that stretching a too-small one is not.

Why does the result look smooth or 'painted'?

Because reconstructing detail and removing noise are the same operation seen from two sides. To decide that a rough patch is skin rather than sensor grain, the model has to commit to an interpretation, and where it guesses wrong you get the waxy, illustrated look. It is most visible on fine natural texture — hair, fabric weave, foliage, film grain — and on faces. The before/after slider is there precisely so you can decide whether the trade landed well on your image, rather than taking the enlargement on faith.