Advertisement
What this tool actually does — and what it does not
Worth being direct, because there is a lot of misleading marketing around this topic. This tool performs image restoration of the smoothing kind: it upscales with high-quality interpolation, removes the hard block edges that pixelation creates, then applies an edge-aware unsharp mask so the result does not look like a blurry smear.
It does not invent detail. If a face in your photo is 12 pixels wide, no algorithm running in a browser can reconstruct a real face — that needs a trained neural reconstruction model, which cannot run privately and instantly on your device. Anyone promising otherwise is showing you a demo that generated a plausible-looking stranger.
What it genuinely fixes
- JPEG block artefacts. The 8×8 grid you see in heavily compressed images. This is where the tool shines.
- Upscaled screenshots. Text and UI elements that turned into staircases.
- Thumbnails blown up for print. Softens the pixelation so a small print looks acceptable.
- Game and retro captures. Smooths the hard blocks while keeping the composition.
- Colour banding and dithering. Large flat areas get much cleaner.
Choosing your settings
| Control | Effect | Sensible range |
|---|---|---|
| Upscale factor | How much larger the output is. | 2x for print, 3x for screens |
| Block smoothing | How hard the square edges are erased. | 40–70% |
| Detail recovery | Unsharp mask that restores perceived crispness. | 30–60% |
Smoothing and detail recovery pull against each other. If the image looks plasticky, lower
smoothing. If it looks soft, raise detail recovery. Aim for the point where the blocks are gone but edges
still read as edges.
Practical workflow
- Start with a 2x upscale and 50% smoothing. Look at a textured area, not a flat one.
- Raise Detail recovery until edges snap back.
- If blocks reappear, nudge smoothing up by 10%.
- Export as PNG so the smoothing is not re-compressed away.
FAQ
How do I unpixelate an image for free?
Upload the blocky image, choose an upscale factor of 2x or 3x, raise Block smoothing until
the square edges soften, then add Detail recovery to bring back crispness. Download as PNG.
Can this recover a face from a blurred or pixelated photo?
No. This tool smooths and upscales existing pixels; it does not invent detail. Recovering real detail
from a heavily pixelated image requires AI reconstruction models — a different class of tool.
What is the difference between smoothing and detail recovery?
Block smoothing averages neighbouring pixels to remove the hard square edges.
Detail recovery is an unsharp mask that puts local contrast back so the result does not look
soft.
Why is my export capped in size?
Output is capped at 3200px on the long side to keep the browser responsive. A clean full-size download costs one credit.
Related tools
Advertisement