The number is not the point. The picture is.
Every compression tool gives you a slider and a file size. Almost none of them let you see what the slider is actually doing to your photograph.
So people pick a number that sounds safe, download it, and only discover the damage later — after it has been uploaded, printed, or sent.
Here the original and the compressed version sit in exactly the same place, and you drag a divider between them. Skies band. Skin goes blotchy. Text edges fur up. All of it is visible before you commit, at the setting you are actually considering.
That is the whole design. Two thumbnails side by side would not work, because at thumbnail size every setting looks fine.
When a file cannot get smaller, this says so
A photograph from a decent camera app has already been compressed once, and most of the redundancy is gone. Re-encode it at high quality and the result can be larger than the original — you are spending new bytes to describe the previous compression's artefacts.
Most tools hide this. They hand back a file named "compressed" and report a saving of zero percent, or quietly return your own file with a new name.
This one tells you the image is already efficient, shows you what it would take to actually go lower, and says plainly that downloading the result would make your file worse. A tool that will not admit failure cannot be trusted when it claims success.
WebP, measured on your image rather than asserted
WebP typically lands 25 to 35 percent below JPEG at matched visual quality. That is a well-known average, and averages are not much use when you have one specific photo in front of you.
So both formats are encoded and weighed, and the real difference for your image is put on screen. Usually WebP wins. Occasionally, on a small or already-optimised JPEG, it does not.
⚠️ WebP is offered, never chosen for you. Some government and exam portals still accept JPG only, and a rejected upload costs far more than the kilobytes saved.
It removes your location data
This is a side effect rather than a feature, but it is worth knowing.
Re-encoding through the browser discards all EXIF metadata. On most phones that includes the GPS coordinates of where the photo was taken, the device model, and the exact timestamp. If you are about to post a picture publicly, that matters.
It also means the orientation tag is discarded, which is why this tool reads the orientation before re-encoding — otherwise portrait photos would silently come out sideways.
What this does not do
- It does not resize. The pixel dimensions stay exactly as they were. If a form demands specific dimensions as well, use resize a photo to exact pixels and file size.
- It cannot make an already-optimised file meaningfully smaller without visible loss, and it will tell you so rather than pretending.
- It does not open HEIC. Convert an iPhone photo with HEIC to JPG first.