Cookies for analytics and advertising
We use cookies for analytics and advertising, both sent to Google. Refusing changes nothing you can see.Read the privacy page
Give a width in pixels or a percentage and get the picture back at that size. The aspect ratio follows automatically, so nothing is stretched, and the whole thing happens on your own machine — the image is never uploaded.
Only applies to lossy formats such as JPG and WebP.
Where it runs
An image is read in the browser. Nothing is uploaded.
No queue, no account
It starts the moment you drop the file, and it never asks who you are.
No watermark
What comes out is the file you made, with nothing added to it.
Two things get called "too big" and they need opposite fixes. Too many kilobytes for an upload limit is a compression problem — the picture stays the same size on screen and is stored more tightly. Too many pixels is a dimensions problem: the image is 4000 across and the box it goes in is 400.
Resizing fixes the second and, as a side effect, fixes the first far more effectively than compression does. Halving the width quarters the pixel count, and the file follows. If your goal is simply a smaller file, resizing first and compressing afterwards will beat compressing alone at any quality setting.
Ask for 4000 pixels from a 900-pixel photograph and every tool that agrees is selling you an idea. The detail was never captured; enlarging produces a larger file containing the same information, blurred. Some tools now dress this up with machine learning, which invents plausible detail rather than recovering real detail — occasionally useful, never faithful, and not what somebody resizing a photo for a form is asking for.
So a request larger than the source is capped at the source, and the page tells you that is what happened rather than handing back an upscaled file and letting you discover it.
The height follows the width, always. Distorting an image by setting both edges independently is something people do by accident far more often than on purpose, and a tool that allows it without comment produces squashed faces that nobody notices until later.
If you genuinely need a fixed rectangle of a different shape — a square avatar from a landscape photo — that is a crop rather than a resize, and it needs a rectangle dragged over a preview to be usable at all. Typing four numbers to choose a crop is worse than not offering it, so it is not here.
No. The resize runs in your browser and the picture never leaves your device. You can watch the network tab while it happens, or disconnect from the internet and resize anyway.
Not beyond its original size, and that is deliberate. The detail was never captured, so enlarging gives you a bigger file containing the same picture, blurred. A request larger than the source is capped and the page says so.
Making an image smaller discards pixels by definition, but at the size it will actually be viewed the result looks identical — you are throwing away detail nobody was going to see. What does cost quality is re-encoding into a lossy format, which is the quality setting rather than the resize.
If the picture is displayed far smaller than it is stored, resize — it saves more than any quality setting and costs nothing visible. If the dimensions are already right and the file is still too large, compress. Doing both, in that order, beats either alone.
Not here. Choosing a crop by typing numbers is worse than not having the feature — it needs a rectangle dragged over a preview to be usable, which is a different piece of interface. Resizing keeps the whole picture and changes its size.