How to resize an image without distorting it
The classic way to ruin an image is to type a width and a height that do not match its shape. Nothing here can do that to you — but understanding why explains most of what the tool does.
The numbers you type are limits, not targets
Give a width alone and the height follows from the image's own proportions. This is the common case and the one to reach for: "make it 1200 pixels wide" is almost always what someone means, and the tool works out the rest.
Give both a width and a height and the image is fitted inside that box rather than stretched to fill it. A 1600×1200 photo asked for 800×800 comes back 800×600 — it fits inside the square without becoming one. That is why nothing here can squash a face or stretch a logo: there is no setting that abandons the aspect ratio.
Because they are limits, a dimension that is already smaller than the limit is simply left alone. Sending a batch through a 1920-pixel cap will shrink the large photos and pass the small ones through untouched, which is exactly what you want when the batch came from more than one camera.
It never enlarges — and that is a feature
Ask for a cap larger than the image and the cap is ignored: a 400-pixel logo asked to become 4000 pixels comes back at 400. Nothing is lost by this, because there was nothing to gain. Enlarging cannot recover detail that was never recorded; it can only guess at pixels between the ones that exist, producing a bigger and softer version of the same picture.
This matters most when it is invisible. A tool that quietly upscaled would report success, hand back a file four times the size, and leave you to discover on a printed page that the logo is mushy. Refusing is the honest behaviour even though it looks like the tool did nothing.
If you genuinely need a larger image, the answer is upstream: export it again from the original, the design file, or the camera. Where the detail still exists is the only place it can come from.
Choosing the width: start from where it will be seen
The default is 1920 pixels wide, and it is chosen rather than inherited: full HD is the width most photos are actually displayed at, and it is a real reduction for anything off a modern phone or camera, which routinely produce images three or four times wider than any screen showing them.
For a website, the useful number is the widest the image will ever be drawn, doubled for high-density screens. An image that occupies a 600-pixel column needs about 1200 pixels and gains nothing above that. For an email attachment or a document, 1600 pixels is generous. For a profile picture, a few hundred is plenty.
The thing not to do is pick a number from the file size you want. Dimensions and file size are related but not the same lever — if the goal is a smaller file at the same dimensions, that is compression, and the quality presets here do it in the same pass.
Resize and convert in a single pass
A resize is always a re-encode: the pixels change, so the file has to be written again. Since that is happening anyway, the format is worth deciding at the same moment. Leave it as the source format, or pick JPEG, PNG, WebP or AVIF and get both changes from one job.
The quality presets govern that re-encode. High quality writes at quality 92 and keeps full colour detail; Balanced writes at 82; Maximum compression at 68 with more effort spent searching for savings. For a photo being scaled down for the web, Balanced plus a sensible width is usually the whole answer.
Doing it in one pass is not just convenience. Resizing to a new file and then converting that file means two lossy encodes stacked on top of each other, and the second one is working from the first one's compromises rather than from your original pixels.
What gets stripped, and why that is the default
Photos carry metadata you did not choose to attach: camera model, exposure settings, timestamps, and often the exact GPS coordinates where the shot was taken. Resizing strips that by default. It is not a size optimisation — it is that publishing a resized photo should not publish your home address along with it.
The colour profile is a different matter and is kept. Discarding it is what turns a correctly exposed photo into a washed-out or oversaturated one when it lands in a different application, so it stays even when everything else goes.
If you need the metadata — a photographer archiving work, a workflow that reads capture dates downstream — there is an option to keep it, and it is off unless you turn it on. Defaults should be the safe choice, not the convenient one.
Step by step
- 01
Add your image
Drop in a JPG, PNG, WebP or AVIF file up to 100 MB — up to 20 at once. Its current dimensions are read in your browser straight away.
- 02
Set a width, a height, or both
A width alone is usually right — the height follows the aspect ratio. Give both to fit the image inside a box. Optionally change the output format in the same job.
- 03
Resize and download
Run the job and download the result. Anything already smaller than your cap comes back unchanged. Files are deleted after 24 hours.
Frequently asked questions
Can I force an exact width and height?
Not if they disagree with the image's shape, no — and that is deliberate. Giving both fits the image inside those bounds with its proportions intact, so a 4:3 photo asked for a square comes back rectangular. Forcing exact dimensions would mean stretching, which is the one result nobody actually wants.
Why did nothing change when I asked for a bigger size?
Because the tool never enlarges. A cap above the source is dropped, and you get your image back at its original dimensions. Enlarging would interpolate pixels rather than recover detail, producing a bigger and softer file — so refusing is the honest outcome.
Does resizing remove my photo's location data?
Yes, by default. EXIF metadata — including GPS coordinates, camera model and timestamps — is stripped unless you explicitly ask to keep it. The colour profile is always preserved, since dropping that is what makes photos look washed out elsewhere.
Can I resize many images at once?
Yes — up to 20 files per job, each up to 100 MB, with the same cap applied to all of them. Images already under the cap pass through untouched, which is what makes a single width safe to use across a mixed batch.
Should I resize or compress to get a smaller file?
If the image is bigger than it will ever be displayed, resize first — removing pixels nobody sees is free. If the dimensions are already right, compress instead. Doing both in one job is fine too: pick a width and a quality preset together.