How it works : Crop an image
Draw the area to keep with a finger, a mouse or the keyboard: every step is announced. The file you get is re-encoded, so all the original metadata disappears, GPS position included.
Cropping is the most ordinary thing anyone does to an image, and it is also the most revealing: people crop a screenshot to show only part of it, a photo to remove what was around. In other words, cropping is almost always about HIDING something, and that something has no reason to travel through a server.
The area is drawn with a finger, a mouse or the keyboard, interchangeably: eight handles, arrow keys to move, Shift and arrows to resize, and every step is announced out loud for a screen reader. That is not an accessibility extra, it is the same path for everybody.
One thing worth knowing: unlike rotating a PDF page, cropping an image REDRAWS pixels. The file is therefore re-encoded, and a re-encoded JPEG loses a little quality even when nothing else changes. The good news about that re-encode is that it carries away ALL the original metadata, including the date, the camera, the GPS position and the embedded preview thumbnail.
How it works
- Drop your image in. It appears with a frame you can grab, move and resize.
- Set what you need: a fixed ratio if you want a square or a screen format, a quarter turn, a mirror flip.
- Click “Crop and download”. The result card states the exact area taken, the size obtained, and the file type READ BACK in the produced file.
Formats accepted
Any format your browser can display can be cropped. The OUTPUT format is limited to what your browser can write, and the type obtained is read back rather than assumed.
JPEG
Re-encoded: there is a loss, adjustable through the quality setting. Pick PNG as output to avoid any loss.
PNG
Re-encoded losslessly, transparency is kept.
WebP
Read everywhere, but NO version of Safari can WRITE a WebP: on an iPhone the output will be a PNG, and the tool will tell you instead of hiding it.
AVIF
Read by Chrome and Firefox. No browser writes it provably: the output will be PNG or JPEG.
Its limits
Cropping always re-encodes the image
A crop changes the file's pixels, so the file has to be rewritten. On a JPEG that re-encode costs a little quality even at maximum setting, because the compression is not reversible. That is a property of the format, not a choice of ours, and it is why PNG is offered as output.
16.7 million pixels maximum
A browser canvas has a ceiling, and the compatibility floor is 16.7 million pixels (iPhones before iOS 18). Beyond that the operation fails silently and returns a blank image. So the tool refuses cleanly, showing the dimensions it read.
All original metadata disappears
That is the consequence of re-encoding, and it is almost always what you want: date, camera, software, GPS position and embedded preview thumbnail are all lost. The browser does however put TWO blocks of its own back into a JPEG (a density tag and a colour profile): they say nothing about anyone, and we would rather name them than hide them.
The output format depends on your browser
The standard lets a browser return a PNG without warning when it cannot write the format asked for. No version of Safari can write a WebP. So the tool READS BACK the type of the file obtained and displays it: that is a hard rule of this project.
Questions about this tool
Is my image sent anywhere?
No, and you can check in thirty seconds. Open the inspector with the F12 key, “Network” tab, clear the list, then drop the image in and crop. Not a single line appears. Cut your connection and do it again: the tool behaves the same.
Can I crop with the keyboard?
Yes, entirely, and it is the same path as the mouse. The frame takes keyboard focus; arrow keys move it by one per cent of the image, Shift and arrows resize it, and every step is announced out loud with the exact size and position in pixels.
Do I lose quality?
A little, if you output JPEG: JPEG compression is not reversible, so re-encoding always costs something, even at maximum quality. Choose PNG as output and there is no loss at all, at the price of a heavier file. This differs from a PDF, where rotating a page touches no pixel.
Does my photo keep its GPS position?
No, and that is deliberate. Cropping re-encodes the image from a canvas, which makes all the original metadata disappear: date, time, camera, software, position, and even the preview thumbnail that cameras slip into the file. If you want to SEE what a photo says before sharing it, the “See and clean metadata” tool in the same corner shows it in detail.
I asked for WebP and got a PNG.
That is your browser, and the standard allows it: when it cannot write the format asked for, it returns a PNG WITHOUT warning. No version of Safari can write a WebP. So the tool reads back the type of the file obtained and shows it to you, instead of letting you believe you have a WebP.
Why is rotation applied after the crop?
Because that is what you expect while looking at your screen: the area you draw is the area of the image as displayed in front of you. Rotation comes next, on what you cut out, so it swaps the width and height of the result for a quarter turn. The result card gives both sizes.
Draw the area to keep with a finger, a mouse or the keyboard: every step is announced. The file you get is re-encoded, so all the original metadata disappears, GPS position included.
VerifyThe proof, step by stepThe network inspector check, explained one gesture at a time, with screenshots.
The cornerThe image cornerA photo says more than it shows. Here you see it, clean it, slim it down, crop it, and blur what is nobody's business.