Image Crop
Crop to a ratio, or to the pixel
Drag a crop box or type exact coordinates, lock it to a common aspect ratio, and export the result.
Image Crop
Runs in your browser · nothing uploaded01 — How to use Image Crop
02 — Getting good results
03 — Questions
No. It is decoded and re-encoded in a background worker inside this tab, and the result is handed straight to a download. No network request happens during the crop.
Because locking a ratio means the box must satisfy it. The box is reshaped around its existing centre and kept as close to its previous size as the image allows, which preserves your framing better than resetting to the middle would.
No — a crop is a decision about one particular image, and applying the same rectangle to a batch rarely produces good results. For batch work with a fixed output size, the Social Media Resizer crops to fill each target automatically.
No. Re-encoding through a canvas drops all metadata, including any location data the original carried. That is usually welcome for anything being published; if you need the metadata, read it first with the EXIF Viewer.
04 — More about Image Crop
Aspect ratios and where they come from
The ratios offered here are not arbitrary. 3:2 is the shape of 35mm film and most interchangeable-lens cameras. 4:3 comes from early television and is still the native shape of many phone sensors and most tablets. 16:9 was chosen as a compromise between the cinema formats and 4:3 when widescreen television was standardised, and it became the default for video and monitors. 1:1 has no photographic ancestry at all — it came from Instagram, which adopted it to make a grid of images look orderly on a small screen, and it has since become a genuine convention for profile images and product photography. Knowing which shape a destination expects saves cropping twice.
Coordinates, display size, and why they differ
The crop box is stored in the image’s own pixel coordinates, not in the coordinates of the picture as displayed. Those are almost never the same: a 4000-pixel-wide photograph shown in a 600-pixel-wide panel is scaled by a factor of nearly seven, so a drag of ten screen pixels moves the box by nearly seventy image pixels. Converting between the two on every pointer event is what allows the numeric fields and the drag interaction to agree, and it is why the exported crop is exactly the region described by the numbers rather than approximately it.
Cropping is lossless; saving is not
Removing pixels from an image loses nothing about the pixels that remain — the crop itself is an exact operation. What loses quality is writing the result back out in a lossy format, because the encoder makes fresh approximations that do not align with the ones already baked into the source. The practical rule is to minimise the number of times an image is encoded: crop once from the original rather than repeatedly from previous exports, and choose a lossless format such as PNG if the file is an intermediate step rather than a final one.
61 more tools
where this came from.
Same shell, same shortcuts, same promise: nothing leaves your browser.