Image Crop
Crop an image to a specified rectangle.
Overview
The Image Crop tool extracts a rectangular region from an uploaded image using x, y, width and height values you specify. Upload a photo, type the crop coordinates, and the tool returns the cropped result in the same format and at the same bit depth as the source.
It is useful when you have to trim a screenshot down to a single panel, prepare a square avatar from a wider portrait, or batch-cut a region out of dozens of camera frames where a manual editor would be too slow. The pixel-accurate inputs also suit workflows where the crop rectangle comes from another tool such as an object detector.
How it works
A crop is the cheapest of all raster operations: the decoder allocates a pixel buffer for the requested rectangle, copies the rows of the source that fall inside that rectangle one scanline at a time, and re-encodes the buffer with the same codec used by the input. No interpolation runs and no pixels are resampled.
Coordinates use the standard image convention — origin at the top-left, x increasing rightwards, y increasing downwards. If the requested rectangle extends past the image edges the tool clamps to the visible bounds and reports the effective crop, so an out-of-range box never produces garbage data.
Examples
Before: screenshot.png (1920x1080)
Crop: x=120, y=80, w=800, h=600
After: screenshot.png (800x600)
Before: portrait.jpg (3000x4500)
Crop: x=750, y=600, w=1500, h=1500
After: portrait-square.jpg (1500x1500)
Before: scan.tiff (5000x7000, full page)
Crop: x=400, y=200, w=4200, h=6600
After: scan-trimmed.tiff with margins removed
FAQ
Does cropping lose quality?
No — the kept pixels are copied byte-for-byte, and the codec re-encodes them at the same quality as the input. Cropping a JPEG and re-saving does introduce one re-encode cycle; if that worries you, use a tool that performs a lossless JPEG crop on 8x8 block boundaries.
What happens if my rectangle is bigger than the image?
The crop is clamped to the image bounds. You get the largest valid subrectangle rather than an error.
Can I crop to a target aspect ratio?
Compute the width and height from the ratio (e.g. for 16:9 with width 1280 use height 720) and provide them as inputs. The tool itself takes literal pixel coordinates.
Will EXIF data survive?
Metadata is preserved except for tags that describe the original frame dimensions, which the encoder rewrites to match the new size.
Does it handle animated GIFs?
Yes — every frame is cropped to the same rectangle and the timing of the resulting GIF is preserved unchanged.