# Crop

Controls how the image is aligned.

# Alignment position &a=

How the image should be aligned when &fit=cover or &fit=contain is set. The &w= and &h= parameters should also be specified.

# Position-based

Controls the starting location of the crop. When &fit=contain is set, it determines how the image is positioned inside its box.

Valid positions:

  • center: default
  • top
  • right
  • bottom
  • left
  • top-left
  • bottom-left
  • bottom-right
  • top-right

For more information, please see the suggestion on our GitHub issue tracker: Issue #24 - Aligning(opens new window).

Alignment position

# Focal point &a=focal

You can be more specific about the alignment using a focal point. This can be set using a horizontal (&fpx=) and vertical (&fpy=) offset decimal value (a float between 0.0 and 1.0, inclusive). The default value is 0.5, or the center of the image.

Focal point

# Smart crop &a=[entropy,attention]

An experimental strategy-based approach to crop the image by removing boring parts. This only works with &fit=cover.

More info: Issue #90 - Add support for smart crop(opens new window).

Valid strategies:

  • entropy: focus on the region with the highest Shannon entropy(opens new window).
  • attention: focus on the region with the highest luminance frequency, color saturation and presence of skin tones.

Smart crop

# Rectangle crop &c[x,y,w,h]=

Crops the image to specific dimensions after any other resize operations.


You can use &precrop, for a pre-resize crop behaviour. See Issue #176 - Combine cropping with resizing(opens new window) for more information.

Rectangle crop

# Trim &trim=

Trim "boring" pixels from all edges that contain values within a similarity of the top-left pixel. Trimming occurs before any resize operation. Use values between 1 and 254 to define a tolerance level to trim away similar color values. You also can specify just &trim, which defaults to a tolerance level of 10.

More info: Issue #39 - able to remove black/white whitespace(opens new window).