Skip to content

URL Options

This document lists and describes the full list of options available to be passed as parameters in the URL.

The server is setup to perform operations based on the following URL pattern.[process-type]/[image_options]/[path_to_image]

for example:


process-type: upload or path

There are 2 main proceses you can do for images.

The first: upload, grabs an image from a URL, transforms it, saves it, and serves it, as an image, with all the appropriate headers.

The second: path, grabs an image from a URL, transforms it, saves it, and returns the absolute path to the image as a string, in the body of the response.


It's the first operation the server does, it will try to get an image from the URI in the path, it can be relative to the local server, or absolute to the internet.



Here you set all the transformations and output settings you want to apply to the image you are fetching.

Most of these options are ImageMagick flags, many can get pretty advanced, use the ImageMagick docs. We put a lot of defaults in place to prevent distortion, bad quality, weird cropping and unwanted paddings.

The script does a lot of sanitizing of the parameters, so many options will not work or have to be carefully escaped. Priority is given to safety and eas of use.

Basic geometry

w : width

int Default: null Description: Sets the target width of the image. If not set, width will be calculated in order to keep aspect ratio.


w_100 :


h : height

int Default: null Description: Sets the target height of the image. If not set, height will be calculated in order to keep aspect ratio.


h_100 :


Using width AND height

example:h_300,w_300 By default setting width and height together, works like defining a rectangle that will define a max-width and max-height and the image will scale propotionally to fit that area without cropping.

By default; width, height, or both will not scale up an image that is smaller than the defined dimensions.

h_300,w_300 :,w_300/


c : crop

bool Default: false Description: When both width and height are set, this allows the image to be cropped, so it fills the width x height area.


c_1,h_400,w_400 :,h_400,w_400/


g : gravity

string Default: Center Description: When crop is applied, changing the gravity will define which part of the image is kept inside the crop area. The basic options are: NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast.


r : rotate

string Default: null Description: Apply image rotation (using shear operations) to the image.

example: r_90, r_-180,...

r_45 :,w_400,h_400/


Output file formats

o : output

string Default: auto Description: Output format requested, for example you can force the output as jpeg file in case of source file is png. The default auto will try to output the best format for the requesting browser, falling back to the same format as the source image or finally with a fallback to jpg.

If input is passed, no "optimal" format will be attempted. Flyimg will try to respond with the source format or fallback to jpg.


q : quality

int (0-100) Default: 90 Description: Sets the compression level for the output image. Your best results will be between 70 and 95.


q_30 :,w_600/


q_100 :,w_600/


webpl : webp-lossless

int Default: 0 Description: If output is set to webp, it will default to lossy compression, but if you want lossless webp encoding you have to set it to 1.


Refresh or re-fetch source image

rf : refresh

Default: false Description: When this parameter is 1, it will force a re-request of the original image and run it through the transformations and compression again. It will delete the local cached copy.


The nginx server will send headers to prevent caching of this request.

It will also send headers with the command done on the image + info returned by the command identity from Imagemagick.

Fancy options

bg : background

color (multiple formats) Default: null Description: Sets the background of the canvas for the cases where padding is added to the images. It supports hex, css color names, rgb. Only css color names are supported without quotation marks. For the hex code, the hash # character should be replaced by %23


  [...] -background red ...
  [...] -background "#ff4455" -> "%23ff4455"
  [...] -background "rgb(255,120,100)" ...,w_400,h_400,bg_red/


st : strip

int Default: 1 Description: removes exif data and additional color profile. Leaving your image with the default sRGB color profile.


ao : auto-orient

int Default: 0 Description: Adjusts an image so that its orientation is suitable for viewing (i.e. top-left orientation).


rz : resize

int Default: null Description: The alternative resizing method to -thumbnail.


moz : mozjpeg

Default: 1 Description: Use moz-jpeg compression library, if 0 it fallsback to the default ImageMagick compression algorithm.

unsh : unsharp

{radius}x{sigma}[+gain][+threshold] Default: null Description: Sharpens an image (despite the unfortunate naming). It basically sharpens via subtracting a blurred, low contrast, version of the image to the image itself. For more details check Imagemagick docs.

example 1:unsh_0x6
example 2:unsh_0.25x0.25+8+0.065

sh : sharpen

{radius}x{sigma} Default: null Description: Use a Gaussian operator of the given radius and standard deviation (sigma). For more details check Imagemagick docs.

example 1:sh_3
example 2:sh_0x5

blr : blur

{radius}x{sigma} Default: null Description: Apply Blur on a image. For more details check Imagemagick docs.

example 1:blr_2
example 2:blr_1x2

f : filter

string Default: Lanczos Description: Resizing algorithm, Triangle is a smoother lighter option


sc : scale

Default: null Description: The "-scale" resize operator is a simplified, faster form of the resize command. Useful for fast exact scaling of pixels.


fc : face-crop

int Default: 0 Description: Using facedetect repository to detect faces and passe the coordinates to ImageMagick to crop.


fc_1 :


fcp : face-crop-position

int Default: 0 Description: When using the Face crop option and when the image contain more than one face, you can specify which one you want get cropped


fcp_2 :,fcp_2/


fb : face-blur

int Default: 0 Description: Apply blur effect on faces in a given image


fb_1 :


e : extract

p1x : extract-top-x

p1y : extract-top-y

p2x : extract-bottom-x

p2y : extract-bottom-y

Default: null

Description: Extract and crop an image with given the x/y coordinates of each booth top and bottom.

sf : sampling-factor

Default: 1x1 Description: ...

ett : extent

Default: null Description: ... not ready

par : preserve-aspect-ratio

int Default: 1 Description: If set to 0, when passing width and height to an image, the image will be distorted to fill the size of the rectangle defined by width and height.

pns : preserve-natural-size

int Default: 1 Description: If set to 0 and if the source image is smaller than the target dimensions, the image will be stretched to the target size.

gf : gif-frame

int Default: 0 Description: ...

PDF options

Requires ghostscript installation in the Dockerfile.

pg : page number

Default: 1
Description: Sets the target page of the PDF. If not set, the default is page 1.


pg_2 :

Video options

Requires ffmpeg installation in the Dockerfile.

tm : time

Default: 00:00:01
Description: Sets the frame capture time duration point in the video. If not set, the default is 1 second. The format is HH:MM:SS OR SS


tm_00:00:05 :

You can also use a shorter syntax for the first 60 seconds.


tm_10 :

Options keys

  moz: mozjpeg
  q: quality
  o: output
  unsh: unsharp
  fc: face-crop
  fcp: face-crop-position
  fb: face-blur
  w: width
  h: height
  c: crop
  bg: background
  st: strip
  rz: resize
  g: gravity
  f: filter
  r: rotate
  sc: scale
  sf: sampling-factor
  rf: refresh
  ett: extent
  par: preserve-aspect-ratio
  pns: preserve-natural-size
  webpl: webp-lossless

Default options values

  mozjpeg: 1
  quality: 90
  output: auto
  unsharp: null
  face-crop: 0
  face-crop-position: 0
  face-blur: 0
  width: null
  height: null
  crop: null
  background: null
  strip: 1
  resize: null
  gravity: Center
  filter: Lanczos
  rotate: null
  scale: null
  sampling-factor: 1x1
  refresh: false
  extent: null
  preserve-aspect-ratio: 1
  preserve-natural-size: 1
  webp-lossless: 0