Processing an image
Processing images with imgproxy is really easy: just send an HTTP GET
request to imgproxy and imgproxy will respond with a processed image.
imgproxy responds with an image file and uses an image/*
Content-Type
header, so you can put imgproxy's URLs right into src
attributes of your <img/>
HTML tags or into url()
CSS values
The request URL path should consist of a signature, processing options, a source URL, and optionally an extension, like this:
http://imgproxy.example.com/%signature/%processing_options/plain/%source_url@%extension
http://imgproxy.example.com/%signature/%processing_options/%encoded_source_url.%extension
http://imgproxy.example.com/%signature/%processing_options/enc/%encrypted_source_url.%extension
Check out the example at the end of this guide.
Signature
The signature part should always be present in a URL. If the signature check is disabled (no key/salt pairs are provided), the signature part may contain anything (for example, unsafe
or _
).
A signature protects your URL from being altered by an attacker. It is highly recommended to sign imgproxy URLs when imgproxy is being used in production.
Once you set up your URL signature, check out the Signing the URL guide to find out how to sign your URLs. Otherwise, since the signature still needs to be present, feel free to use any string here.
Processing options
Processing options should be specified as URL parts divided by slashes (/
). A processing option has the following format:
%option_name:%argument1:%argument2:...:%argumentN
You can redefine the arguments separator with the the IMGPROXY_ARGUMENTS_SEPARATOR config. For example, if you set IMGPROXY_ARGUMENTS_SEPARATOR=","
, processing options will have the following format:
%option_name,%argument1,%argument2,...,%argumentN
The list of processing options does not define imgproxy's processing pipeline. Instead, imgproxy already comes with a specific, built-in image processing pipeline for maximum performance. Read more about this in the About processing pipeline guide.
imgproxy supports the following processing options:
Resize
resize:%resizing_type:%width:%height:%enlarge:%extend
rs:%resizing_type:%width:%height:%enlarge:%extend
This is a meta-option that defines the resizing type, width, height, enlarge, and extend. All arguments are optional and can be omitted to use their default values.
Size
size:%width:%height:%enlarge:%extend
s:%width:%height:%enlarge:%extend
This is a meta-option that defines the width, height, enlarge, and extend. All arguments are optional and can be omitted to use their default values.
Resizing type
resizing_type:%resizing_type
rt:%resizing_type
Defines how imgproxy will resize the source image. Supported resizing types are:
fit
: resizes the image while keeping aspect ratio to fit a given size.fill
: resizes the image while keeping aspect ratio to fill a given size and crops projecting parts.fill-down
: the same asfill
, but if the resized image is smaller than the requested size, imgproxy will crop the result to keep the requested aspect ratio.force
: resizes the image without keeping the aspect ratio.auto
: if both source and resulting dimensions have the same orientation (portrait or landscape), imgproxy will usefill
. Otherwise, it will usefit
.
Default: fit
Resizing algorithm pro
resizing_algorithm:%algorithm
ra:%algorithm
Defines the algorithm that imgproxy will use for resizing. Supported algorithms are nearest
, linear
, cubic
, lanczos2
, and lanczos3
.
Default: lanczos3
Width
width:%width
w:%width
Defines the width of the resulting image. When set to 0
, imgproxy will calculate width using the defined height and source aspect ratio. When set to 0
and resizing type is force
, imgproxy will keep the original width.
Default: 0
Height
height:%height
h:%height
Defines the height of the resulting image. When set to 0
, imgproxy will calculate resulting height using the defined width and source aspect ratio. When set to 0
and resizing type is force
, imgproxy will keep the original height.
Default: 0
Min width
min-width:%width
mw:%width
Defines the minimum width of the resulting image.
When both width
and min-width
are set, the final image will be cropped according to width
, so use this combination with care.
Default: 0
Min height
min-height:%height
mh:%height
Defines the minimum height of the resulting image.
When both height
and min-height
are set, the final image will be cropped according to height
, so use this combination with care.
Default: 0
Zoom
zoom:%zoom_x_y
z:%zoom_x_y
zoom:%zoom_x:%zoom_y
z:%zoom_x:%zoom_y
When set, imgproxy will multiply the image dimensions according to these factors. The values must be greater than 0.
Can be combined with width
and height
options. In this case, imgproxy calculates scale factors for the provided size and then multiplies it with the provided zoom factors.
Unlike the dpr
option, the zoom
option doesn't affect gravities offsets, watermark offsets, and paddings.
Default: 1
Dpr
dpr:%dpr
When set, imgproxy will multiply the image dimensions according to this factor for HiDPI (Retina) devices. The value must be greater than 0.
The dpr
option affects gravities offsets, watermark offsets, and paddings to make the resulting image structures with and without the dpr
option applied match.
Default: 1
Enlarge
enlarge:%enlarge
el:%enlarge
When set to 1
, t
or true
, imgproxy will enlarge the image if it is smaller than the given size.
Default: false
Extend
extend:%extend:%gravity
ex:%extend:%gravity
- When
extend
is set to1
,t
ortrue
, imgproxy will extend the image if it is smaller than the given size. gravity
(optional) accepts the same values as the gravity option, exceptsm
,obj
, andobjw
. Whengravity
is not set, imgproxy will usece
gravity without offsets.
Default: false:ce:0:0
Extend aspect ratio
extend_aspect_ratio:%extend:%gravity
extend_ar:%extend:%gravity
exar:%extend:%gravity
- When
extend
is set to1
,t
ortrue
, imgproxy will extend the image to the requested aspect ratio. gravity
(optional) accepts the same values as the gravity option, exceptsm
,obj
, andobjw
. Whengravity
is not set, imgproxy will usece
gravity without offsets.
Default: false:ce:0:0
Gravity
gravity:%type:%x_offset:%y_offset
g:%type:%x_offset:%y_offset
When imgproxy needs to cut some parts of the image, it is guided by the gravity option.
type
- specifies the gravity type. Available values:no
: north (top edge)so
: south (bottom edge)ea
: east (right edge)we
: west (left edge)noea
: north-east (top-right corner)nowe
: north-west (top-left corner)soea
: south-east (bottom-right corner)sowe
: south-west (bottom-left corner)ce
: center
x_offset
,y_offset
- (optional) specifies the gravity offset along the X and Y axes:- When
x_offset
ory_offset
is greater than or equal to1
, imgproxy treats it as an absolute value. - When
x_offset
ory_offset
is less than1
, imgproxy treats it as a relative value.
- When
Default: ce:0:0
Special gravities:
gravity:sm
: smart gravity.libvips
detects the most "interesting" section of the image and considers it as the center of the resulting image. Offsets are not applicable here.gravity:obj:%class_name1:%class_name2:...:%class_nameN
: pro object-oriented gravity. imgproxy detects objects of provided classes on the image and calculates the resulting image center using their positions. If class names are omited, imgproxy will use all the detected objects. Also, you can use theall
pseudo-class to use all the detected objects.gravity:objw:%class_name1:%class_weight1:%class_name2:%class_weight2:...:%class_nameN:%class_weightN
: pro object-oriented gravity with weights. The same asgravity:obj
but with custom weights for each class. You can use theall
pseudo-class to set the weight for all the detected objects. For example,gravity:objw:all:2:face:3
will set the weight of all the detected objects to2
and the weight of the detected faces to3
. The default weight is1
.gravity:fp:%x:%y
: the gravity focus point.x
andy
are floating point numbers between 0 and 1 that define the coordinates of the center of the resulting image. Treat 0 and 1 as right/left forx
and top/bottom fory
.
Objcts position pro
objects_position:%type:%x_offset:%y_offset
obj_pos:%type:%x_offset:%y_offset
op:%type:%x_offset:%y_offset
When imgproxy needs to cut some parts of the image, and the obj
/objw
gravity is used, the objects_position
option allows you to adjust the position of the detected objects on the resulting image.
type
- specifies the position type. Available values:no
: north (top edge)so
: south (bottom edge)ea
: east (right edge)we
: west (left edge)noea
: north-east (top-right corner)nowe
: north-west (top-left corner)soea
: south-east (bottom-right corner)sowe
: south-west (bottom-left corner)ce
: center
x_offset
,y_offset
- (optional) specifies the position offset along the X and Y axes.
Default: ce:0:0
Special positions:
objects_position:fp:%x:%y
: the focus point position.x
andy
are floating point numbers between 0 and 1 that define the coordinates of the center of the objects' area in the resulting image. Treat 0 and 1 as right/left forx
and top/bottom fory
.objects_position:prop
: the proportional position. imgproxy will try to set object offsets in the resulting image proportional to their offsets in the original image. This position type allows the picture scene to be maintained after cropping.
Crop
crop:%width:%height:%gravity
c:%width:%height:%gravity
Defines an area of the image to be processed (crop before resize).
width
andheight
define the size of the area:- When
width
orheight
is greater than or equal to1
, imgproxy treats it as an absolute value. - When
width
orheight
is less than1
, imgproxy treats it as a relative value. - When
width
orheight
is set to0
, imgproxy will use the full width/height of the source image.
- When
gravity
(optional) accepts the same values as the gravity option. Whengravity
is not set, imgproxy will use the value of the gravity option.
Trim
trim:%threshold:%color:%equal_hor:%equal_ver
t:%threshold:%color:%equal_hor:%equal_ver
Removes surrounding background.
threshold
- color similarity tolerance.color
- (optional) a hex-coded value of the color that needs to be cut off.equal_hor
- (optional) set to1
,t
ortrue
, imgproxy will cut only equal parts from left and right sides. That means that if 10px of background can be cut off from the left and 5px from the right, then 5px will be cut off from both sides. For example, this can be useful if objects on your images are centered but have non-symmetrical shadow.equal_ver
- (optional) acts likeequal_hor
but for top/bottom sides.
Trimming requires an image to be fully loaded into memory. This disables scale-on-load and significantly increases memory usage and processing time. Use it carefully with large images.
If you know background color of your images then setting it explicitly via color
will also save some resources because imgproxy won't need to automatically detect it.
Use a color
value of FF00FF
for trimming transparent backgrounds as imgproxy uses magenta as a transparency key.
The trimming of animated images is not supported.
Padding
padding:%top:%right:%bottom:%left
pd:%top:%right:%bottom:%left
Defines padding size using CSS-style syntax. All arguments are optional but at least one dimension must be set. Padded space is filled according to the background option.
top
- top padding (and for all other sides if they haven't been explicitly set)right
- right padding (and left if it hasn't been explicitly set)bottom
- bottom paddingleft
- left padding
Padding is applied after all image transformations (except watermarking) and enlarges the generated image. This means that if your resize dimensions were 100x200px and you applied the padding:10
option, then you will end up with an image with dimensions of 120x220px.
Padding follows the dpr option so it will also be scaled if you've set it.
Auto rotate
auto_rotate:%auto_rotate
ar:%auto_rotate
When set to 1
, t
or true
, imgproxy will automatically rotate images based on the EXIF Orientation parameter (if available in the image meta data). The orientation tag will be removed from the image in all cases. Normally this is controlled by the IMGPROXY_AUTO_ROTATE configuration but this processing option allows the configuration to be set for each request.
Rotate
rotate:%angle
rot:%angle
Rotates the image on the specified angle. The orientation from the image metadata is applied before the rotation unless autorotation is disabled.
Only 0, 90, 180, 270, etc., degree angles are supported.
Default: 0
Background
background:%R:%G:%B
bg:%R:%G:%B
background:%hex_color
bg:%hex_color
When set, imgproxy will fill the resulting image background with the specified color. R
, G
, and B
are the red, green and blue channel values of the background color (0-255). hex_color
is a hex-coded value of the color. Useful when you convert an image with alpha-channel to JPEG.
With no arguments provided, disables any background manipulations.
Default: disabled
Background alpha pro
background_alpha:%alpha
bga:%alpha
Adds an alpha channel to background
. The value of alpha
is a positive floating point number between 0
and 1
.
Default: 1
Adjust pro
adjust:%brightness:%contrast:%saturation
a:%brightness:%contrast:%saturation
This is a meta-option that defines the brightness, contrast, and saturation. All arguments are optional and can be omitted to use their default values.
Brightness pro
brightness:%brightness
br:%brightness
When set, imgproxy will adjust brightness of the resulting image. brightness
is an integer number ranging from -255
to 255
.
Default: 0
Contrast pro
contrast:%contrast
co:%contrast
When set, imgproxy will adjust the contrast of the resulting image. contrast
is a positive floating point number, where a value of 1
leaves the contrast unchanged.
Default: 1
Saturation pro
saturation:%saturation
sa:%saturation
When set, imgproxy will adjust saturation of the resulting image. saturation
is a positive floating-point number, where a value of 1
leaves the saturation unchanged.
Default: 1
Monochrome pro
monochrome:%intencity:%color
mc:%intencity:%color
When intencity
is greater than zero, imgproxy will convert the resulting image to monochrome.
intencity
- a positive floating-point number between0
and1
that defines the intencity of the monochrome effect.color
- (optional) a hex-coded value of the color that will be used as a base for the monochrome palette.
Default: 0:b3b3b3
Duotone pro
duotone:%intencity:%color1:%color2
dt:%intencity:%color1:%color2
When intencity
is greater than zero, imgproxy will convert the resulting image to duotone.
intencity
- a positive floating-point number between0
and1
that defines the intencity of the duotone effect.color1
,color2
- (optional) hex-coded values of the colors that will be used as a base for the duotone palette.color1
is the color for the dark areas,color2
is the color for the light areas.
Default: 0:000000:ffffff
Blur
blur:%sigma
bl:%sigma
When set, imgproxy will apply a gaussian blur filter to the resulting image. The value of sigma
defines the size of the mask imgproxy will use.
Default: disabled
Sharpen
sharpen:%sigma
sh:%sigma
When set, imgproxy will apply the sharpen filter to the resulting image. The value of sigma
defines the size of the mask imgproxy will use.
As an approximate guideline, use 0.5 sigma for 4 pixels/mm (display resolution), 1.0 for 12 pixels/mm and 1.5 for 16 pixels/mm (300 dpi == 12 pixels/mm).
Default: disabled
Pixelate
pixelate:%size
pix:%size
When set, imgproxy will apply the pixelate filter to the resulting image. The value of size
defines individual pixel size.
Default: disabled
Unsharp masking pro
unsharp_masking:%mode:%weight:%divider
ush:%mode:%weight:%divider
Allows redefining unsharp masking options. All arguments have the same meaning as Unsharp masking configs. All arguments are optional and can be omitted.
Blur detections pro
blur_detections:%sigma:%class_name1:%class_name2:...:%class_nameN
bd:%sigma:%class_name1:%class_name2:...:%class_nameN
imgproxy detects objects of the provided classes and blurs them. If class names are omitted, imgproxy blurs all the detected objects.
The value of sigma
defines the size of the mask imgproxy will use.
Draw detections pro
draw_detections:%draw:%class_name1:%class_name2:...:%class_nameN
dd:%draw:%class_name1:%class_name2:...:%class_nameN
When draw
is set to 1
, t
or true
, imgproxy detects objects of the provided classes and draws their bounding boxes, classes, and confidences. If class names are omitted, imgproxy draws the bounding boxes of all the detected objects.
Colorize pro
colorize:%opacity:%color:%keep_alpha
col:%opacity:%color:%keep_alpha
Places a color overlay on the processed image.
opacity
: specifies the overlay opacity. When set to0
, overlay is not applied.color
: (optional) a hex-coded value of the overlay color. Default:000
(black).keep_alpha
: (optional) when set to1
,t
ortrue
, imgproxy will keep the alpha channel of the original image. Default:false
Gradient pro
gradient:%opacity:%color:%direction:%start:%stop
gr:%opacity:%color:%direction:%start:%stop
Places a gradient on the processed image. The placed gradient transitions from transparency to the specified color.
opacity
: specifies gradient opacity. When set to0
, gradient is not applied.color
: (optional) a hex-coded value of the gradient color. Default:000
(black).direction
: (optional) specifies the direction of the gradient. The direction can be specified in two ways:- An angle in degrees (clockwise). For example,
0
will create a gradient from top to down (the top side is transparrent, the bottom side is opaque), and90
will create a gradient from right to left (the right side is transparrent, the left side is opaque). - A string value. Available values:
down
: (default) the top side of the gradient is transparrent, the bottom side is opaqueup
: the bottom side of the gradient is transparrent, the top side is opaqueright
: the left side of the gradient is transparrent, the right side is opaqueleft
: the right side of the gradient is transparrent, the left side is opaque
- An angle in degrees (clockwise). For example,
start
,stop
: floating point numbers that define relative positions of where the gradient starts and where it ends. Default values are0.0
and1.0
respectively.
Watermark
watermark:%opacity:%position:%x_offset:%y_offset:%scale
wm:%opacity:%position:%x_offset:%y_offset:%scale
Places a watermark on the processed image.
opacity
: watermark opacity modifier. Final opacity is calculated likebase_opacity * opacity
.position
: (optional) specifies the position of the watermark. Available values:ce
: (default) centerno
: north (top edge)so
: south (bottom edge)ea
: east (right edge)we
: west (left edge)noea
: north-east (top-right corner)nowe
: north-west (top-left corner)soea
: south-east (bottom-right corner)sowe
: south-west (bottom-left corner)re
: repeat and tile the watermark to fill the entire imagech
: pro same asre
but watermarks are placed in a chessboard order
x_offset
,y_offset
- (optional) specify watermark offset by X and Y axes:- When
x_offset
ory_offset
is greater than or equal to1
or less than or equal to-1
, imgproxy treats it as an absolute value. - When
x_offset
ory_offset
is less than1
and greater than-1
, imgproxy treats it as a relative value. - When using
re
orch
position, these values define the spacing between the tiles.
- When
scale
: (optional) a floating-point number that defines the watermark size relative to the resultant image size. When set to0
or when omitted, the watermark size won't be changed.
Default: disabled
Watermark URL pro
watermark_url:%url
wmu:%url
When set, imgproxy will use the image from the specified URL as a watermark. url
is the URL-safe Base64-encoded URL of the custom watermark.
Default: blank
Watermark text pro
watermark_text:%text
wmt:%text
When set, imgproxy will generate an image from the provided text and use it as a watermark. text
is the URL-safe Base64-encoded text of the custom watermark.
By default, the text color is black and the font is sans 16
. You can use Pango markup in the text
value to change the style.
If you want to use a custom font, you need to put it in /usr/share/fonts
inside a container.
Default: blank
Watermark size pro
watermark_size:%width:%height
wms:%width:%height
Defines the desired width and height of the watermark. imgproxy always uses fit
resizing type when resizing watermarks and enlarges them when needed.
When %width
is set to 0
, imgproxy will calculate the width using the defined height and watermark's aspect ratio.
When %height
is set to 0
, imgproxy will calculate the height using the defined width and watermark's aspect ratio.
This processing option takes effect only when the scale
argument of the watermark
option is set to zero.
Default: 0:0
Watermark rotate pro
watermark_rotate:%angle
wm_rot:%angle
wmr:%angle
Rotates the watermark on the specified angle (clockwise). The orientation from the image metadata is applied before the rotation.
Default: 0
Watermark shadow pro
watermark_shadow:%sigma
wmsh:%sigma
When set, imgproxy will add a shadow to the watermark. The value of sigma
defines the size of the mask imgproxy will use to blur the shadow.
Default: disabled
Style pro
style:%style
st:%style
When set, imgproxy will prepend a <style>
node with the provided content to the <svg>
node of a source SVG image. %style
is URL-safe Base64-encoded CSS-styles.
Default: blank
Strip metadata
strip_metadata:%strip_metadata
sm:%strip_metadata
When set to 1
, t
or true
, imgproxy will strip the metadata (EXIF, IPTC, etc.) on JPEG and WebP output images. This is normally controlled by the IMGPROXY_STRIP_METADATA configuration but this processing option allows the configuration to be set for each request.
Keep copyright
keep_copyright:%keep_copyright
kcr:%keep_copyright
When set to 1
, t
or true
, imgproxy will not remove copyright info while stripping metadata. This is normally controlled by the IMGPROXY_KEEP_COPYRIGHT configuration but this processing option allows the configuration to be set for each request.
DPI pro
dpi:%dpi
When set, imgproxy will replace the image's DPI metadata with the provided value. When set to 0
, imgproxy won't change the image's DPI or will reset it to the default value if the image's metadata should be stripped.
This processing option takes effect whether imgproxy should strip the image's metadata or not.
Default: 0
Strip color profile
strip_color_profile:%strip_color_profile
scp:%strip_color_profile
When set to 1
, t
or true
, imgproxy will transform the embedded color profile (ICC) to sRGB and remove it from the image. Otherwise, imgproxy will try to keep it as is. This is normally controlled by the IMGPROXY_STRIP_COLOR_PROFILE configuration but this processing option allows the configuration to be set for each request.
Enforce thumbnail
enforce_thumbnail:%enforce_thumbnail
eth:%enforce_thumbnail
When set to 1
, t
or true
and the source image has an embedded thumbnail, imgproxy will always use the embedded thumbnail instead of the main image. Currently, only thumbnails embedded in heic
and avif
are supported. This is normally controlled by the IMGPROXY_ENFORCE_THUMBNAIL configuration but this processing option allows the configuration to be set for each request.
Quality
quality:%quality
q:%quality
Redefines quality of the resulting image, as a percentage. When set to 0
, quality is assumed based on IMGPROXY_QUALITY and format_quality.
Default: 0.
Format quality
format_quality:%format1:%quality1:%format2:%quality2:...:%formatN:%qualityN
fq:%format1:%quality1:%format2:%quality2:...:%formatN:%qualityN
Adds or redefines IMGPROXY_FORMAT_QUALITY values.
Autoquality pro
autoquality:%method:%target:%min_quality:%max_quality:%allowed_error
aq:%method:%target:%min_quality:%max_quality:%allowed_error
Redefines autoquality settings. All arguments have the same meaning as Autoquality configs. All arguments are optional and can be omitted.
Autoquality requires the image to be saved several times. Use it only when you prefer the resulting size and quality over the speed.
Max bytes
max_bytes:%bytes
mb:%bytes
When set, imgproxy automatically degrades the quality of the image until the image size is under the specified amount of bytes.
Applicable only to jpg
, webp
, heic
, and tiff
.
When max_bytes
is set, imgproxy saves image multiple times to achieve the specified image size.
Default: 0
JPEG options pro
jpeg_options:%progressive:%no_subsample:%trellis_quant:%overshoot_deringing:%optimize_scans:%quant_table
jpgo:%progressive:%no_subsample:%trellis_quant:%overshoot_deringing:%optimize_scans:%quant_table
Allows redefining JPEG saving options. All arguments have the same meaning as the Advanced JPEG compression configs. All arguments are optional and can be omitted.
PNG options pro
png_options:%interlaced:%quantize:%quantization_colors
pngo:%interlaced:%quantize:%quantization_colors
Allows redefining PNG saving options. All arguments have the same meaning as with the Advanced PNG compression configs. All arguments are optional and can be omitted.
WebP options pro
webp_options:%compression:%smart_subsample
webpo:%compression:%smart_subsample
Allows redefining WebP saving options. All arguments have the same meaning as with the Advanced WebP compression configs. All arguments are optional and can be omitted.
Format
format:%extension
f:%extension
ext:%extension
Specifies the resulting image format. Alias for the extension part of the URL.
Default: jpg
Page pro
page:%page
pg:%page
When a source image supports pagination (PDF, TIFF) or animation (GIF, WebP), this option allows specifying the page to use. Page numeration starts from zero.
If both the source and the resulting image formats supoprt animation, imgproxy will ignore this option and use all the source image pages. Use the disable_animation option to make imgproxy treat all images as not animated.
Default: 0
Pages pro
pages:%pages
pgs:%pages
When a source image supports pagination (PDF, TIFF) or animation (GIF, WebP), this option allows specifying the number of pages to use. The pages will be stacked vertically and left-aligned.
If both the source and the resulting image formats supoprt animation, imgproxy will ignore this option and use all the source image pages. Use the disable_animation option to make imgproxy treat all images as not animated.
Default: 1
Disable animation pro
disable_animation:%disable
da:%disable
When set to 1
, t
or true
, imgproxy will treat all images as not animated. Use the page and the pages options to specify which frames imgproxy should use.
Default: false
Video thumbnail second pro
video_thumbnail_second:%second
vts:%second
Allows redefining IMGPROXY_VIDEO_THUMBNAIL_SECOND config.
Video thumbnail keyframes pro
video_thumbnail_keyframes:%keyframes
vtk:%keyframes
Allows redefining IMGPROXY_VIDEO_THUMBNAIL_KEYFRAMES config.
Video thumbnail tile pro
video_thumbnail_tile:%step:%columns:%rows:%tile_width:%tile_height:%extend_tile:%trim:%fill:%focus_x:%focus_y
vtt:%step:%columns:%rows:%tile_width:%tile_height:%extend_tile:%trim:%fill:%focus_x:%focus_y
When step
is not 0
, imgproxy will generate a tiled sprite using the source video frames.
step
: the step of timestamp (in seconds) between video frames that should be used for the sprite generation:- When
step
value is positive, imgproxy will use it as an absolute value - When
step
value is negative, imgproxy will calculate the actual step asvideo_duration / (columns * rows)
- When
columns
: the number of columns in the spriterows
: the number of rows in the spritetile_width
,tile_height
: the width and height of each tile in the sprite. imgproxy will resize each used frame to fit the provided sizeextend_tile
: (optional) when set to1
,t
ortrue
, imgproxy will extend each tile to the requested size using a black backgroundtrim
: (optional) when set to1
,t
ortrue
, imgproxy will trim the unused space from the spritefill
: (optional) when set to1
,t
ortrue
, imgproxy will use thefill
resizing type for the tilesfocus_x
,focus_y
: (optional) floating point numbers between 0 and 1 that define the coordinates of the center of the resulting tile (as in thefp
gravity type). Treat 0 and 1 as right/left forx
and top/bottom fory
. Applicable only whenfill
is set. Default:0.5:0.5
The timestamp of the first source video frame can be set using the IMGPROXY_VIDEO_THUMBNAIL_SECOND config or the video_thumbnail_second option.
You can make imgproxy use only keyframes with the IMGPROXY_VIDEO_THUMBNAIL_KEYFRAMES config or the video_thumbnail_keyframes option. Also, IMGPROXY_VIDEO_THUMBNAIL_TILE_AUTO_KEYFRAMES config makes imgproxy automatically use keyframes when the step
value is greater than the interframe interval.
If the step
value is less than the interframe interval, some frames may be used for more than one tile.
Default: 0:0:0:0:0
Video thumbnail animation pro
video_thumbnail_animation:%step:%delay:%frames:%frame_width:%frame_height:%extend_frame:%trim:%fill:%focus_x:%focus_y
vta:%step:%delay:%frames:%frame_width:%frame_height:%extend_frame:%trim:%fill:%focus_x:%focus_y
When step
is not 0
, imgproxy will generate an animated image using the source video frames.
step
: the step of timestamp (in seconds) between video frames that should be used for the animation generation:- When
step
value is positive, imgproxy will use it as an absolute value - When
step
value is negative, imgproxy will calculate the actual step asvideo_duration / frames
- When
delay
: the delay between animation frames in millisecondsframes
: the number of animation framesframe_width
,frame_height
: the width and height of animation frames. imgproxy will resize each used frame to fit the provided sizeextend_frame
: (optional) when set to1
,t
ortrue
, imgproxy will extend each animation frame to the requested size using a black backgroundtrim
: (optional) when set to1
,t
ortrue
, imgproxy will trim the unused frames from the animationfill
: (optional) when set to1
,t
ortrue
, imgproxy will use thefill
resizing type for the animation framesfocus_x
,focus_y
: (optional) floating point numbers between 0 and 1 that define the coordinates of the center of the resulting animation frame (as in thefp
gravity type). Treat 0 and 1 as right/left forx
and top/bottom fory
. Applicable only whenfill
is set. Default:0.5:0.5
The timestamp of the first video source frame can be set using the IMGPROXY_VIDEO_THUMBNAIL_SECOND config or the video_thumbnail_second option.
You can make imgproxy use only keyframes with the IMGPROXY_VIDEO_THUMBNAIL_KEYFRAMES config or the video_thumbnail_keyframes option. Also, IMGPROXY_VIDEO_THUMBNAIL_TILE_AUTO_KEYFRAMES config makes imgproxy automatically use keyframes when the step
value is greater than the interframe interval.
If the step
value is less than the interframe interval, some video frames may be used for more than one animation frame.
Default: 0:0:0:0:0
Fallback image URL pro
You can use a custom fallback image by specifying its URL with the fallback_image_url
processing option:
fallback_image_url:%url
fiu:%url
The value of url
is the URL-safe Base64-encoded URL of the custom fallback image.
Default: blank
Skip processing
skip_processing:%extension1:%extension2:...:%extensionN
skp:%extension1:%extension2:...:%extensionN
When set, imgproxy will skip the processing of the listed formats. Also available as the IMGPROXY_SKIP_PROCESSING_FORMATS configuration.
Processing can only be skipped when the requested format is the same as the source format.
Video thumbnail processing can't be skipped.
Default: empty
Raw
raw:%raw
When set to 1
, t
or true
, imgproxy will respond with a raw unprocessed, and unchecked source image. There are some differences between raw
and skip_processing
options:
- While the
skip_processing
option has some conditions to skip the processing, theraw
option allows to skip processing no matter what - With the
raw
option set, imgproxy doesn't check the source image's type, resolution, and file size. Basically, theraw
option allows streaming of any file type - With the
raw
option set, imgproxy won't download the whole image to the memory. Instead, it will stream the source image directly to the response lowering memory usage - The requests with the
raw
option set are not limited by the IMGPROXY_WORKERS config
Default: false
Cache buster
cachebuster:%string
cb:%string
Cache buster doesn't affect image processing but its changing allows for bypassing the CDN, proxy server and browser cache. Useful when you have changed some things that are not reflected in the URL, like image quality settings, presets, or watermark data.
It's highly recommended to prefer the cachebuster
option over a URL query string because that option can be properly signed.
Default: empty
Expires
expires:%timestamp
exp:%timestamp
When set, imgproxy will check the provided unix timestamp and return 404 when expired.
Default: empty
Filename
filename:%filename:%encoded
fn:%filename:%encoded
Defines a filename for the Content-Disposition
header. When not specified, imgproxy will get the filename from the source URL.
filename
: escaped or URL-safe Base64-encoded filename to be used in theContent-Disposition
headerencoded
: (optional) identifies iffilename
is Base64-encoded. Set it to1
,t
, ortrue
if you encoded thefilename
value with URL-safe Base64 encoding.
Default: empty
Return attachment
return_attachment:%return_attachment
att:%return_attachment
When set to 1
, t
or true
, imgproxy will return attachment
in the Content-Disposition
header, and the browser will open a 'Save as' dialog. This is normally controlled by the IMGPROXY_RETURN_ATTACHMENT configuration but this processing option allows the configuration to be set for each request.
Preset
preset:%preset_name1:%preset_name2:...:%preset_nameN
pr:%preset_name1:%preset_name2:...:%preset_nameN
Defines a list of presets to be used by imgproxy. Feel free to use as many presets in a single URL as you need.
Read more about presets in the Presets guide.
Default: empty
Hashsum pro
hashsum:%hashsum_type:%hashsum
hs:%hashsum_type:%hashsum
When hashsum_type
is not none
, imgproxy will calculate the hashsum of the source image and compare it with the provided hashsum. If they don't match, imgproxy will respond with 422.
hashsum_type
: specifies the hashsum type. Available values:none
: don't check the hashsummd5
: use MD5 hashsumsha1
: use SHA1 hashsumsha256
: use SHA256 hashsumsha512
: use SHA512 hashsum
hashsum
: the expected hex-encoded hashsum of the source image. Ifhashsum_type
isnone
, this argument is ignored and can be omitted
Default: none
Checking video file hashsums is not supported as it would require downloading the whole video file. imgproxy won't throw an error if you try to check the hashsum of a video file but it will skip the hashsum check
Max src resolution
max_src_resolution:%resolution
msr:%resolution
Allows redefining IMGPROXY_MAX_SRC_RESOLUTION config.
Since this option allows redefining a security restriction, its usage is not allowed unless the IMGPROXY_ALLOW_SECURITY_OPTIONS config is set to true
.
Max src file size
max_src_file_size:%size
msfs:%size
Allows redefining IMGPROXY_MAX_SRC_FILE_SIZE config.
Since this option allows redefining a security restriction, its usage is not allowed unless the IMGPROXY_ALLOW_SECURITY_OPTIONS config is set to true
.
Max animation frames
max_animation_frames:%size
maf:%size
Allows redefining IMGPROXY_MAX_ANIMATION_FRAMES config.
Since this option allows redefining a security restriction, its usage is not allowed unless the IMGPROXY_ALLOW_SECURITY_OPTIONS config is set to true
.
Max animation frame resolution
max_animation_frame_resolution:%size
mafr:%size
Allows redefining IMGPROXY_MAX_ANIMATION_FRAME_RESOLUTION config.
Since this option allows redefining a security restriction, its usage is not allowed unless the IMGPROXY_ALLOW_SECURITY_OPTIONS config is set to true
.
Source URL
Plain
The source URL can be provided as is, prepended by the /plain/
segment:
/plain/http://example.com/images/curiosity.jpg
If the source URL contains a query string or @
, you'll need to escape it.
When using a plain source URL, you can specify the extension after @
:
/plain/http://example.com/images/curiosity.jpg@png
Base64 encoded
The source URL can be encoded with URL-safe Base64. The encoded URL can be split with /
as desired:
/aHR0cDovL2V4YW1w/bGUuY29tL2ltYWdl/cy9jdXJpb3NpdHku/anBn
When using an encoded source URL, you can specify the extension after .
:
/aHR0cDovL2V4YW1w/bGUuY29tL2ltYWdl/cy9jdXJpb3NpdHku/anBn.png
Encrypted with AES-CBC pro
The source URL can be encrypted with the AES-CBC algorithm, prepended by the /enc/
segment. The encrypted URL can be split with /
as desired:
/enc/jwV3wUD9r4VBIzgv/ang3Hbh0vPpcm5cc/VO5rHxzonpvZjppG/2VhDnX2aariBYegH/jlhw_-dqjXDMm4af/ZDU6y5sBog
When using an encrypted source URL, you can specify the extension after .
:
/enc/jwV3wUD9r4VBIzgv/ang3Hbh0vPpcm5cc/VO5rHxzonpvZjppG/2VhDnX2aariBYegH/jlhw_-dqjXDMm4af/ZDU6y5sBog.png
Extension
Extension specifies the format of the resulting image. Read more about image formats support here.
The extension can be omitted. In this case, imgproxy will use the source image format as resulting one. If the source image format is not supported as the resulting image, imgproxy will use jpg
. You also can enable AVIF, WebP, or JPEG XL support detection to use it as the default resulting format when possible.
Best format pro
You can use the best
value for the format option or the extension to make imgproxy pick the best format for the resultant image. Check out the Best format guide to learn more.
Example
A signed imgproxy URL that uses the sharp
preset, resizes http://example.com/images/curiosity.jpg
to fill a 300x400
area using smart gravity without enlarging, and then converts the image to png
:
http://imgproxy.example.com/AfrOrF3gWeDA6VOlDG4TzxMv39O7MXnF4CXpKUwGqRM/preset:sharp/resize:fill:300:400:0/gravity:sm/plain/http://example.com/images/curiosity.jpg@png
The same URL with shortcuts will look like this:
http://imgproxy.example.com/AfrOrF3gWeDA6VOlDG4TzxMv39O7MXnF4CXpKUwGqRM/pr:sharp/rs:fill:300:400:0/g:sm/plain/http://example.com/images/curiosity.jpg@png
The same URL with a Base64-encoded source URL will look like this:
http://imgproxy.example.com/AfrOrF3gWeDA6VOlDG4TzxMv39O7MXnF4CXpKUwGqRM/pr:sharp/rs:fill:300:400:0/g:sm/aHR0cDovL2V4YW1w/bGUuY29tL2ltYWdl/cy9jdXJpb3NpdHku/anBn.png