dezoomify-rs is a tiled image downloader. Some webpages present high-resolution zoomable images without a way to download them. These images are often tiled: the original large image has been split into smaller individual image files called tiles. The only way to download such an image is to download all the tiles separately and then stitch them together. This process can be automated by a tiled image downloader.
dezoomify-rs is a tiled image downloader.
Some webpages present high-resolution zoomable images without a way to download them.
These images are often tiled: the original large image has been split into smaller individual image files called tiles.
The only way to download such an image is to download all the tiles separately and then stitch them together.
This process can be automated by a tiled image downloader.
The most common tiled image downloader is probably dezoomify,
an online tool which is very easy to use.
The goal of this project is not to replace the traditional dezoomify.
However, some images are so large that they can't be efficiently downloaded and displayed inside a web browser.
Other times, a website tries to protect its tiles by refusing access to them when certain
HTTP headers are not set to the right values.
dezoomify-rs is a desktop application for Windows, MacOs and linux that does not have the same limitations as the online zoomify.
dezoomify-rs also lets the user choose between
several image formats,
whereas in dezoomify, you can only save the image as PNG.
dezoomify-rs supports several zoomable image formats, each backed by a dedicated dezoomer.
The following dezoomers are currently available:
generic For when the tile URLs follow a simple pattern.
custom for advanced users.
It allows you to specify a custom tile URL format that can contain multiple variables. This gives you the most flexibity, but requires some manual work.
download the version that matches your operating system (Windows, MacOS, or Linux),
Extract the binary from the compressed file.
On some operating systems, you may have to authorize the application execution
before being able to launch it. See how to do
in MacOS.
Install via Homebrew
As an alternative to installing the binary directly, on macOS and Linux dezoomify-rs is available via the Homebrew package manager. It can be installed with the command brew install dezoomify-rs.
Supported output image formats
Dezoomify-rs supports multiple output image formats.
The format to use is determined by the name of the output file.
For instance, entering dezoomify-rs http://example.com/ my_image.png on the command line
will create a PNG image.
Each image format encoder has a distinct set of features and limitations :
PNG images are compressed losslessly, which means that the output image quality
is (very slightly) better than JPEG, at the expense of much larger file sizes.
The PNG encoder in dezoomify-rs can create very large images;
it is not limited by the available memory on your computer.
This format is chosen by default when the image is very large,
or its size is not known in advance.
JPEG is the most common image format.
JPEG images cannot be more than 65,535 pixels wide or high.
This format is chosen be default for images that fit within this limit.
The JPEG encoder in dezoomify-rs requires the whole image to fit in memory on your computer.
IIIF, which allows you to re-create a zoomable image locally.
This is the recommended output format when your image is very large
(multiple hundreds of megapixels), since most image viewers do not accept huge PNGs or JPEGs.
If the output path ends with .iiif, a folder will be created instead of a single file,
with its structure following the IIIF specification.
A file called viewer.html will be created inside this folder,
which you can open in your browser to view the image.
Tile cache
By default, dezoomify-rs works entirely in memory, which is very fast.
However, the latest versions added the possibility to use a "tile cache".
When you launch dezoomify-rs from the commandline with dezoomify-rs --tile-cache my_caching_folder http://myurl.com,
it will save all the image tiles it downloads to the specified folder.
If the download is interrupted before the end, you will be able to resume it later by specifying the same tile cache folder.
A tile cache also allows you to manually get the individual tiles if you want to stitch them manually.
You have to give dezoomify-rs an url to the ImageProperties.xml file.
You can use dezoomify-extension to
find the URL of this file.
Alternatively, you can find it out manually by opening your network inspector.
If the image tile URLs have the form
http://example.com/path/to/TileGroup1/1-2-3.jpg,
then the URL to enter is
http://example.com/path/to/ImageProperties.xml.
IIIF
The IIIF dezoomer takes the URL of an
info.json file as input.
Alternatively, you can find this url in your browser's network inspector when loading the image.
IIIF Manifest Support
dezoomify-rs also supports processing IIIF Presentation API manifests directly, which is particularly useful for downloading entire manuscripts or multi-page documents. When processing manifests, dezoomify-rs extracts metadata to generate meaningful filenames that include document titles and page/section labels rather than generic numbered files.
DeepZoom
The DeepZoom dezoomer takes the URL of a dzi file as input, which you can find using
dezoomify-extension.
You can find this url in your browser's network inspector when loading the image.
If the image tile URLs have the form
http://test.com/y/xy_files/1/2_3.jpg,
then the URL to enter is
http://test.com/y/xy.dzi.
Zoomify PFF
PFF
is an old zoomable image file format format developed by zoomify.
You can give a pff meta-information URL (one that contains requestType=1)
to dezoomify-rs and it will download it.
Krpano
Krpano is a zoomable image format often used
for panoramas, virtual tours, photoshperes, and other 3d zoomable images.
dezoomify-rs supports downloading individual image planes from such images.
You need to provide the xml meta-information file for the image.
Nypl
The digital collections of New York's Public Library
use their own zoomable image format, which dezoomify-rs supports.
Some images have a high-resolution version available, and work with this software.
Others do not, and can be downloaded by simply right-clicking on them in your browser.
To download an image, just enter the URL of its viewer page in dezoomify-rs, like for example:
IIPImage is an image web server that implements
the Internet Imaging Protocol.
Such images are easily recognizable by their tile URLs, which contain FIF=.
You can pass an URL containing FIF= to dezoomify-rs to let it download the image.
Generic
You can use this dezoomer if you know the format of tile URLs.
For instance, if you noticed that the URL of the first tile is
http://example.com/my_image/image-0-0.jpg
and the second is
http://example.com/my_image/image-1-0.jpg
then you can guess what the general format will be, and give dezoomify-rs
the following:
http://example.com/my_image/image-{{X}}-{{Y}}.jpg
If the numbers have leading zeroes in the URL
(such as image-01-00.jpg instead of image-1-0.jpg),
then you can specify them in the url template as follows:
The custom yaml dezoomer
is a powerful tool that lets you download tiled images in many different formats, including formats that are not explicitly
supported by dezoomify-rs.
In order to use this dezoomer, you'll need to create a tiles.yaml file, which is a little bit technical.
However, we have a a tutorial for the custom YAML dezoomer
to help you.
If you are having troubles understanding the tutorial or adapting it to your use-case, you should get in touch by
opening a new github issue.
Command-line options
When using dezoomify-rs from the command-line
Allows downloading zoomable images. Supports several different formats such as zoomify, iiif, and deep zoom images.
Usage: dezoomify-rs [OPTIONS] [INPUT_URI] [OUTFILE]
Arguments:
[INPUT_URI] Input URL or local file name. By default, the program will ask for it interactively
[OUTFILE] File to which the resulting image should be saved. By default the program will generate a name based on the image metadata if available. Otherwise, it will generate a name in the format "dezoomified[_N].{jpg,png}" depending on which files already exist in the current directory, and whether the target image size fits in a JPEG or not
Options:
-?, --help
Displays this help message
-d, --dezoomer
Name of the dezoomer to use. "auto" will try to detect the format automatically [default: auto]
-l, --largest
If several zoom levels are available, select the largest one (highest resolution)
-w, --max-width
If several zoom levels are available, select the one with the largest width that does not exceed this value (in pixels)
-h, --max-height
If several zoom levels are available, select the one with the largest height that does not exceed this value (in pixels)
--zoom-level
Select a specific zoom level by its index (0-based). Use 0 for the smallest, 1 for the next level up, etc. If the specified level doesn't exist, falls back to the highest available level
--image-index
Select a specific image by its index (0-based) when multiple images are found. If not specified, the program will ask interactively when multiple images are available. If the specified index doesn't exist, falls back to the last one
-n, --parallelism
Degree of parallelism to use. At most this number of tiles will be downloaded at the same time [default: 16]
-r, --retries
Number of new attempts to make when a tile load fails before giving up. Setting this to 0 is useful to speed up the generic dezoomer, which relies on failed tile loads to detect the dimensions of the image. On the contrary, if a server is not reliable, set this value to a higher number [default: 1]
--retry-delay
Amount of time to wait before retrying a request that failed. Applies only to the first retry. Subsequent retries follow an exponential backoff strategy: each one is twice as long as the previous one [default: 2s]
--compression
A number between 0 and 100 expressing how much to compress the output image. For lossy output formats such as jpeg, this affects the quality of the resulting image. 0 means less compression, 100 means more compression. Currently affects only the JPEG and PNG encoders [default: 5]
-H, --header
Sets an HTTP header to use on requests. This option can be repeated in order to set multiple headers. You can use `-H "Referer: URL"` where URL is the URL of the website's viewer page in order to let the site think you come from the legitimate viewer
--max-idle-per-host
Maximum number of idle connections per host allowed at the same time [default: 32]
--accept-invalid-certs
Whether to accept connecting to insecure HTTPS servers
-i, --min-interval
Minimum amount of time to wait between two consequent requests. This throttles the flow of image tile requests coming from your computer, reducing the risk of crashing the remote server of getting banned for making too many requests in a short succession [default: 50ms]
--timeout
Maximum time between the beginning of a request and the end of a response before the request should be interrupted and considered failed [default: 30s]
--connect-timeout
Time after which we should give up when trying to connect to a server [default: 6s]
--logging
Level of logging verbosity. Set it to "debug" to get all logging messages [default: info]
-c, --tile-cache
A place to store the image tiles when after they are downloaded and decrypted. By default, tiles are not stored to disk (which is faster), but using a tile cache allows retrying partially failed downloads, or stitching the tiles with an external program
--bulk
URL or path to a text file containing a list of URLs to process in bulk mode. Each line in the file should contain one URL, optionally followed by a custom title. Format: URL [custom title] Lines starting with # are treated as comments and ignored. Accepts both local file paths and HTTP(S) URLs. Can also directly process IIIF manifests to download all images with enhanced metadata-based filenames. In bulk mode, if no level-specifying argument is defined (such as --max-width), then --largest is implied
-V, --version
Print version
Multi-Image Selection
Many sources contain multiple zoomable images rather than just one. Dezoomify-rs can handle these cases intelligently:
Automatic Detection
IIIF Manifests: When you provide a manifest URL, dezoomify-rs will extract all images and let you choose which one to download
Krpano Scenes: Multi-scene Krpano files (like panoramic tours) are processed as separate images
Bulk Processing: Text files with multiple URLs are processed sequentially
Interactive Selection
When multiple images are found, dezoomify-rs will show you a list with titles and descriptions:
Found 3 images:
[0] Front cover (2000x3000 pixels)
[1] f. 1r - Gospel of Matthew begins (4000x6000 pixels)
[2] Back cover (2000x3000 pixels)
Enter the image number to download (0-2):
Non-Interactive Selection
For automated workflows, use the --image-index option:
# Download the second image (0-based indexing)
dezoomify-rs --image-index 1 https://example.com/iiif/manifest.json
# In bulk mode, the first image is selected automatically
dezoomify-rs --bulk urls.txt
Examples
# Interactive selection from an IIIF manifest
dezoomify-rs https://library.example.edu/iiif/manuscript/manifest.json
# Select specific image non-interactively
dezoomify-rs --image-index 2 https://library.example.edu/iiif/manuscript/manifest.json output.jpg
# Bulk process with automatic first-image selection
dezoomify-rs --bulk manuscript-urls.txt
Documentation
For documentation specific to this tool, see the dezoomify-rs wiki. Do not hesitate to contribute to it by creating new pages or modifying existing ones.
For general purpose documentation about zoomable images, the dezoomify wiki may be useful.
Bulk mode
dezoomify-rs supports bulk processing of multiple URLs using the --bulk option. This allows you to process multiple zoomable images in a single command. The bulk source can be either a local file path or a URL.
Using bulk mode
Option 1: Text file with URLs
Create a text file containing one URL per line, optionally followed by a custom title:
# urls.txt - Lines starting with # are comments and will be ignored
https://example.com/image1/ImageProperties.xml My First Image
https://example.com/image2/info.json Custom Title for Second Image
https://example.com/image3.dzi
# You can also include local file paths
/path/to/local/tiles.yaml Local Manuscript
The format for each line is: URL [custom title]
The URL is required and must be valid
The custom title is optional - if not provided, a title will be generated from the URL
Everything after the first space following the URL is treated as the title
Empty lines and lines starting with # are ignored as comments
Then run dezoomify-rs with the --bulk option:
./dezoomify-rs --bulk urls.txt
Option 2: Direct IIIF manifest processing
You can also pass a URL directly to the --bulk option to process IIIF manifests:
This is particularly useful for downloading entire manuscripts or collections from IIIF-compatible repositories. The tool will automatically extract all images from the manifest and generate meaningful filenames using metadata from the manifest.
Enhanced filename generation
When processing IIIF manifests, dezoomify-rs now creates much more descriptive filenames by leveraging metadata:
Metadata titles: Uses the "Title" field from manifest metadata (e.g., "Gospel-book ('Lindisfarne Gospels')")
Canvas labels: Incorporates specific page/section labels (e.g., "Front cover", "f. 1r", "Inside back cover")
Smart fallbacks: Falls back to manifest labels or generic page numbers when metadata isn't available
Progress tracking: Each URL is processed sequentially with progress indicators ([1/5], [2/5], etc.)
Automatic level selection: If no level-specifying arguments (--max-width, --max-height, --zoom-level) are provided, --largest is automatically implied
Output file naming: If you specify an output file with --outfile, each image will be saved with a suffix (_0001, _0002, etc.)
Error handling: Failed downloads don't stop the entire process; the tool continues with the next URL and reports a summary at the end
Metadata preservation: IIIF manifest metadata is extracted and used for intelligent filename generation
Examples
Process multiple images and save them with a common prefix:
./dezoomify-rs --bulk urls.txt my_collection.jpg
# Creates: my_collection_0001.jpg, my_collection_0002.jpg, etc.
Process with specific size constraints:
./dezoomify-rs --bulk urls.txt --max-width 2000
# Uses max-width constraint instead of auto-selecting largest
Process an IIIF manifest directly:
./dezoomify-rs --bulk https://library.example.edu/iiif/manuscript123/manifest.json
# Downloads all pages with meaningful names based on manifest metadata
Process a bulk file from a URL:
./dezoomify-rs --bulk https://example.com/collection-urls.txt
# Downloads and processes the URL list from the remote file
Alternative: Using shell scripts
For more advanced processing, you can still use traditional shell scripting approaches with xargs:
xargs -d '\n' -n 1 ./dezoomify-rs < ./urls.txt
Technical Architecture
Dezoomify-rs works by first analyzing the structure of a zoomable image to understand its tiling scheme,
then systematically downloads all individual tiles and reassembles them into the complete high-resolution image.
We can automatically detect which type of zoomable image format is being used (such as Zoomify, IIIF, or DeepZoom)
and apply the appropriate strategy to locate and download tiles.
During the download process, dezoomify-rs can preserve important image metadata:
ICC color profiles from individual tiles are transferred
to the final output image to maintain accurate color representation,
and EXIF metadata is preserved when saving to PNG format (though it's lost with other formats due to encoder limitations).
The PNG encoder uses a streaming approach that writes image data progressively to disk, allowing it to handle extremely large images without being limited by available system memory. In contrast, other format encoders like JPEG must keep the entire assembled image in memory before writing, which can be a constraint for very large images.
