Skip to content


Repository files navigation

Unreal Archive

Scans, categorises and produces metadata for Unreal, Unreal Tournament, Unreal Tournament 2003/4, and Unreal Tournament 3 content, and builds a static browsable website of the content, currently published at


  • Java JRE 17 for running
  • Java JDK 17 for building

Tested with both OpenJDK 17 on Linux and Azul's Zulu Java 17 on Windows.


The project is build with Gradle. The provided gradlew wrapper may be invoked as follows to produce an executable file:

On Linux

./gradlew jlink

To run, execute:


On Windows

gradlew.bat jlink

To run, execute:


Components and Modules

The project is arranged into several modules, and published as Java Modules so components may be reused by other software projects.

These are:

  • common: the usual assortment of shared boilerplate code and utility classes and functions.
  • content: the definition of the content types, Maps, Map Packs, Skins, Game Types, etc. These are the elements which are serialiased to and from YAML format stored in the archive data repository, and projects wishing to work with the Unreal Archive data types should re-use these if possible. In addition, the code for loading, saving, and generally working with the data sets are here too.
  • storage: the remote storage/mirror implementations, for S3 and Azure and associated interfaces.
  • www: the code and templates responsible for generating the static website output.
  • root project: the content indexing and mirroring implementation and management. This will likely move into its own module(s) later.

Usage and Functionality

TODO: complete this section

Run with no arguments to see input arguments and additional help for each command.

Browsing and Information:

  • ls: List indexed content filtered by game, type or author.
  • filter: List indexed content filtered by attribute=value pairs.
  • show: Show data for the content items specified.
  • summary: Show stats and counters for the content index.

General Content Management

  • scan: Dry-run scan the contents of files or paths, comparing to known content where possible.
  • index: Index the contents of files or paths, writing the results to the content path.
  • edit: Edit the metadata for the provided.
  • set: Convenience, set an attribute for the provided. Eg: set <hash> author Bob.
  • sync: Sync managed files' local files to remote storage.

Gametype Management

  • All commands prefixed by gametype:
  • init <game> <gametype name>: Create a skeleton gametype file structure.
  • locate <game> <gametype name>: Show the local file path to the provided gametype.
  • index <game> <game type name> <release name>: Indexes the content of the release specified.
  • add <game> <game type name> <release name> <file>: Convenience, which adds a gametype if it does not yet exist, adds a release, and indexes the release. A sync command afterwards is still required to sync download files to mirrors.
  • addmirror <game> <game type name> <release name> <url>: Adds a secondary mirror to the gametype specified.


  • local-mirror: Create a local mirror of all file content.
  • mirror: Create a remote mirror of all file content and add mirror links.

Utilities and Tools:

  • unpack: Unpack the contents of a umod file to a directory.
  • install: Unpack the contents of a file, URL or hash, and place within an Unreal game's standard directory layout (Maps, System, Textures etc).

Website Build:

  • www: Generate the HTML website for browsing content.

Content Submission and Indexing Pipeline

  • For any given file, a Submission instance is created.
    • If the file's directory contains an _override.yml file, that file is loaded as a SubmissionOverride instance and associated with the Submission.
  • The Submission is used to instantiate an Incoming instance, which contains additional information like the file's SHA1 hash.
  • If the content is new (hash is not yet known), the file is unpacked onto disk (via system-installed archiving tools 7z and unrar).
  • The Incoming instance is passed to ContentType.classify() for content type classification, where the type of content is determined (see Content Identification).
  • If the content type can be identified, the content type will be used as a factory to instantiate a new Content element of the appropriate type with common content information and metadata.
  • The content type will again be used as a factory to instantiate an IndexHandler instance, which will take the new Content instance and populate it with more type-specific metadata extracted from the actual content being indexed (eg. extract author and screenshot information from map packages, or determine whether or not a skin has team colours defined).
  • Once indexed, the result is returned via a Consumer<>, and this is then "checked into" the ContentManager, which will write the content metadata to disk in YAML format, as well as sync the package and any additional files to the nominated data store.

Most of this process uses Consumer<>s for providing feedback, with the intention of eventually being able to make this process more parallelised for better performance. Currently the performance is "good enough" that such parallel behaviour is not quite worth the implementation time.


If you have storage capacity available and would like to contribute some of it as a public mirror for archive content, the following steps should be taken.

  1. Fork and clone the unreal-archive-data repository.
    • This dataset will be updated during the mirroring process.
  2. Download or build the unreal-archive project binary
  3. Execute unreal-archive mirror with the following command-line options:
    • --content-path=/path/to/unreal-archive-data
    • --store=[dav|s3|az] and appropriate configuration and credentials.
    • --concurrency=3 with an appropriate concurrency value for your bandwidth and processing power (3 is default)
    • --since=yyyy-mm-dd [optional] - only mirror content added after the date specified
  4. Wait while the mirror process completes. If you want to abort, just Ctrl+C the process and whatever content has been mirrored so far can be used as-is or resumed later.
  5. Ensure that the URLs added to the data files are publicly accessible by doing a spot-check on some of them.
  6. Once the mirror is complete (or partially complete if you're only doing a partial mirror) and some URLs have been eyeballed, git commit your changes to the unreal-archive-data repository and push to your fork, then use GitHub to open a Pull Request to the main repository master branch.

Storage Configuration

A storage backend must be configured in order to store content during indexing.

When interrogating, mirroring or downloading content, no store needs to be specified, the default no-op store (--store=nop) will be used.

Environment variables may be used for configuration, eg.: replace --s3-key-images=key with S3_KEY_IMAGES=key.


This is mostly for testing purposes, but in some circumstances may be useful.

  • --store=dav
    • --store-[images|attachments|content]=dav
  • --dav-url=http://hostname/path/
    • --dav-url-[images|attachments|content]=...

S3 Bucket Storage

Supports S3-compatible storage implementations.

  • --store=s3
    • --store-[images|attachments|content]=bs3
  • --s3-key=key-id
    • --s3-key-[images|attachments|content]=key-id
  • --s3-secret=secret
    • --s3-secret-[images|attachments|content]=secret
  • --s3-bucket=bucket-id
    • --s3-bucket-[images|attachments|content]=bucket-id
  • --s3-endpoint=
    • (provide the root of your storage API)
    • --s3-endpoint-[images|attachments|content]=
  • --s3-url=
    • (provide the public URL of your storage bucket in the appropriate region. __BUCKET__ and __NAME__ will be replaced by the bucket and uploaded filenames respectively)

Note: Amazon S3 bucket policy to allow public downloads:

    "Version": "2012-10-17",
    "Statement": [
            "Sid": "PublicRead",
            "Effect": "Allow",
            "Principal": "*",
            "Action": "s3:GetObject",
            "Resource": "arn:aws:s3:::testing-unreal-archive-mirror/*"

Azure Blob Storage

  • --store=az
    • --store-[images|attachments|content]=az
  • --az-acc=storage-account-name
    • --az-acc-[images|attachments|content]=storage-account-name
  • --az-container
    • --az-container-[images|attachments|content]=container-name
  • --az-sas
    • --az-sas-[images|attachments|content]="shared-access-signature"
  • --az-endpoint (Optional, default="")
    • --az-endpoint-[images|attachments|content]=endpoint-suffix

Note: The shared access signature (SAS) should be created with full permissions to the specified container. Use double quotes around the SAS when specifying it via the command line.