Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


Server Components for the AcousticBrainz project

GitHub last commit GitHub pull requests

WebsiteDocumentationBug tracker


The AcousticBrainz project aims to crowd source acoustic information for all music in the world and to make it available to the public. This acoustic information describes the acoustic characteristics of music and includes low-level spectral information and information for genres, moods, keys, scales and much more. The goal of AcousticBrainz is to provide music technology researchers and open source hackers with a massive database of information about music.

AcousticBrainz is a joint effort between Music Technology Group at Universitat Pompeu Fabra in Barcelona and the MusicBrainz project.

AcousticBrainz organizes the data on a recording basis, indexed by the MusicBrainz ID for recordings. If you know the MBID for a recording, you can easily fetch from AcousticBrainz. For details on how to do this, visit our API documentation.

Changes and other important announcements about the AcousticBrainz project will be announced on our blog.


If you are interested in helping out, consider donating to the MetaBrainz Foundation.

If you are interested in contributing code or documentation, please have a look at the issue tracker or come visit us in the #metabrainz IRC channel on

Installation and Running


We use docker and docker-compose to run the AcousticBrainz server. Ensure that you have these tools installed, following the installation instructions.


Copy the following two configuration files:

  1. to
  2. to in the ./hl_extractor/ directory In you need to set the models_essentia_git_sha value. Unless you know what you are doing, this value should be v2.1_beta1

Running docker-compose commands

For convenience, we provide a script which calls docker-compose. We also have some additional subcommands for commonly used commands. Some of these subcommands take no arguments:

./ bash   # open a bash shell in a new container in the webserver service
./ psql   # run psql, connecting to the database
./ shell  # run a flask shell in ipython

And some subcommands take arguments, passing them to the underlying program:

./ npm    # run npm in a new container in the webserver service
./ manage # run python in a new container in the webserver service
./ ...    # run docker-compose

If you want to run docker-compose yourself you are welcome to do so, however keep in mind that we call it in the following way, to standardise the project name:

docker-compose -f docker/ -p acousticbrainz-server <args>

Build and initial configuration

Build the docker containers needed for AcousticBrainz by running the following:

./ build


Start the webserver and other required services with:

./ up

The first time you install AcousticBrainz, you will need to initialize the AcousticBrainz database. Run in a separate terminal:

./ manage init_db

You will be able to view your local AcousticBrainz server at http://localhost:8080

Development notes


In order to load a psql session, use the following command:

./ psql

Building static files

We use webpack as our JavaScript/CSS build system.

First-time npm setup

For development, the first time that you install acousticbrainz you must install node packages in your local directory.

./ npm install

This has the effect of creating a node_modules directory in your local code checkout.

To build stylesheets and javascript bundles, run webpack:

./ npm run build:dev

You will need to rebuild static files after you modify JavaScript or CSS. If you want to rebuild these source files as you change them then you can run webpack in watch mode:

./ npm run build:watch


To use the dataset tools you need to configure OAuth with MusicBrainz. Log in to your MusicBrainz account (or create one if needed) and create a new application.

Choose a name (for example, "AcousticBrainz development"), set Type to "Web Application" and set the Callback URL to http://localhost:8080/login/musicbrainz/post

Copy the OAuth Client ID and OAuth Client Secret values to as MUSICBRAINZ_CLIENT_ID and MUSICBRAINZ_CLIENT_SECRET.

You should now be able to use the menu in the top corner of your AcousticBrainz server to log in.

Admin interface

Once you have logged in, you can make your user an admin, by running

./ manage add_admin <your user>

You should now be able to access the admin section at http://localhost:8080/admin

Working with data


Before you import or export data, make sure you understand how docker bind mounts work. The following commands will work if you specify paths in the current directory, but if you want to specify paths somewhere else (e.g. a Downloads or tmp directory) you must specify an additional --mount flag.

AcousticBrainz provides data dumps that you can import into your own server. Latest database dump is available at You need to download full database dump from this page and use it during database initialization:

./ run --rm webserver python2 init_db path_to_the_archive

you can also easily remove existing database before initialization using --force option:

./ run --rm webserver python2 init_db --force path_to_the_archive

or import archive after database is created:

./ run --rm webserver python2 import_data path_to_the_archive

You can also import dumps that you created yourself. This process is described below (see dump full_db command).


There are several ways to export data out of AcousticBrainz server. You can create full database dump or export only low-level and high-level data in JSON format. Both ways support incremental dumping.


Full database dump:

./ manage dump full_db

JSON dump:

./ manage dump json

Creates two separate full JSON dumps with low-level and high-level data.

Incremental dumps:

./ manage dump incremental

Creates new incremental dump in three different formats: usual database dump, low-level and high-level JSON.

Previous incremental dumps:

./ manage dump incremental --id 42

Same as another one, but recreates previously created incremental dump.

Test your changes with unit tests

Unit tests are an important part of AcousticBrainz. It helps make it easier for developers to test changes and help prevent easily avoidable mistakes later on. Before commiting new code or making a pull request, run the unit tests on your code.


This will start a set of docker containers separate from your development environment, run the tests, and then stop and remove the containers. To run tests more rapidly without having to bring up and take down containers all the time, you can run each step individually. To bring up containers in the background:

./ -u

Then run your tests when you need with:

./ [optional arguments to pass to py.test]

Stop the test containers with:

./ -s

This will stop but not delete the containers. You can delete the containers with:

./ -d

We use the -p flag to docker-compose to start the test containers as a new project, acousticbrainztest so that containers don't conflict with already running development containers. You can access containers directly while they are running (e.g. with docker exec) with this name (e.g. acousticbrainztest_db_1)

The database has no separate volume for data, this means that any data in the test database will disappear when the containers are deleted (at the end of standalone ./, or after ./ -d)

We forward the port from postgres to localhost:15431, so you can connect to it with psql on your host if you want to inspect the contents of the database.