Music archive manager
Java C++ Perl
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Locutus is a software that compares audio files with the MusicBrainz database,
updates the metadata in the files and stores the files in a structured
This project is still quite young, there may be bugs and I won't guarantee that
it won't erase any files. That said, I still haven't lost a single file due to
Locutus, so the risk is low.

Some of Locutus' features:
- Capable of handling extremely large music archives with low CPU and memory
- Clever metadata matching that's extensively configurable. As long as the
  metadata found in audio files are correct Locutus mismatch files very rarily.
  It is however possible to configure Locutus in a way that will cause
  mismatched files, use precaution!
- Locutus is split in 3 parts, the daemon, the database and a user interface.
  Each part can run on its own computer, or they can all run on the same
  computer. While the daemon compares files, updates metadata and store the
  files, you can manually match the files the daemon was unable to automatically
  match in a user interface (currently there's a web interface and a Java
- Albums retrieved from MusicBrainz are stored in a local cache, reducing amount
  of queries sent, and thus also reducing load generated on MusicBrainz.
  When manually matching files with tracks no requests are sent to MusicBrainz.
- Grouping files in assumed albums. To reduce amount of queries sent to
  MusicBrainz the files are placed in groups that are assumed to be albums. This
  makes Locutus look up the first file in a group, fetch the best matching
  albums for that track and compare the other files in the group with those
  albums. If some of the files have a poor match, those files will be looked up
  separately and albums returned from those will also be matched with the other
  files in the group.
- Automatically fetch tags from AudioScrobbler as "genre". Currently it only
  supports storing the most popular tag. Since it needs to look up each file
  this feature significantly slows down Locutus as it may only look up tags once
  per second at most. In an effort to speed this up, only files with no genre
  set are looked up.

This project is still in an early alpha stage. It's not ready for the public yet
and the installation prosess requires some knowledge with Linux.

Locutus currently supports PostgreSQL version 8.3 or later (not tested on
earlier versions).

1. Become the 'postgres' user:
     sudo su - postgres
2. Create a user for Locutus:
     createuser -S -D -R -P -E locutus
3. Create a database for Locutus
     createdb -O locutus locutus
4. Log out the 'postgres' user and goto the "sql" directory of Locutus where
   you issue the following command:
     psql -h localhost -f locutus.sql -U locutus locutus
5. You'll need to grant Locutus access to the database. Edit (most likely)
   "/etc/postgresql/8.3/main/pg_hba.conf" and add the following entry after
   "#TYPE    DATABASE    USER        CIDR-ADDRESS          METHOD":
     host    locutus     locutus          md5
   Change the IP if you're running the daemon/web interface on another computer.
6. By default PostgreSQL only listen on If you're accessing the
   database from somewhere else, you'll need to edit
     listen_addresses = '*'
7. Restart PostgreSQL:
     /etc/init.d/postgresql restart

The daemon requires GNU Common C++, TagLib (version 1.5 or later) and the
PostgreSQL C client library.
For Debian users:
  aptitude install libpq-dev libtag1-dev libcommoncpp2-dev build-essential

1. Create a directory for building Locutus and enter the directory:
     mkdir build
     cd build
2. Generate Makefile:
     cmake ..
3. Compile Locutus:

If all goes well, you should now have a "locutus" executable in the "build"
directory. Before you can run Locutus you'll need to set up a configuration file
for connecting to PostgreSQL. Copy or symlink the "locutus.conf" file from the
"conf" directory to the build directory and modify its settings if necessary.
You should now be able to run Locutus. If you execute Locutus with "-d" then
Locutus will run as a daemon untill you send a TERM signal to its process.

The first time you run Locutus it will populate the "setting" table with default
values. You'll need to set the "dry_run" setting in the database to "false" when
you want Locutus to actually sort your files.

Java interface:
Java 1.6 is needed for the Java UI.
More information will come when it is slightly more usable.

Web interface:
Perl, Apache2 and Template::Toolkit are required for the web interface.
This guide is written for Debian Linux, if you're using another distro you may
need to do things differently. I leave it as an exercise to those readers to
figure this out :)
For Debian users:
  aptitude install libapache2-mod-speedycgi libapache-dbi-perl \
  templatetoolkit-perl libdbd-pgsql libclass-dbi-pg-perl

1. Open "/etc/apache2/sites-enabled/000-default" in your favourite editor and
   add the following lines somewhere in the VirtualHost (changing the paths to
   where you got the "web" directory):
     Alias /locutus/ "/home/canidae/projects/locutus/web/"
     <Directory /home/canidae/projects/locutus/web/>
         Options ExecCGI Indexes FollowSymLinks MultiViews
         AllowOverride None
         Order allow,deny
         allow from all
         AddHandler cgi-script .pl
2. Restart Apache:
     /etc/init.d/apache2 restart
3. Set the connection parameters in the "locutus.conf" file residing in the
4. The interface should now be up at "http://localhost/locutus/" (or another URL
   if you installed the web interface on another computer). If it isn't, then
   check logs in "/var/log/apache2/" and "/var/log/postgresql/".

The project is MIT licensed, so you're pretty free to use it as you wish.
If you improve the project then we would greatly appreciate patches, though.

The web interface is currently GPLv3 licensed. It is deprecated and will
eventually be removed or replaced.

Vidar Wahlberg, <>
  Daemon, database, Java interface, web interface

Kristian Klette, <>
  Web interface
Per Øyvind Øygard, <>
  Web interface