ListenBrainz and MusicBrainz: The Wikipedia of Music Metadata
Why I moved from Last.fm to ListenBrainz, what the MetaBrainz databases offer, and how my old site showed what I was listening to.
Reviewed by adrian
on
My old website had a small widget on the home page: a pulsing dot showed what I was listening to, or what I last played when I was offline. A consumption page showed a grid of albums from songs I had liked. The data came from ListenBrainz, my replacement for Last.fm, which I used for years. Behind it is a whole set of open databases.
MusicBrainz
MusicBrainz is an open, collaborative music database, run by the non-profit MetaBrainz Foundation. Every artist, album, recording and release gets a unique, permanent ID, the MBID (MusicBrainz Identifier).
The comparison with Wikipedia fits: a worldwide community maintains the data by clear rules, with an open API. Today it is one of the most complete music databases there is, with millions of artists, albums and tracks, links to lyrics, relations between bands, remixes and release dates.
The MetaBrainz projects
- ListenBrainz is a free, open-source successor to Last.fm scrobbling: a player plugin reports each track, and ListenBrainz keeps it in your profile, permanently and exportable. My whole Last.fm history imported without trouble. There is no subscription for statistics, and the API is public, including other people's public profiles. On top come recommendations, year-in-review summaries and playlist export.
- Cover Art Archive, run with the Internet Archive, stores album covers linked to the release MBIDs. With an MBID you get the cover in several sizes from a single URL.
- AcoustID identifies a music file from its sound alone: it computes an acoustic fingerprint and matches it against a database linked to MusicBrainz IDs.
- AcousticBrainz collected acoustic properties (tempo, key, mood) for millions of tracks. It stopped in 2022; the data is still available.
- BookBrainz does the same for books: authors, publishers, editions and translations, with permanent IDs and an open API.
The software around it
- MusicBrainz Picard is the official tag editor. It fingerprints local files, finds the matching release and writes clean tags back, MBIDs and covers included.
- Beets manages music libraries on the command line and uses MusicBrainz to identify and sort imported files. Plugins extend it.
- Scrobbler clients exist for nearly every player: Spotify, VLC, MPD, Plex, Kodi, Jellyfin and more.
How my old site used it
Now playing. The widget polled the ListenBrainz endpoint /user/altner/playing-now every 60 seconds, cached. ListenBrainz sends no stop signal when a track ends. A track that had been "playing now" for more than 15 minutes counted as a missing stop, and the widget fell back to the last entry of the listening history:
const PLAYING_NOW_MAX_MS = 15 * 60 * 1000;
if (
cache.playingNow.data &&
playingSince &&
Date.now() - playingSince > PLAYING_NOW_MAX_MS
) {
return null; // No longer "now playing", stale signal
}
Album grid. The /feedback API with score=1 returns every track I marked with a heart. The field mbid_mapping gives the releases. For each release the site fetched the year from the MusicBrainz API and the cover from the Cover Art Archive. Both were cached on the server: covers as files with a one-hour TTL, metadata in a JSON file. The result was a horizontal row of albums, each linked to its MusicBrainz page.
Why I like it
- Open data instead of silos. Music metadata is shared culture. A non-profit keeps it free under open licences with a stable API. Spotify, Apple Music and the others sit on large metadata stores and give nothing back.
- Stable IDs. An MBID does not change, and no streaming service can delete or rename it. The same URL will return the same data in ten years.
- Quality from people. The entries are kept by people who care about music and correct data. In many areas that beats generated metadata.
- No lock-in. ListenBrainz exports all listening data, as JSON or JSPF playlists. If you want to leave, you can.
If you have only ever scrobbled to Last.fm, ListenBrainz is worth a look. And if your music folder is a mess, run Picard over it once.
Based on
-
ListenBrainz & MusicBrainz: The Wikipedia of Music Metadata
(
raw/2026/2026-06-14/listenbrainz-musicbrainz-the-wikipedia-of-music-metadata.md) [post]
Topic
- Open Data Data that anyone may use, share and build on, maintained in open databases with stable identifiers and public interfaces.