Appearance
Sources
Media aggregates roughly a dozen upstreams. Each contributes something different, and none is individually required.
| Source | Contributes |
|---|---|
| Plex | What you own, and when it was added |
| Tautulli | Richer play data from that Plex server |
| Last.fm | Scrobble history — the basis of discoveries and trending |
| Spotify | Playlists, and metadata |
| Deezer | Wishlist and metadata |
| Discogs | Release detail, pressings, credits |
| MusicBrainz | Canonical music metadata |
| AcousticBrainz | Audio characteristics, behind moods |
| iTunes | Cover art and metadata |
| Wikipedia | Artist biographies |
| Lyrics | Track lyrics |
What happens when a source disagrees
Sources are reconciled onto canonical records, and where they conflict the canonical layer picks. Sources are not equal — a release identifier from MusicBrainz outranks a guess from a filename.
Manual corrections get overwritten
If you correct something a source supplies, the next enrichment run will overwrite it. Corrections belong upstream, in the source, or in the matching rules — not in the reconciled record.
This is worth knowing before spending an evening fixing album titles by hand.
Enrichment costs upstream calls
Several sources are rate-limited or paid. Enrichment runs on a schedule and as explicit admin actions rather than on every page view, which is why a newly added album may take a while to acquire its cover and credits.