A popular archive hub says it has published a Spotify backup as bulk torrents totaling 300TB or roughly 86 million music files, grouped by popularity. If you’re expecting a clean, complete mirror of Spotify’s catalog that you can download in one shot, that’s not what’s live today.
What is live is the catalog data: SQLite databases that the group says contain the largest publicly available music metadata database, covering 256 million tracks and 186 million unique ISRCs.
Anna’s Archive says it usually focuses on text because it’s dense, but its mission is preserving knowledge and culture across media. It also claims it found a way to scrape Spotify at scale and sees this as a start at building a preservation-focused music archive. If this archive is too much, perhaps Spotify’s own offline feature is for you.
What the database release includes
In its write-up, the group argues music is already fairly well preserved, but points to three gaps: a long tail that only gets saved when someone cares enough (and torrents can be poorly seeded), an audiophile tilt toward huge lossless files that makes “everything” hard to keep, and the lack of an authoritative torrent list meant to represent all recorded music.
Its Spotify metadata dump is positioned as the fix. It claims metadata coverage for about 99.9% of artists, albums, and tracks, with the core artist, album, and track dataset under 200GB compressed, plus a separate audio analysis dataset listed at 4TB compressed.
Audio comes in batches
Audio is the part many readers will care about most, and it’s the part still rolling out. Anna’s Archive says it has archived around 86 million music files, representing around 99.6% of listens, but those music files are slated to release in popularity order, not as a single drop.
It also calls out quality choices. For popularity greater than 0, it says it pulled original OGG Vorbis at 160kbit/s without reencoding. For popularity equal to 0, it says it reencoded to OGG Opus at 75kbit/s, and notes a ReplayGain tag bug affecting many files.
What to watch next
The group lists a cutoff of 2025-07, meaning releases after July 2025 may not be present. It also lays out the next steps: music files, then extra file metadata (paths and checksums), then album art and patch files meant to reconstruct originals. The practical takeaway is that this backup is metadata only for now, with audio coming in later.
