According to Anna’s Archive, their project is a "repository for preserving music," which includes 86 million of the most popular tracks, covering about 99.6% of all listens on Spotify. Priority was given to compositions that Spotify classifies as popular. In the first phase, the team has already released a separate torrent with metadata containing information about approximately 256 million tracks and 186 million unique ISRC codes used to identify recordings in the music industry.
Spotify has confirmed the start of an investigation and described the methods currently being examined: a company representative stated that a third party collected public metadata and used illegal means to bypass DRM and gain access to some audio files. However, Spotify cautiously phrases its statements, not confirming the scale of the leak mentioned by Anna’s Archive, indicating that it concerns only "some" audio recordings.
The statistics that the pirate group attaches to its announcement are of particular interest. Spotify has a track popularity index — from 0 to 100 — which is formed based on the number of listens and their relevance. Anna’s Archive claims that this indicator helped them determine which tracks to upload first and in what quality.
It is reported that for tracks with a popularity above zero, almost everything was preserved in its original quality from Spotify (OGG Vorbis at about 160 kbps). At the same time, less popular recordings, which account for about half of all listens, were re-encoded to OGG Opus at around 75 kbps to reduce the overall size of the archive. The group also notes that the "long tail" of the catalog, where popularity is zero, was hardly affected, as such tracks account for only a negligible share of listens, among which there is much questionable content, including hard-to-filter AI-generated material.
Despite claims of "preserving cultural heritage," from a legal standpoint, this looks like a classic case of mass data extraction and distribution of protected content.
This behavior almost certainly violates the terms of service and copyright, which could lead to demands to remove the torrents and strict measures from rights holders. Additionally, such a large set of music and metadata could theoretically serve as a basis for creating alternative pirate streaming services or be used for training models, which has already sparked debates about AI and author consent.
