“We discovered a way to scrape Spotify at scale,” writes the Anna’s Archive team in a blog post announcing they have posted a 300TB torrent of 86 million audio files—an estimated 99.6% of streams on the platform. The shadow library, typically focused on sharing books and papers, frames the scrape as the world‘s first fully open “preservation archive” for music, and includes extensive analysis of listening patterns, genre distributions, and song metadata.