Date created: | Last Updated:

Category: Data

Description: MagnetDB is a large-scale, open dataset that catalogs extensive BitTorrent metadata. It contains nearly 30 million torrent entries in a torrents table, along with over 950 million file-level entries in a files table. For each torrent, MagnetDB tracks core attributes such as its name, size, and discovery timestamps. For each file, it stores detailed metadata like file paths, sizes, video/audio codecs, matching confidence scores for video content to the non-commercial IMDb dataset, and more. By providing a unified schema across billions of rows, MagnetDB offers researchers, data scientists, and developers a powerful resource to study the supply side of the BitTorrent ecosystem, analyze file distribution patterns, or build content aggregation tools. It is distributed in a single SQLite database (over 100 GB uncompressed), along with sample extracts for quick exploration.