Bencode Encoder / Decoder

Bencode is a simple serialisation format designed for BitTorrent by Bram Cohen. It encodes four types: byte strings (4:spam), integers (i42e), lists (l...e), and dictionaries (d...e). BitTorrent uses it for .torrent metainfo files and DHT messages because it is unambiguous (no encoding ambiguity like JSON has for numbers), compact, and trivial to parse in any language without a library.

Bencode encodes integers as ie (e.g., i42e), byte strings as : (e.g., 4:spam), lists as le, and dictionaries as de where keys must be byte strings in sorted order. Nesting is supported arbitrarily. There is no native boolean or float type — booleans are represented as integers (0/1) and floats are typically stored as strings.

A .torrent file is a Bencoded dictionary containing: the announce URL of the tracker, an info dictionary with the torrent name, piece length, SHA-1 hashes of all pieces, and either a single file length or a list of files with paths and lengths. Optional fields include announce-list (backup trackers), creation date, comment, and created-by. Magnet links encode a subset of this as a URI.

Yes. Bencode byte strings are raw binary sequences prefixed with their byte length. This is why .torrent files store SHA-1 piece hashes as raw 20-byte binary strings rather than hex. When converting Bencode to JSON, binary strings are typically represented as hex or Base64 since JSON has no native binary type.