CrateMate
A Python CLI tool that cleans, tags, and organizes downloaded music files into a DJ-ready library.
Challenge
As a DJ focused on electronic music, I deal with files from a variety of sources, each with different naming conventions and metadata quality. Existing music taggers like beets and MusicBrainz Picard consistently fail with electronic music, mismatching remixes with originals, requiring manual confirmation for every file, and producing incorrect metadata. I needed a tool that could get me from messy downloads to a clean, tagged, DJ-ready library, fast and with no manual matching.
Challenge
As a DJ focused on electronic music, I deal with files from a variety of sources, each with different naming conventions and metadata quality. Existing music taggers like beets and MusicBrainz Picard consistently fail with electronic music, mismatching remixes with originals, requiring manual confirmation for every file, and producing incorrect metadata. I needed a tool that could get me from messy downloads to a clean, tagged, DJ-ready library, fast and with no manual matching.
Approach
CrateMate is built on a simple insight: the correct metadata is almost always already in the filename. An 11-step regex pipeline extracts artist, title, and mix information, with Google Gemini as a fallback for filenames too chaotic for pattern matching. Cover art and genre data are pulled from Spotify and Discogs. Beyond import, the tool includes spectral analysis for detecting fake high-bitrate files, FLAC to MP3 conversion, AI genre classification, duplicate detection, and a full undo system. The entire application is a single 3,200 line Python module. All API integrations are optional. Built with Python, numpy, Pillow, mediafile, and FFmpeg. Developed using Claude Code.
Approach
CrateMate is built on a simple insight: the correct metadata is almost always already in the filename. An 11-step regex pipeline extracts artist, title, and mix information, with Google Gemini as a fallback for filenames too chaotic for pattern matching. Cover art and genre data are pulled from Spotify and Discogs. Beyond import, the tool includes spectral analysis for detecting fake high-bitrate files, FLAC to MP3 conversion, AI genre classification, duplicate detection, and a full undo system. The entire application is a single 3,200 line Python module. All API integrations are optional. Built with Python, numpy, Pillow, mediafile, and FFmpeg. Developed using Claude Code.