Dig into your stacks.
Type "late night psychedelic soul" and get a playlist. Filter by female artists from the 70s. Discover you've scrobbled Stereolab 247 times but don't own a single album. Runs on your NAS. Your data stays yours.
Your files already know the truth. MusicMind just asks them.
Synapse now identifies vocal vs. instrumental straight from the audio — no guessing, no tags to trust.
Resolves "Various Artists" tracks to their real performer, and finds true duplicates regardless of filename or bitrate.
No more pip, no more permissions wrangling. One container, one command.
Plex has the data. MusicMind surfaces it intelligently.
Type any mood or vibe. "Quiet and watery." "Those cats are jumpin'." AI expands your prompt into specific genre tags and scores your library against them.
Your entire library gets tagged with specific subgenres โ shoegaze, dance-punk, balearic, cosmic disco โ and written back into Plex. Genre browsing finally makes sense. Tags are grounded in real artist data, not free-association.
Your scrobbles drive real play counts, listening patterns, and loved track ratings. Plex's play counts are unreliable. Last.fm's aren't.
"Your Afternoon" โ what you actually play 1-5pm. "Weekend Flow." "Often Together" โ tracks that appear in the same listening session, cross-artist.
Artists you've scrobbled 50+ times that aren't in your library. Categorized into Worth Acquiring, Classical, Ambient/Meditation, and Unknown.
AI classifies every track as instrumental or vocal. Filter playlists to instrumentals only, or auto-detect from your prompt.
Filter by gender, country, era, decade, genre, year range, play count, or rating. "Female artists from Brazil in the 70s" is a valid query.
Your library and listening history at a glance. Scrobbles by year, top artists, genre breakdown, library by era and country, gender split.
Full sync pipeline in one button: scan Plex for new files โ ingest โ sync Last.fm โ AI tag new tracks. Deleted files get reconciled out automatically. Runs on your NAS, streams live output.
MusicMind listens to your actual files. BPM, key, scale, and danceability extracted locally with Essentia โ no API, no cost, no cloud. Your library's sonic fingerprint, computed on your NAS.
Artist country and gender come from MusicBrainz, the open music encyclopedia โ real data, not guesses. AI fills in only what MusicBrainz can't match.
Full documentation and a support & feedback channel โ all inside the app. No hunting through GitHub wikis.
Dry-run the enrichment pipelines and verify your data with one-click integrity checks. It caught two real data bugs on its very first run.
Four steps from setup to your first AI-generated playlist.
Python app runs alongside Plex. No cloud required. Your music never leaves your network.
Pulls all tracks from Plex into SQLite. AI tags every track with specific genre and mood tags via OpenAI.
Sync your scrobble history. Real play counts, listening patterns, loved tracks โ all imported.
Type a mood, set filters, hit Preview. Click Create in Plex โ your playlist appears in Plexamp instantly.
A look inside MusicMind for Plex.
Natural language playlist generator with filters
The Collection โ your library at a glance
Missing From Your Collection โ gap analysis
Back Office โ one-click sync pipeline
Synapse โ local audio analysis with live progress
Built-in User Guide โ docs where you need them
Test Suite โ pipeline dry runs and data integrity checks
MusicMind for Plex is free to download and self-host. The only additional cost is your OpenAI API usage for AI tagging โ see OpenAI pricing for current rates.
free. No subscription. Your data stays on your hardware.