PlexMix
github.com
- Linux
- Web
- Self-hosted
What PlexMix is
PlexMix is free, open-source software for creating personalized playlists from a Plex music library with natural-language descriptions. It synchronizes Plex metadata into a local SQLite database and stores semantic embeddings for search. Playlist generation can use filters for genre, year, rating, artist, environment, instrument, tempo, energy, key, and danceability. Ordering choices include similarity, random, alternating-artist, and energy-curve modes, with saved templates available. Optional Essentia analysis extracts tempo, key, energy, danceability, and loudness, while MusicBrainz enrichment can add community-curated genres, artist identifiers, and recording-type information. Playlists can be imported from files with track matching or exported as JSON and M3U. PlexMix offers a command-line interface and an alpha web UI; its README recommends the CLI for production. It can be installed from PyPI or run with Docker Compose on Linux and self-hosted setups. Cloud AI providers need an API key and internet access, while local models support offline generation. The project is licensed under GNU AGPL v3 and currently supports one music library at a time.
Who it suits
Plex users and music lovers who want prompt-based playlist creation and local control may find it suitable. It also fits self-hosters comfortable with a CLI, Docker Compose, or an alpha web interface.
Good to know, and what to watch
- Natural-language playlist prompts
- Many playlist filters and ordering modes
- JSON and M3U import and export
- Local models can run offline
- Open-source GNU AGPL v3 license
- Supports one music library currently
- Web interface is alpha
- Cloud providers require keys and internet
- README recommends CLI for production
The verdict
PlexMix offers detailed, prompt-driven playlist generation around Plex libraries, with optional local analysis and broad export options. Its one-library limit and alpha web interface make the command line the clearer production path.
What the maker publishes
- AI providers
- It supports Google Gemini, OpenAI GPT, Anthropic Claude, Cohere Command, local Gemma and Mistral models, and OpenAI-compatible local endpoints. source
- AI providers
- Supported providers include Google Gemini, OpenAI, Anthropic, Cohere, local models, and OpenAI-compatible endpoints. source
- Audio analysis
- Optional Essentia-based analysis extracts features such as tempo, key, energy, danceability, and loudness. source
- Audio analysis
- Optional Essentia analysis extracts features including tempo, key, energy, danceability, loudness, and spectral measures. source
- Cost note
- The README estimates default Gemini playlist creation at ~$0.01 per playlist and local embeddings as completely free. source
- Credential storage
- Credentials can be stored in the system keyring or passed as environment variables. source
- Deployment
- It can be installed from PyPI or run with Docker Compose, with prebuilt amd64 and arm64 images described in the README. source
- Deployment
- The project can be installed from PyPI or source and deployed with Docker Compose; it publishes amd64 and arm64 Docker images. source
- Export and import
- Playlists can be exported as JSON or M3U and imported from files with track matching. source
- Interfaces
- PlexMix provides a command-line interface and an alpha web UI; the README recommends the CLI for production use. source
- Library limit
- The README says PlexMix currently supports one music library at a time; multi-library support is on its roadmap. source
- License
- The project is licensed under the GNU Affero General Public License v3.0. source
- License and audience
- The project is licensed under GNU AGPL v3 and describes itself as made for music lovers. source
- Local library data
- It syncs Plex music metadata to a local SQLite database and generates semantic embeddings for tracks. source
- Local storage
- It syncs Plex music metadata to a local SQLite database and stores semantic embeddings for search. source
- Music metadata
- Optional MusicBrainz enrichment provides community-curated genres, canonical artist IDs, and recording-type detection. source
- Music metadata
- Optional MusicBrainz enrichment provides community-curated genres, canonical artist IDs, and recording type detection. source
- Network requirements
- Cloud AI providers require an API key and internet access, while the README documents local models for offline playlist and tag generation. source
- Ordering
- It offers similarity, random, alternating-artist, and energy-curve track ordering modes. source
- Playlist controls
- Playlist generation supports filters including genre, year, rating, artist, environment, instrument, tempo, energy, key, and danceability. source
- Playlist tools
- Playlist generation supports filters, four ordering modes, saved templates, and JSON or M3U import and export. source
- Plex access
- PlexMix reads Plex metadata and can create Plex playlists when enabled; it says it does not modify or delete tracks, albums, or artists. source
- Privacy option
- Local embedding models run on the user's machine, and the README says their embedding calls run offline and privately. source
- Purpose
- PlexMix creates personalized music playlists from a Plex library using natural-language mood descriptions. source
Checked against the maker's pages on 9 Oct 2026.