Summary AudioMuse AI automates intelligent playlist generation for Jellyfin and Navidrome users by performing deep sonic analysis on music libraries. It leverages AI and machine learning models to understand track characteristics like tempo, mood, and genre, creating dynamic playlists based on user preferences or automated discovery methods.

How It Works The system analyzes audio tracks using Librosa and TensorFlow's MusiCNN models to extract features and generate 200-dimensional embeddings. These features or embeddings are then clustered using algorithms like K-Means, DBSCAN, or GMM, optimized via an evolutionary Monte Carlo approach. AI models (Ollama, Gemini) can provide descriptive playlist names. The generated playlists are directly created within Jellyfin or Navidrome via their respective APIs. Similarity searches are powered by the Voyager Approximate Nearest Neighbors index.

Quick Start & Requirements

• Installation: Deploy via Docker Compose for local setups or using the provided Helm chart for Kubernetes (K3S tested).
• Prerequisites: A running Jellyfin or Navidrome server is essential. Requires Docker/Docker Compose or Kubernetes/Helm.
• Hardware: Recommended: 4-core CPU (Intel/ARM, 2015+, AVX support), 8GB RAM, SSD. Older CPUs without AVX are incompatible.
• Optional: NVIDIA GPU with CUDA 12.9+ for accelerated analysis. AI models require Gemini API key or Ollama instance.
• Docs: Helm Chart available at https://github.com/NeptuneHub/AudioMuse-AI-helm.

Highlighted Details

• AI-Powered Features: Generates playlists via natural language chat, finds similar songs using sonic fingerprints, and creates "song paths" between tracks.
• Intelligent Naming: Utilizes Ollama or Gemini for creative playlist naming.
• Advanced Clustering: Supports multiple algorithms (K-Means, DBSCAN, GMM, Spectral) with an evolutionary tuning process for optimal playlist configurations.
• Collection Sync: Experimental feature to sync analysis data to a centralized cloud database.
• Architecture Support: Compatible with amd64 and arm64 platforms.

Maintenance & Community AudioMuse AI is an open-source, community-driven project in BETA. Contributions and issue reporting are encouraged. Specific community channels (like Discord/Slack) or a public roadmap are not detailed in the provided README.

Licensing & Compatibility The specific open-source license is not explicitly stated in the README. Compatibility is confirmed for amd64 and arm64 architectures and integrates with Jellyfin, Navidrome, and other Subsonic API-compatible media servers.

Limitations & Caveats This project is in BETA and intended for testing, not production environments; users proceed at their own risk. Full library analysis can take several days. Older CPUs lacking AVX support are incompatible. Some features, like Collection Sync and Nvidia Worker support, are experimental.