Calypso Music AI is an experimental system that classifies and groups music based on musical attributes rather than user behavior. It analyzes rhythm, harmony, timbre, and structure to create meaningful clusters and recommendations. The project explores how genre, mood, and arrangement can be inferred from audio features, enabling discovery that feels musical instead of statistical. The pipeline emphasizes clean preprocessing, robust feature extraction, and interpretable outputs. It is designed as a research friendly tool for exploring audio similarity and building playlists that sound coherent and intentional. It balances experimentation and rigor, making results easier to trust and apply later confidently.