Documentation
Core Musica: Similarity engine
How Core Musica scores tracks across ten weighted dimensions to drive playlist discovery.
The similarity engine evaluates every pair of tracks across ten dimensions — Genre, Subgenre, Tone, Mood, Vocal Style, Era, Tempo, Instrumentation, Themes, and Production Style — and lets you adjust each dimension's weight to steer discovery exactly where you want it.
- Ten dimensions: Genre, Subgenre, Tone, Mood, Vocal, Era, Tempo, Instrumentation, Themes, Production Style.
- Each dimension has an independent weight slider from 0 (ignored) to full influence.
- Similarity scores are cached after first computation for fast playlist generation.
- Semantic embedding fallback improves quality for tracks with sparse metadata.