music-emotion

Audio emotion analysis identifies the primary mood, energy level, valence, arousal, and detected genre of a track. It extracts structured metadata fields like primary_emotion, overall_energy, and mood_tags for music tagging and playlist curation.

Audio mood detection produces structured fields including primary_emotion, secondary_emotions, overall_energy, energy_curve, valence, arousal, genre, and mood_tags. These outputs enrich audio libraries and enable mood-based recommendations across streaming workflows.

Genre and mood profiling classifies the detected genre alongside the dominant mood and energy trajectory across tracks. This generates the structured metadata required for organizing large audio libraries and curating mood-based playlists.

Valence and arousal detection maps the emotional content of audio files using a CLAP-based model or heuristic features. It outputs structured valence and arousal scores alongside primary_emotion and overall_energy for archival metadata enrichment.

Audio mood tagging and energy curve extraction processes standard audio files to generate mood_tags and energy_curve metadata. It applies heuristic features or CLAP-based models directly to the audio file without requiring external dependencies.

Automated emotion tagging is ideal for enriching archival metadata, generating mood-based recommendations, and scaling playlist curation workflows. It classifies primary_emotion and overall_energy across large audio libraries without manual listening.