Transkr V4 treats audio as musical structure rather than simply a stream of sound. It analyzes notes, polyphonic combinations, melody, chords, and meter, with real-time processing for live performance as well as analysis of recorded audio. The analysis works directly from the instrument’s audio signal, requiring no specialized pickup or hardware, and runs locally without cloud processing or reliance on large training datasets. Its model customization system allows users to adapt transcription models to different instruments and playing styles. The combination of polyphonic transcription, musical structure analysis, real-time processing, and user-adaptable models makes Transkr a flexible software-only approach to extracting musical information from audio.