The Semantic Layer
The annotation layer that turns the corpus from files on disk into something that can be trained on with precision and searched through with intent.
Three Modes of Search
Precision, audio similarity, and narrative — three ways to ask the corpus for music, each suited to a different way of knowing what you want.
What CORPUS Knows About Each Track
The dimensions the annotation pipeline captures — tempo, key, instrumentation, mood, vocal treatment, energy, narrative function, and originality.