How it works
From first sketch to final tag, every Royalty.red track goes through a deliberate process designed to make music that does a specific job — not just music that sounds good.
Composition
Tracks are built — AI-assisted generation plus human composers — specifically targeting one of the four modes' audio characteristics. We do not write music generically and sort it afterward. A Move track is composed with interval structures and BPM ranges in mind from the first bar. A Sleep track is designed to slow gradually, with no sudden dynamic shifts that could pull a listener back to alertness.
Human creative pass
Every AI-assisted track goes through human editing, arrangement, and refinement before release. This matters for quality, and it matters legally: purely AI-generated audio has uncertain copyright status in many jurisdictions without meaningful human creative input. Our composers, arrangers, and editors shape the material until it meets our standards — and until the human contribution is clear enough to stand up to licensing scrutiny.
Mode tagging
Tracks are tested and tagged into Move, Focus, Unwind, or Sleep — including sub-category — based on measurable traits: tempo, dynamics, harmonic stability, structure, and energy curve. This is not a vibe-based sort. A track lands in Focus because it maintains a steady rhythmic pulse and avoids melodic hooks that compete for attention. A track lands in Sleep because its BPM decreases gradually and its dynamic range stays contained.
Listener feedback loop
Skip rates, completion rates, and replay patterns per mode feed back into what gets produced next. If Sleep listeners consistently skip a track at the twelve-minute mark, we look at what changed in the arrangement. If Focus listeners replay a specific track for deep-work sessions, we study its structure and apply those traits to future compositions. The catalog evolves based on how people actually use it, not just what sounds good in the studio.
How AI and humans share the work
Royalty.red tracks are produced using a mix of AI tools and human composers, arrangers, and editors — every release gets human creative direction before it is published.
In practice: we use generative tools to sketch ideas, explore harmonic directions, and speed up production. Then human composers and editors shape the material — choosing what stays, refining arrangements, mixing, and signing off on the final master. We do this for two reasons. First, quality: AI-only output doesn't reliably hit the specific functional targets each mode requires. Second, rights: purely AI-generated audio has uncertain copyright status in many jurisdictions without meaningful human creative input, and our licensing depends on a clear human contribution.
We'd rather tell you this up front than bury it in a footer.