Your Music Powers What We Build.
You Own a Share of What It Becomes.
CORPUS builds AI music technology — adaptive, context-responsive, trained on a shared library scored for quality, originality, and diversity. Contributors keep their rights and hold a lasting stake in the system.
Scroll to explore
Why This Exists
Your music may already be training AI models — without your knowledge, your consent, or your compensation. Most training data is scraped or bought out. Musicians lose control. The resulting models reflect a narrow slice of the world's music.
CORPUS takes a different approach
Musicians contribute under explicit consent. The library grows more valuable as it grows more diverse — and powers applications where music has never had an economic infrastructure: mobility, healthcare, interactive environments.
How the Scoring System Works
Every contribution is scored once, at the moment it enters the library. Your score determines how you participate in the revenue from AI products built on the library.
Whether it's a finished track, a stem, a field recording, or a set of detailed annotations — every contribution is evaluated for what it adds to the library as a whole.
Contributions are scored on:
- •Quality — production standards, metadata completeness, technical integrity.
- •Originality — how your work positions itself relative to what's already in the library.
- •Diversity — contributions that expand the library into underrepresented territory earn more.
Points translate into revenue shares. As the library generates income, it is distributed proportionally among contributors. Higher scores mean a larger share — and because scores are fixed at the moment of entry, your position is locked in.
For Musicians
How contribution works — and what it earns you.
Upload your music under explicit, revocable consent. You keep your rights. Every work is evaluated for how it strengthens the library.
Your score determines your share of revenue from every AI product built on the library. Royalties are ongoing — not a one-time payment.
Every contribution also generates a lasting participation right in the system itself. As the library grows and more products are built, your stake grows with it.
For Product Teams
Music that responds to context — adaptive, licensed, and ready to integrate into vehicles, games, and interactive environments.
Sound that evolves with use — not playlists, not pre-produced loops. Continuous musical response shaped by what's happening in the product.
Design the sonic behavior without composing every note. Set parameters for mood, intensity, and narrative arc — the model handles variation and continuity.
Built on licensed contributions with auditable provenance. Safe to ship in regulated industries, brand-sensitive products, and consumer-facing applications.
What We're Building
The library powers real products. These are the first.
A real-time music model that generates continuous sound responding to context — driving dynamics, player behavior, user interaction. Not playback. Musical behavior.
Describe a scene, a mood, a narrative arc. Get musical ideas that match — connecting creative intent to the library through semantic understanding, not keyword search.
High-resolution semantic analysis of music: mood, narrative function, instrumentation, cultural context, production characteristics. Structured descriptions, not just tags.
Built from Artistic Practice.
Co-funded by the EU.
CORPUS is developed by Sofilab, a Munich-based sound design and AI lab, and co-financed by Creative Europe.
The core team are practicing musicians and sound artists. CORPUS is not a tech platform proposing to fix music — it's a proposal from within the practice.
Data hosted in Germany under EU law. All AI models run on our own servers — no third-party APIs, no external processing. Your music never leaves our network. Built for regulatory compliance. Designed for global reach.
The Beta Is Almost Here
The platform is almost ready. The scoring system is set. Early contributors will be the first to shape the library — and the first to earn.
Questions? info@sofilab.art
Cookie Settings
We use cookies to enhance your browsing experience, serve personalized content, and analyze our traffic. By clicking "Accept All", you consent to our use of cookies.