01
Musicians can't prove it.
Recordings and compositions train AI models without consent, credit, or payment. Opt-out signals don’t scale, and there’s no shared record of who made what, when.
AI music is built on real music — yours.
Proof, Rights & Royalties.
For musicians, rights-holders and the AI companies that train on their work.
Source Layer establishes where recordings and compositions came from, who holds the rights, and how creators are credited when their music trains AI models. Starting with music — expanding to other creative industries next.
The infrastructure gap
The UK's £145.8bn creative economy has no shared layer for music provenance, rights, or attribution.
02
AI developers can't license it.
Teams training on music want a lawful route, but there’s no standard way to establish provenance, clear the rights, or verify what’s already in a dataset.
03
Labels & publishers can't attribute it.
Catalogues are ingested at scale. Rights-holders have no infrastructure to trace usage, attribute works, or license to AI at the speed the models move.
04
Regulation demands it.
The EU AI Act (Art. 53, August 2026) mandates training data disclosure. The UK rejected broad TDM exceptions. Compliance is no longer optional — but the infrastructure doesn’t exist yet.
Source Layer is the infrastructure connecting every side of music: the artists and rights-holders who own it, the AI developers who need lawful access to it, and the regulations that now require a record. Starting with music — expanding to other creative industries next.
How it works
Three layers. One platform. Provenance, rights, and attribution for music.
01 · PRV
Prove It
Provenance
Register a recording or composition with a cryptographic fingerprint and a C2PA content credential — verifiable, timestamped proof of where the music came from. The foundation every rights and attribution decision is built on.
02 · RTS
License It
Rights
Establish who holds the rights and open a clear licensing pathway for AI training on that music. Standardised terms, usage-based royalties, and EU AI Act documentation — through a single integration.
03 · ATT
Protect It
Attribution
Track where registered music surfaces in training datasets and model outputs, and credit the rights-holders and creators behind it — evidence-grade, and at catalogue scale.
For creators · starting with music
Your music. Your rights. Your terms.
01 · MUSIC
Musicians & Songwriters
Your catalogue has value. Register recordings and compositions, see when AI trains on them, and get credited and paid when it does.
02 · LABELS
Labels & Publishers
Bulk-register your catalogue via API. Establish rights, receive attribution reports, and license to AI developers through standardised infrastructure — not one-off deals.
Soon
Visual Artists & Illustrators
Provenance and attribution for images — no poisoning, just verifiable proof.
Soon
Authors & Writers
Establish where your text came from, and prove it when a model trains on it.
Soon
Games & Film
The same provenance, rights and attribution for interactive and screen work.
Source Layer is built for music first. Visual art, writing, games and film come next — the same provenance, rights and attribution, one industry at a time.
Available now · For rights holders
The Catalogue Audit
A paid audit for rights holders. We cross-reference your catalogue against the major public AI training datasets and return a structured PDF risk report — a clear, evidence-backed read on where your music stands.
01 · LICENSING READINESS
AI licensing readiness
We score how cleanly each work can be identified, owned, registered and licensed — six dimensions in all — and hand you a prioritised fix list. Every line is evidence-backed against independent rights and credits sources, so your team knows exactly what to resolve first.
02 · DATASET PRESENCE
Dataset presence
An evidence-based indication of which of your recordings appear in the datasets now circulating among AI developers. It’s inference, not a legal opinion — presence isn’t proof a model trained on a track, and absence isn’t proof it didn’t — but it gives rights and licensing teams something concrete to work from.
For large catalogues we don’t price per work. We assess a representative sample — typically 50–200 works — and report a catalogue-level result, so cost tracks scope.
The Catalogue Audit is available today, offered as a no-obligation sample run to start. An ongoing provenance and monitoring platform is being built behind it — audit clients are first in line when it goes live.
Request a sample audit →For AI developers
Licensed music training data. One integration.
01 · COMPLY
EU AI Act ready
Art. 53 mandates training-data disclosure from August 2026. Source Layer gives you provenance records and an audit trail for the music in your datasets — before the deadline.
02 · LICENSE
License music at scale
Access properly licensed music through a single API. Standardised terms, usage-based pricing, and rights cleared at the source — no chasing individual rights-holders.
03 · VERIFY
Prove your dataset is clean
Run your training data against our fingerprint database. Get a certificate showing which music is licensed, which isn’t, and what needs attention.
Register interest
Join the early access programme.
Be first to prove, license and protect your music.
Whether you make music or build AI that trains on it, we'd like to hear from you. Starting with music — other creative industries next.
We'll never share your data. Source Layer is built on transparency and creator control.