AI-generated music:yes, but transparent.
We accept AI-generated and AI-assisted music on the platform. We don't reject it — but we require transparency, both toward listeners and within our own risk models. This manifesto is binding for every distribution agreement and every advance contract.
1. Per-track transparency requirement
On every track in the release wizard you flag two fields: `audio_ai_usage` (for the audio stem) and `composition_ai_usage` (for the composition). Values: `none` / `partial` / `full`. Misrepresentation is a contract breach and can result in immediate account suspension.
2. What is `partial` vs `full`
`partial`: AI tools used for individual elements (e.g. AI vocals on a human-produced beat, or vice versa). `full`: track was predominantly or entirely created by generative AI without meaningful human editing.
3. DSP compliance
Spotify, Apple Music, YouTube and other DSPs continuously update their AI policies. We forward your track flags via DDEX metadata — we cannot guarantee that every DSP will accept your track. If a DSP rejects on AI-policy grounds there is no refund of the distribution fee.
4. Impact on advance
When generating an advance offer, our internal AI pre-check factors in the percentage of AI-generated tracks. Over 50% AI content → automatic manual review. Over 80% AI content → advance not available in v1, because DSP risk is too unbounded.
5. Royalty treatment
Royalties for AI-generated tracks are treated 1:1 like royalties for human-produced tracks. No percentage cut difference. You keep 100% of your masters, you get the full payout.
6. Training data
We train our internal risk models (AI pre-check, fraud detection) on platform-internal signals — track metadata, streaming curves, royalty statements. We NEVER use your audio files as training data for generative AI.
7. Changes
This manifesto is a self-commitment and may be changed at any time without prior notice. The version currently published on this page is the one that applies.