Creator Guide: ML Training & Industrial Earnings
Welcome to the Derived Audio Data SKU. This guide explains how you earn money when your voice or data is used for Machine Learning (ML), AI training, and industrial applications.
The Democratic Payout Model
Unlike the standard marketplace where your price might fluctuate based on subjective popularity, the ML Training SKU uses a Cohort Pro-Rata Model. This ensures that everyone in the same "Industrial Bucket" is paid equitably for their contribution based on the volume of data provided.
1. The 70/30 Human-First Split
We believe the humans providing the data should receive the majority of the value.
- 70% of Gross Revenue goes directly into the Human Pool.
- 30% is retained by the platform to cover verification, security, and infrastructure costs.
2. Understanding "Industrial Buckets"
Your payout is determined by which "Bucket" your data is placed in by the buyer. Each bucket has a multiplier that increases your share of the Human Pool.
| Industrial Bucket | What happens to your data? | Your Payout Weight |
|---|---|---|
| I. Analysis | Researchers extract "features" (like pitch or tone) or use LLMs to label your files. | 1.0x (Base) |
| II. Transformative | Your voice is used for style transfer or watermarked streaming. | 2.0x |
| III. Foundational | Your voice is used to train AI models or create clones. | 10.0x |
Why is this better for you?
Cohort Pro-Rata Distribution
Within a specific bucket, payouts are distributed pro-rata based on the volume of data contributed. If a buyer licenses a "Foundational" training set, revenue is split among creators in proportion to the seconds of audio they provided to that cohort. This ensures fair compensation for larger contributions.
Evolutionary Access (Sync Fees)
When you join the ML Training SKU, you aren't just selling a snapshot of your voice. You are providing access to your "Evolution."
- Initial Payout: You get paid when you are first included in a bundle.
- Sync & Update Fees: If you upload higher-quality data or new samples that improve the buyer's model, you may trigger additional "Sync" payments as the buyer updates their dataset.
Legal Protection
The platform acts as your agent. We ensure that:
- Buyers are verified (KYC).
- Your data is never "scraped" without consent.
- Your Intellectual Property (IP) is respected through clear, enforceable digital contracts.
FAQ for Creators
Q: Can I opt-out of the "Foundational" bucket? A: Yes. You control your permissions. If you only want your data used for "Analysis," you will only be included in those bundles. However, the Foundational bucket offers a 10x higher payout weight.
Q: How do I increase my earnings? A: Earnings in this SKU are driven by Quality and Data Volume. Ensuring you have at least 60 minutes of high-fidelity (Tier 3) audio makes you eligible for the highest-paying "Foundational" bundles.