The question most platforms cannot answer.
What was the model trained on? As platforms tighten their standards and regulators define new requirements, that question has become the one that separates AI music systems built to last from those built to move fast.
We built HitZERØ so that we could answer it clearly, back it with documentation, and open that documentation to review. We hold ourselves to a framework we call the IAH Provenance Framework: eight criteria, each designed to be verifiable through documentation, not merely asserted. HitZERØ maintains verified conformance with all eight.
Read the IAH Provenance Framework
Four categories. Documented sources.
The IAH Resonance Engine™ that powers HitZERØ was trained on data drawn from four categories. Each is recorded in our Master Training Data Provenance Registry.
Original in-house works.
Compositions produced inside iah.fit inc. and wholly owned by the company.
Commissioned works.
Music created by independent creators and delivered to iah.fit inc. under platform terms of service that assign intellectual property rights in the delivered work to the buyer upon payment. All creators were compensated in full. Each commission is documented through platform records and executed IP declarations.
Licensed for machine learning.
Datasets acquired under paid commercial licenses that expressly grant rights to use the content for machine learning training and to commercially exploit models and outputs trained on the data. This includes a machine learning license agreement with Rightsify Group, LLC, a commercial music data licensor whose agreement is purpose-built for AI model training and under which Rightsify indemnifies the licensed dataset against infringement claims.
Open academic research.
Vocal and musical datasets published by research institutions under open academic and Creative Commons terms, including VocalSet (Wilkins et al., 2018), licensed under CC BY 4.0.
What we leave out matters as much as what we put in.
We built our acquisition practice on clear lines, enforced from the first day of development.
We acquire no audio through scraping. We acquire no rips from Spotify, Apple Music, YouTube, or any streaming platform. Third-party datasets enter only under written license. Customer data stays out of training entirely. The prompts you write and the music you create are yours. They do not become training material for the model.
Documentation, not promises.
Every source is recorded in our Master Training Data Provenance Registry. For each source, the registry captures the rights holder, the acquisition terms, and the rights basis. The integrity of that registry is supported by executed declarations and documentation maintained by the company.
For partners, distributors, and platforms that need to go deeper, we open our provenance documentation to legal counsel under non-disclosure agreement. We are prepared to extend contractual indemnification for training data provenance through partnership agreements.
Partners, distributors, and platforms can request our provenance documentation for review under non-disclosure agreement.
You own what you create.
Every track you create on HitZERØ comes with worldwide, royalty-free commercial rights. Under current U.S. copyright law, works generated solely by AI are not eligible for copyright protection, so those rights are secured by license under your Commercial Use and Rights Certificate rather than by copyright. The full scope is defined in that certificate and our Terms of Service.
Documented training data is the foundation that makes that ownership defensible.
Questions, answered.
What was HitZERØ trained on?
The IAH Resonance Engine™ that powers HitZERØ was trained on data from four categories: works commissioned under platform terms of service that assign IP rights to iah.fit inc. upon payment, datasets licensed under paid commercial agreements that expressly grant machine learning training rights, open academic research datasets including VocalSet (Wilkins et al., 2018) under CC BY 4.0, and original works produced in house. Every source is recorded in our Master Training Data Provenance Registry.
Did you scrape music from the internet?
No. We acquire no audio through scraping. Every source enters under documented terms.
Do you use my music or my prompts to train your model?
No. Customer prompts, recordings, and generations stay out of the training corpus. What you create on HitZERØ is yours. It does not become raw material for the model.
Can you prove any of this?
Yes. Every source is documented in our Master Training Data Provenance Registry, which records the rights holder, the acquisition terms, and the rights basis for each one. The registry is supported by executed declarations and documentation maintained by iah.fit inc. We open this documentation to qualified partners and their counsel under non-disclosure agreement.
Read the full Training & Ethics FAQ
We did the work. The records are here.
AI music is under scrutiny. Platforms are raising the bar on documentation. Regulators are defining new requirements. HitZERØ was built on a foundation designed to meet that standard of review.
Fuel Your Frequency.
This page describes the provenance and rights of HitZERØ training data. It does not describe model architecture or generation methods, which remain proprietary. The sources described on this page are recorded in the Master Training Data Provenance Registry maintained by iah.fit inc.