Any musician who creates and owns their music. We mean that literally, we want to hear your unique sound, your voice, your style. Just hit record and upload your work to CORPUS.
That said, we're launching in phases. The legal situation in music is complex, and if we waited for every legal detail to be perfectly resolved, we'd never get started. So here's how we begin:
We're onboarding early contributors who meet the following criteria:
- You write and produce your music independently, without ties to a label or a royalty collecting society (like GEMA, SACEM or ASCAP).
- You either played all the instruments yourself, or you personally know the other musicians involved and can get their permission. The CORPUS platform will support you in handling those agreements.
As we grow, we'll expand who can join. But for now, we're building a clean foundation — together with artists who can ensure full control over their rights.
By contributing your music to CORPUS, you're helping build a new kind of AI music system: a controllable, real-time, modulatable text-to-music model.
This isn't just a tool that spits out music. It's designed to generate adaptive, dynamic music that can respond to inputs in real time — for use in games, interactive environments, and beyond. Imagine your track becoming fully responsive: shifting in intensity, density, or tone depending on what's happening in the experience. All without the tedious and expensive process of manually creating adaptive music assets.
That kind of technology is usually reserved for high-budget productions. CORPUS aims to make it accessible for a much wider range of creators and projects.
For composers and producers, this opens up new creative possibilities. In the future, you might fine-tune your own model inside a game engine like Unity or Unreal, feed it your sound world, set a few behavioral rules, and let the system adapt your music live to gameplay.
What we're not doing: replacing you. Quite the opposite. Our goal is to build tools that expand your creative reach — not automate you out of the process.
All CORPUS-based projects will be communicated transparently. We're working on a governance model that gives contributors a say. You'll be able to decide whether a project aligns with your values — and even collectively reject things that don't.
You can read more about our values on our Culture page.
We use a combination of proprietary AI tools and human input to assess contributions. Our technology helps analyze musical features automatically — including aspects like structure, timbre, and novelty — but it doesn't make decisions on its own. Outputs from these systems are reviewed through a crowd-sourced checking process to ensure fairness and context sensitivity.
We're committed to transparency: you'll be able to see how your points were calculated and why your contribution was valued the way it was.
We believe diversity makes the library stronger — and more inspiring. That's why we actively seek out unique voices, hidden gems, and underrepresented artists. Weird is welcome!
Our dynamic attribution system is designed to reward originality. Every new contribution is compared to what already exists in the corpus. If your music brings something new — a sound, a structure, a style that increases diversity — it earns more points. For the first time, there's a licensing model that doesn't reward mass appeal, but uniqueness.
Who can contribute music to CORPUS?
Any musician who creates and owns their music. We mean that literally, we want to hear your unique sound, your voice, your style. Just hit record and upload your work to CORPUS. That said, we're launching in phases. The legal situation in music is complex, and if we waited for every legal detail to be perfectly resolved, we'd never get started. So here's how we begin: We're onboarding early contributors who meet the following criteria: • You write and produce your music independently, without ties to a label or a royalty collecting society (like GEMA, SACEM or ASCAP). • You either played all the instruments yourself, or you personally know the other musicians involved and can get their permission. The CORPUS platform will support you in handling those agreements. As we grow, we'll expand who can join. But for now, we're building a clean foundation — together with artists who can ensure full control over their rights.
What is CORPUS actually building — and why should I contribute?
By contributing your music to CORPUS, you're helping build a new kind of AI music system: a controllable, real-time, modulatable text-to-music model. This isn't just a tool that spits out music. It's designed to generate adaptive, dynamic music that can respond to inputs in real time — for use in games, interactive environments, and beyond. Imagine your track becoming fully responsive: shifting in intensity, density, or tone depending on what's happening in the experience. All without the tedious and expensive process of manually creating adaptive music assets. That kind of technology is usually reserved for high-budget productions. CORPUS aims to make it accessible for a much wider range of creators and projects. For composers and producers, this opens up new creative possibilities. In the future, you might fine-tune your own model inside a game engine like Unity or Unreal, feed it your sound world, set a few behavioral rules, and let the system adapt your music live to gameplay. What we're not doing: replacing you. Quite the opposite. Our goal is to build tools that expand your creative reach — not automate you out of the process. All CORPUS-based projects will be communicated transparently. We're working on a governance model that gives contributors a say. You'll be able to decide whether a project aligns with your values — and even collectively reject things that don't. You can read more about our values on our Culture page.
How are contributions evaluated in CORPUS?
We use a combination of proprietary AI tools and human input to assess contributions. Our technology helps analyze musical features automatically — including aspects like structure, timbre, and novelty — but it doesn't make decisions on its own. Outputs from these systems are reviewed through a crowd-sourced checking process to ensure fairness and context sensitivity. We're committed to transparency: you'll be able to see how your points were calculated and why your contribution was valued the way it was.
How does CORPUS promote originality and diversity in the dataset?
We believe diversity makes the library stronger — and more inspiring. That's why we actively seek out unique voices, hidden gems, and underrepresented artists. Weird is welcome! Our dynamic attribution system is designed to reward originality. Every new contribution is compared to what already exists in the corpus. If your music brings something new — a sound, a structure, a style that increases diversity — it earns more points. For the first time, there's a licensing model that doesn't reward mass appeal, but uniqueness.
Do I keep the rights to my music after uploading it to CORPUS?
Yes. Under German law, the ownership of your music stays with you — it's non-transferable, and we fully respect that. What we ask for are non-exclusive usage rights to use your music for AI training. That's not hidden in fine print, it's the core of the project. In addition, your tracks will be available for listening on the CORPUS platform. You can revoke your agreement at any time. However, if your music has already been used to train AI models, that part can't be undone — it's technically irreversible. But from the moment you withdraw, your music will no longer be used in any future training runs.
How can I upload tracks?
Once contributions open, you'll be able to upload your music via the CORPUS Dashboard. Subscribe to our newsletter to be the first to know when submissions open.
When will contributions open?
We're planning to launch in October 2025.
Can I contribute unfinished tracks or experimental pieces?
Absolutely. We actively encourage you to share raw, unusual, or experimental material. AI systems can only learn to generate unexpected or unconventional results if they're trained on that kind of input. So please — send us your sketches, fragments, and sound experiments. That's exactly the kind of uniqueness CORPUS is built for.
Can I contribute stems or other materials like midi or scores?
Yes, we highly value them. Stems, MIDI files, and scores provide detailed insights that are extremely useful for training AI systems. If you include these materials alongside your full mix, you'll earn more points as a contributor. The richer your submission, the more it helps — and the more you're rewarded.
Can I upload my tracks to CORPUS and still distribute them elsewhere?
Yes. The usage license you grant to CORPUS is non-exclusive. That means you're free to release and distribute your music on any other platform — as long as you don't sign an exclusive deal elsewhere that would prevent CORPUS from continuing to use your tracks.
How is the value of a point calculated and when can I expect payouts?
First things first: this system is still in active development. But here's how we currently design it: Every contribution to CORPUS is rewarded with a certain number of points, based on our dynamic value attribution protocol. The more valuable or detailed your contribution (e.g. stems, scores), the more points you earn. As new music is added, the total number of points across the entire CORPUS grows. When it comes to payouts — currently planned on a yearly basis — your share is calculated by dividing your individual number of points by the total number of points in the system. The profit CORPUS has made (after covering project costs) is then distributed proportionally based on that share.
Can I remove my music from CORPUS later on?
You can revoke your contribution at any time. However, if your music has already been used to train AI models, that part can't be undone — it's technically irreversible. But from the moment you withdraw, your music will no longer be used in any future training runs.
How can my company legally license music from the CORPUS library for AI or creative projects?
The legal framework and the dataset are still in development, but we're already in conversation with potential partners. If you're interested in using CORPUS for AI training, research, or creative applications, we'd love to hear from you. Get in touch to discuss your needs and expectations — your input helps us shape the platform and ensure it works in practice, not just in theory.
Is the CORPUS library suitable for training AI models, and are the rights fully cleared?
Yes — that's exactly what CORPUS is built for. The entire purpose of the project is to offer a high-quality, diverse music dataset with fully cleared rights. With CORPUS, you can train AI models with confidence — no grey zones, no hidden risks.
What kind of music content is available in the CORPUS library – genres, formats, and diversity?
Our long-term vision is ambitious: we want CORPUS to represent the full range of musical expression — all genres, all cultures, all formats. Of course, we have to start somewhere. As a Europe-based, EU-funded project, our early focus is on underrepresented and regionally rooted music from across the continent. We actively seek contributors who share ethnic, local, or otherwise marginalized musical traditions. In addition, you'll find a growing body of music suited for creative applications like film, games, and interactive media — including instrumental works, experimental pieces, and hybrid genres.
How is the music in CORPUS evaluated and organized to ensure quality and compliance?
At the core of CORPUS is a hybrid evaluation system that combines automated analysis with crowd-sourced review. Our technology checks for originality, structure, and other musical features, while human curators help validate quality and verify that all rights are properly cleared. We're building this system to be transparent and accountable. More detailed insights into how it works will be shared through upcoming blog posts and documentation.
I want to use the CORPUS dataset for my company, how can I do it?
Please reach out to us.