Component-level content provenance for AI-generated, human-made, and hybrid music.

AI created a new music economy. Rights infrastructure didn't follow.
We built it.

The problem

Generative AI platforms create music by combining melodies, vocals, rhythms, stems, and other musical components that originate from real artists and rights holders. Yet today's infrastructure cannot reliably determine what was used, where it came from, who controls it, or how rights and revenue should be allocated. The generation event happens. The provenance is lost.

The challenge extends far beyond AI-generated music. Unauthorized sampling, voice cloning, fraud, and hybrid works where AI and human creativity coexist all face the same limitation. The industry lacks the infrastructure to establish content provenance, identify rights at the component level, and support licensing, remuneration, and enforcement.


  • $B+

    Value untracked every year

  • $B+

    Global music copyright economy

  • The gap

    Between creation and monetization

How we solve it

Every piece of music, whether AI-generated, human-made, or hybrid, is built from components: melody, rhythm, vocals, bass lines, instrumentation, and lyrics. Every component has an origin. Every component has rights attached to it.

We built the infrastructure to identify those components, establish their provenance, connect them to the relevant rights holders, and record their usage.

Music is decomposed into atomic components, the smallest meaningful musical units. For every atomic component, musicDNA identifies its origin, its transformation, its associated rights information, and its contribution to the final output.

During generation, provenance is recorded in real time. For existing recordings, musicDNA reconstructs that provenance by identifying the same atomic components and matching them against a reference database. The result is a complete, auditable provenance record.


Provenance recorded across four categories

Musical Work

Sound Recording

Vocal Stem

Lyrics

MixAudio

B2C

Licensed AI remix platform with provenance at generation.

Users select a track from a fully licensed catalog and remix it. Every stem, vocal, melody, lyric, and licensed component used during generation is recorded in real time. Nothing is reconstructed afterwards. Rights holders receive detailed attribution reports and are compensated based on actual component usage.

MixAudio workspace screenshot

Three revenue flows

01

Component usage during generation

02

Streaming of the original work

03

Streaming of the remix

Usage-based micro-settlement. Not pool-based. Every rights holder is paid according to what was actually used.
Also available as a white-label solution for labels, DSPs, and music platforms.

Two layers

Layer 1 - Licensed Input Track

Preserves original melody

StemsVocalsMelodyDrumsBass

Licensing attached to the work defines the applicable rights and how revenue is allocated.

Layer 2 - Licensed Component Library

Modify arrangementChange tempoAdd licensed componentsTransform vocalsOwn performances

Output · provenance recorded across

Musical WorkSound RecordingVocal StemLyrics

musicDNA

B2B

Component-level content provenance for any audio output.

Submit any AI-generated, human-made, or hybrid recording. musicDNA decomposes audio into atomic components, typically 2 to 4 bars or 8 to 16 bars depending on the analysis, and matches each component against a reference database of musical works. Publishing and recording rights are analyzed simultaneously.

01 Submit

Submit Audio

Any AI-generated, human-made, or hybrid recording.

02 Decompose

Decompose

Into atomic components - 2-4 or 8-16 bars.

03 Match

Match

Each component against a reference database.

04 Report

Report

A complete attribution & compliance record.

musicDNA attribution analysis

Output & applications

Report includes

  • Copyrighted works identified
  • Atomic components detected
  • Contribution levels
  • Degree of transformation
  • Copyright infringement indicators
  • Compliance status

Audio is processed transiently. Audio enters the system, embeddings are extracted, and the original audio is discarded. No centralized audio storage is required. Available through API.

Applications

Copyright infringement detectionFraud detectionAI & hybrid music analysisPre-distribution screeningVoice-cloning detectionRights managementReporting & compliance

The foundation

ISBC

International Standard Sound Block Code

ISBC introduces a granular identifier for atomic musical components, connecting ISRC and ISWC at the component level. It is designed in alignment with DDEX standards.

ISRCISBCISWC

BlockDB

Provenance-aware registry

A registry connecting atomic musical components, usage events, rights data, and revenue logic at scale.

ComponentsUsage eventsRights dataRevenue logic

Standards & policy

We don't just build for current standards. We help write the next ones.
DDEX – AI WG · MRT WG · Studio WGC2PAWIPO AIII TEN

Recognition

CES Innovation Award

2023

German Design Award

2024

Korea Ministry of Culture, Sports and Tourism

Recognition

Backed by

Lee Soo-man

Founder of SM Entertainment and A2O Entertainment

InterVest

Top-tier Korean venture capital

Kakao Ventures

Strategic investor

Research

About Neutune

Neutune was co-founded in 2020 by Dr. Jongpil Lee and researchers from KAIST's Music and Audio Computing Lab, Korea's equivalent of MIT or Stanford. We were built on a single conviction: music cannot be fairly monetized in the AI era without infrastructure that operates at the component level.

Active members of DDEX across three working groups, C2PA, and the WIPO AIII Technical Exchange Network. Backed by Lee Soo-man, founder of SM Entertainment, InterVest, and Kakao Ventures. We have worked with KOMCA and the Korean Copyright Office on attribution-related initiatives, and our previous MixAudio platform generated 1.3 million tracks across 220 countries, with commercial deployments across automotive, smart TV, mobile, and gaming platforms including Renault, Hyundai, and LG Electronics.

AI researchers & ML engineers

Patents filed

2020

Founded

Contact

Jongpil Lee

Jongpil Lee

CEO

Technology and research

Virginie Berger

Virginie Berger

Chief Industry and Rights Officer

Partnerships and industry

© 2026 Neutune. Building attribution, provenance and rights infrastructure for AI music.