Beatoven.ai launches fully licensed generative AI music model
Beatoven.ai's new foundation model, Maestro, positions itself as the first generative AI music model built entirely on licensed data.
Music technology startup Beatoven.ai has unveiled Maestro, a new generative AI foundation model designed to create instrumental music with future extensions for sound effects and vocals. The company says it is the first such model to be fully licensed, ensuring ongoing royalty payments to artists, composers, and rightsholders whose works contribute to AI-generated outputs.
Unlike other AI music tools that rely on scraped datasets, Maestro has been trained through official licensing partnerships managed by Musical AI, a rights management platform that attributes outputs to specific tracks. Current data partners include Rightsify, Soundtrack Loops, Symphonic Music, Bobby Cole, Vadi Sound, and Pro Sound Effects, with more expected.
Beatoven.ai, which has two million registered users who have already generated more than 15 million tracks on its earlier systems, said Maestro can be fine-tuned for new genres and styles. The platform also includes Musical Intelligence tools for catalogue owners, allowing labels and publishers to analyse their music, generate metadata, and improve the discovery of back-catalogue tracks.
Mansoor Rahimat Khan, Co-Founder and CEO of Beatoven.ai, described the launch as a step toward building an “AI-powered music ecosystem” that rewards human creativity. “Most tools try to mimic humans, whereas AI should push human creativity forward by generating what we’ve never heard before,” he said.
Industry voices welcomed the approach.
Ed Newton-Rex, CEO of Fairly Trained and a long-time advocate for ethical AI training, called it “compelling evidence that generative AI doesn’t need to be built by scraping the music of the world’s musicians.” Sean Power, CEO of Musical AI, said the model demonstrates what a fair AI licensing deal looks like, combining attribution, respect for rights, and payouts every time a contribution is used.
Alongside its direct-to-creator offerings, Beatoven.ai continues to operate its generative music API, currently described as the most widely used in the sector. The company said it plans to expand API access further in 2025 to serve industries such as gaming, film, and virtual production, providing additional revenue streams to musicians and rights holders.
Co-founded by musician and entrepreneur Mansoor Rahimat Khan in San Francisco and CTO Siddharth Bhardwaj in Bengaluru, Beatoven.ai positions Maestro as both a creator tool and a business platform for the music industry. The company said the model aims to bring professional-quality, ethically sourced generative music to a broad user base while ensuring that artists share in the economic benefits of AI.
How Beatoven.ai is differentiating with licenses
Licensing data, rather than scraping it, has become a central issue in the debate over how generative AI should be built. At its core, the difference is about legality, fairness, and sustainability. Scraping refers to collecting large amounts of content from the internet — often copyrighted works such as songs, images, or writing — without explicit permission from the creators.
This practice raises significant copyright concerns, since the material is used to train AI systems that can produce new outputs but generate no compensation for the original rightsholders.
When companies license datasets, they enter into formal agreements with labels, publishers, or artists. This means the use of creative works in AI training is lawful and transparent, avoiding the legal risks that have already sparked lawsuits against several AI developers.
More importantly, licensing ensures that creators receive compensation when their work contributes to AI outputs. For musicians, that can mean ongoing royalty payments whenever an AI-generated track draws on elements from their licensed material.
Licensing also signals respect for human creativity. Scraping can feel exploitative, treating artists’ works as free raw material. In contrast, licensed models show that AI innovation can coexist with creative rights, building trust with the very industry it hopes to serve. That trust is crucial for long-term adoption. An AI company that pays creators is more likely to attract partnerships with record labels, publishers, and other players in the ecosystem.
There are practical benefits as well. Scraped datasets are often noisy, with mislabeled or poor-quality data, leading to inconsistent AI outputs. Licensed datasets tend to be curated and come with rich metadata such as genre, tempo, or instrumentation. This not only improves model performance but also makes it possible to attribute AI outputs to specific tracks — a key step in enabling royalties and accountability.
Ultimately, while scraping data may allow for rapid scaling, licensing offers a more ethical and sustainable path forward. It balances the promise of AI-generated music with the rights and livelihoods of human artists, ensuring that technological progress does not come at the expense of those whose creativity fuels it.
Edited by Jyoti Narayan