ElevenLabs has introduced a major update to its ElevenMusic platform, enabling the creation of musical compositions using custom voices and maintaining a consistent musical style.

!image.webp&w=3840&q=80)
What Happened
As part of the ElevenMusic platform update, the Vocals and Styles features were introduced. The Vocals feature allows users to create songs using their own voice by uploading samples to train the Music v2 model, or by choosing ready-made options from the Vocal Library. The Styles feature ensures sonic consistency, allowing users to maintain a unified musical "signature" and style across different tracks.
Context
With the release of Music v2, ElevenLabs is moving from simple generation of disconnected audio clips to tools for full musical identity management. These new capabilities enable one-shot/few-shot learning mechanisms for precise cloning of timbre and performance manner, addressing the problem of consistency—one of the key challenges in generative audio.
Why It Matters for the Industry
ElevenLabs is transforming the music creation process, moving from random generation to the management of digital identity. The vertical integration of voice and style management creates an ecosystem for developing full-fledged AI performers. This intensifies competition in the AI music segment and sets new standards for professional content production.
Why It Matters for Users
For independent musicians and content creators, this opens up opportunities to create unique and repeatable digital characters with their own recognizable timbre. These tools allow for rapid song prototyping and the creation of branded audio content, significantly lowering the barrier to entry for high-quality AI production.
What Is Not Yet Known / Limitations
Despite the creative possibilities, the implementation of such technologies carries engineering and operational risks, such as high inference costs, difficulties in managing generation quality, and legal issues related to the use of biometric voice data.
Sources
Author
Look at AI, Editorial Team
