ElevenLabs has introduced References for its Music v2 model, a new feature that allows users to upload existing music tracks to guide the style and feel of AI-generated songs. Rather than copying the uploaded audio, References uses it as a creative guide, enabling Music v2 to produce original compositions that capture elements such as mood, instrumentation, tempo, and production style. The feature is now available across ElevenMusic, ElevenCreative, and the ElevenLabs API, giving musicians, brands, and developers greater control over AI-generated music.
The launch expands the capabilities of Music v2, which already supports long-form composition, multilingual lyrics, section-by-section editing, and advanced inpainting. By allowing creators to steer music generation with reference tracks, ElevenLabs aims to bridge the gap between prompt-based generation and more precise creative direction while maintaining safeguards against copyright infringement.
References Brings Style-Guided Music Generation
With References, users can upload an existing audio track to influence the musical characteristics of newly generated content.
The AI can draw inspiration from attributes such as:
- Musical style.
- Overall mood.
- Instrumentation.
- Tempo and rhythm.
- Production quality.
- Sonic atmosphere.
The generated output remains a new composition rather than a duplicate of the uploaded reference. All uploaded tracks are screened for copyright compliance before they can be used.
References Feature Overview
| Feature | Details |
|---|---|
| Input | User-uploaded reference track |
| AI Model | Music v2 |
| Purpose | Guide musical style and feel |
| Platforms | ElevenMusic, ElevenCreative, ElevenAPI |
| Copyright Protection | Uploaded tracks undergo copyright screening |
More Creative Control for Music Generation
References gives creators a more intuitive way to communicate creative intent than relying solely on text prompts.
Potential use cases include:
- Developing songs inspired by a particular musical style.
- Maintaining a consistent sonic identity across projects.
- Creating branded audio for advertising campaigns.
- Producing music for games, videos, and podcasts.
- Building music generation tools through the ElevenLabs API.
By combining prompts with reference tracks, users can achieve more predictable and consistent results when generating music.
Part of the Music v2 Ecosystem
References builds on the broader capabilities introduced with Music v2.
Recent additions include:
- Improved vocals and instrumentation.
- Long-form, section-by-section composition.
- Better multilingual lyric generation.
- Mid-track genre transitions.
- Enhanced inpainting for selective editing.
- Embedded sound effects within songs.
Together, these features aim to make AI-generated music more flexible for professional creators as well as casual users.
Music v2 Feature Comparison
| Capability | Music v2 |
|---|---|
| Style guidance with References | Yes |
| Long-form composition | Yes |
| Section-by-section editing | Yes |
| Multilingual lyrics | Yes |
| Improved inpainting | Yes |
| API support | Yes |
Focus on Copyright Compliance
ElevenLabs says every uploaded reference track is checked for copyright infringement before being processed.
The safeguard is intended to:
- Protect artists’ rights.
- Prevent unauthorized use of copyrighted recordings.
- Encourage responsible AI music creation.
- Support commercially usable outputs generated with Music v2.
The company has also stated that Music v2 is trained on licensed data and is designed for commercial use across supported platforms.
Growing Competition in AI Music
The launch comes as AI music platforms continue to compete by offering creators greater control over generated content.
Across the industry, developers are expanding capabilities such as:
- Style-guided generation.
- Voice cloning for singing.
- AI-assisted composition.
- Music editing and remixing.
- Creator customization.
- Enterprise APIs for music generation.
As AI-generated music becomes more sophisticated, precision and creative control are becoming key differentiators.
Why It Matters
The introduction of References represents a significant step toward making AI music generation more controllable and predictable. Instead of depending entirely on descriptive prompts, creators can now guide Music v2 with real audio references that communicate stylistic intent more naturally. This approach enables musicians, brands, and developers to produce original tracks that better align with their creative vision while preserving originality through reference-based generation rather than direct replication.
The feature also reflects the broader evolution of generative AI from simple content generation toward professional creative tools. By combining reference-guided generation with copyright screening, licensed training data, and API integration, ElevenLabs is positioning Music v2 as a platform suitable for commercial music production, marketing, game development, and enterprise creative workflows.
Looking Ahead
ElevenLabs is expected to continue expanding Music v2 with additional features that provide creators with greater consistency and control, including deeper customization options and enhanced workflow integrations. Future updates may further strengthen how users combine prompts, reference tracks, vocals, and editing tools to produce high-quality music tailored to specific creative needs.
Looking ahead, reference-guided music generation is likely to become an increasingly important capability across the AI music industry. As creators demand more precision without sacrificing originality, tools like References could help establish a new standard for AI-assisted music production, enabling faster workflows while supporting responsible use of copyrighted material and commercially deployable content.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
