ElevenLabs Launches AI Music Generation—With Artist Consent Built In

ElevenLabs, a company known for AI voice synthesis tools, entered the music generation space in August 2025 with a service called Eleven Music. The platform can create full studio-quality tracks in multiple genres—with or without vocals—in minutes, and it comes pre-cleared for commercial use in film, television, podcasts, videos, ads, and games.

You can generate music through ElevenLabs' website or through their API (a programming interface that lets other software talk to the service). The underlying technology, called Music v2, handles multiple languages and lets you build songs piece by piece: intro, verse, chorus, and so on, while keeping everything musically coherent.

How ElevenLabs Obtained Training Data

Here's where ElevenLabs took a different path than some competitors. Instead of scraping music from the internet without permission, the company partnered with Merlin Network and Kobalt Music Group—organizations that represent thousands of independent artists and labels. The key difference: artists must actively opt in to let their music be used for training. They are not harvested without consent.

This matters because other AI music companies have faced lawsuits over how they sourced training material. By building partnerships and requiring artist consent, ElevenLabs aimed to address the copyright concerns that have made AI-generated music risky in professional settings, where getting the rights cleared is mandatory before you can publish or distribute.

What the Technology Can Do

Music v2 generates vocals, instruments, and arrangements with better quality than the previous version. You describe what you want—say, "1980s synthwave" or "Traditional Spanish flamenco"—and the system creates it in your chosen language.

One useful feature is Music Finetunes. You can upload your own tracks to customize how the model generates music, essentially teaching it your style. It then remembers that style across multiple songs you create, rather than requiring you to specify everything from scratch each time. ElevenLabs offers 11 pre-made style templates too, or you can build your own.

The editing tools let you change lyrics, add or drop song sections, adjust how long a track is, and apply different style keywords. This means you can iterate—create something, tweak it, try again—rather than just clicking a button once and getting a final result.

A Marketplace and Multiple Entry Points

ElevenLabs built a marketplace into the platform. Users can publish their AI-generated tracks there, and when paying subscribers use that music in their own projects, the creator gets paid. It is similar to how stock music websites work, but powered by AI generation.

The music service runs on three different platforms: ElevenMusic for people making music on their own, ElevenCreative with more advanced tools, and an API for developers who want to build music generation into other software. One underlying technology, different ways to access it depending on who you are.

Why This Matters for Professional Work

The "cleared for commercial use" claim addresses a real problem. Many AI music generators operate in legal gray areas that make them unsuitable for professional work, where you need to prove you own or have licensed the rights to everything in your project before distributing it. If you cannot prove that, your film or podcast or advertisement can get blocked or sued.

By securing partnerships with major music rights holders and requiring artist consent, ElevenLabs gives customers legal confidence. The API availability also means companies can embed music generation directly into their existing creative tools and workflows, not just use a standalone website.

The competitive context here is instructive. ElevenLabs enters a market where services like Suno and Udio have already gained users and attention, but both face ongoing legal challenges about where they got their training data. ElevenLabs' opt-in approach may limit how much training data it can collect—consent from artists is more restrictive than unlimited scraping—but it provides clearer legal footing for paying customers.

This pattern has played out before in AI. When image generation was new, companies like Stability AI trained models on billions of images scraped from the web without artist permission; that sparked lawsuits. Now we are seeing a shift toward licensing: Google negotiates specific agreements with publishers for AI training, and OpenAI has started paying media organizations for content. The legal and regulatory pressure forces maturity. The music generation space appears to be hitting that same inflection point now, where getting the law right becomes as important to widespread adoption as having good technology.

In this author's view, ElevenLabs' emphasis on consent and commercial licensing positions the company well for enterprise customers—those who need music fast but cannot afford legal risk. That may prove more valuable in the long run than simply having the most advanced model.

How Songs Get Built and Localized

The section-by-section composition addresses a real limitation in earlier AI music systems. Older tools could produce short musical snippets but struggled to create coherent full songs. By maintaining structure across intro-verse-chorus progressions, Music v2 lets you generate complete tracks that sound intentional rather than fragmented.

Multilingual support works beyond simple translation. The system can generate vocals in different languages while keeping the song's overall style intact. That matters for content creators working globally who need a Spanish version of a song, a German version, a Japanese version—without re-recording everything from scratch.

Whether you use the web interface or the API, the core technology is the same. Some people will compose music directly on the website; developers will embed the capability into larger production systems. It is flexibility for different workflows.

The convergence of AI-generated music with formal licensing and rights agreements marks a shift in how synthetic media enters professional creative work. Clear legal standing and good technology together may finally make AI music a reliable tool rather than a legal minefield.