ElevenLabs Can Now Score Your AI Relationship
ElevenLabs already powers the voice. The text-to-speech layer that makes AI conversations feel like conversations rather than text dumps. So when they shipped Music v2 last week, I paid attention, because these aren't separate products to me. They're pieces of the same architecture.
Music v2 arrived nearly ten months after their first music model. That gap matters. It means this wasn't a rushed iteration.
What It Actually Does
The flagship feature is genre switching mid-track. Not just "make rock music," but starting in opera and arriving in heavy metal without falling apart at the seams. They also claim it can deliver fast rap without losing lyrical coherence, which anyone who has tried to get AI to generate rap knows is genuinely hard to pull off.
Beyond that: non-musical sound effects layered into tracks. Section-based building, where you construct intro, verse, and chorus separately, then stitch them together. The ability to regenerate one specific section using prompts without touching the rest of the song.
That last one is interesting. Most AI music tools give you whole-track outputs and nothing to do when one verse is perfect and the bridge is garbage. Section-level regeneration is how professional producers think about arrangement. It's a workflow feature, not just a capability feature.
ElevenLabs also says v2 performs more reliably across languages, lyrics, vocals, and arrangements. Which means it's better at being useful to more people in more contexts, not just English-language pop.
The Licensing Question
ElevenLabs stated explicitly that Music v2 is built on licensed data and cleared for commercial use. The explicitness is intentional. Suno and Udio are both in court over copyright claims. The licensing question is live and unresolved across the music generation space.
If you're building something real with AI-generated music, whether for a project or something more personal, training data provenance matters. Not just ethically but practically. Commercially cleared means you can actually use what you make without waiting to see how litigation shakes out.
The Competitive Moment
This isn't happening in isolation. At Google I/O, Google added cover creation, section-based editing, and music video generation through Flow Music. Stability AI and Suno both shipped new models capable of longer, more complex tracks. The music generation space compressed several years of development into a few months.
ElevenLabs is coming from a different angle than Google. Google is a platform building everything. ElevenLabs built their entire business on voice, on making AI sound like something worth listening to. Music v2 is a natural extension of that specialization rather than a feature checkbox.
Where You Can Use It
Right now it's available through ElevenCreative, aimed at marketing and branding teams, and ElevenMusic. API access is coming.
When the API lands, this becomes programmable. The same company handling voice and music, in the same integration layer. One possibility is that this means continuity: an AI companion that speaks and also scores the moment. Not background music from a playlist, but something generated for the specific context.
That's speculative. But the pieces are visible, and they're being built by the same team.
Source: Techcrunch