ElevenLabs just shipped Music v2, a new music-generation model that can switch genres in the middle of a single track. According to TechCrunch AI, the voice AI company built the model to handle both vocal complexity and intricate composition, and it arrives nearly 10 months after the startup’s first music model. The headline trick: a song can move from opera to heavy metal and back without falling apart.
This is a notable jump because mid-track genre changes have been one of the hardest problems in AI music. Keeping a vocal coherent while the style shifts underneath it takes real control over arrangement. ElevenLabs says v2 manages it, and adds a few capabilities that matter more to working creators than to casual listeners.
What Music v2 can actually do
- Genre switching mid-track. Opera to metal and back, in one continuous song. That flexibility is the model’s main selling point.
- Fast rap without losing coherence. Dense, rapid vocals tend to break AI models. ElevenLabs claims v2 holds together at speed.
- Section-by-section building. Instead of spitting out short clips, you can build a song in parts, intro, verse, chorus, then stitch them into a full track. That’s closer to how songs are actually written.
- Targeted regeneration. Pick one part of a song and recreate it with a prompt while leaving everything else untouched. No more regenerating the whole thing to fix one section.
- Non-musical sound effects. The model can drop sound effects into a track, not just notes and vocals.
- Better multilingual performance. ElevenLabs says v2 is more reliable across languages, lyrics, vocals, and arrangements.
The licensing angle is the real story
Here’s what stands out to me. ElevenLabs stressed that Music v2 is built on licensed data and cleared for commercial use, so users can run with the tracks freely. That’s not a throwaway detail.
As TechCrunch AI notes, rivals Suno and Udio have both faced court cases over copyright. Training music models on unlicensed catalogs is a legal minefield, and the labels have lawyers. By striking deals up front, ElevenLabs is selling something its competitors can’t fully promise yet: output you can use in a paid campaign without wondering who might sue. For marketing and branding teams, that peace of mind may matter more than any single audio feature.
Where it fits in a crowded race
ElevenLabs isn’t alone. TechCrunch AI points out that AI labs have been racing to ship professional-grade music tools, with Google, Stability AI, and Suno all releasing models that generate longer, more complex tracks. At Google I/O, Google rolled out easy cover creation, section-based editing, and music video generation through its Flow Music tool.
So the feature set, section editing, targeted regeneration, longer compositions, is becoming table stakes. What separates these players now is data provenance and where the tools live. ElevenLabs is betting that licensed, commercially safe audio plus its existing foothold in voice gives it an edge with business users.
Who can use it and when
- ElevenCreative, the company’s tool aimed at marketing and branding teams.
- ElevenMusic, a newly launched platform for creating AI-generated songs.
Access through the ElevenAPI is coming soon, which is the piece developers will want. The article didn’t detail specific pricing tiers, so anyone planning to build on top of v2 should check ElevenLabs directly before committing.
Why it matters
Mid-track genre switching is a flashy demo, but the licensing story is the strategic move. AI music is heading toward a split: tools that are creatively impressive but legally shaky, and tools that are safe to ship in commercial work. ElevenLabs is planting its flag firmly in the second camp.
If the API launch delivers and the commercial clearance holds up, v2 could become the default choice for teams that need original music without legal risk. The next thing worth watching is how Suno, Udio, and Google respond on the licensing front. For full details, check the original report at TechCrunch AI.