ElevenLabs, the three-year-old audio-generation startup valued at more than $3 billion, has introduced “Eleven Music”, an artificial-intelligence model that turns written prompts into complete songs, jingles and background tracks. The company says the service is available to individuals and businesses and that all output is pre-cleared for commercial use. To support that claim, ElevenLabs says it trained the model only on licensed material and has struck agreements with digital rights managers Merlin Network and Kobalt Music Group. The startup has also built safeguards to block requests that could reproduce protected works or offensive content. The launch moves ElevenLabs beyond its core text-to-speech technology and into a field that has drawn heightened legal scrutiny. Competitors Suno and Udio were sued last year by the Recording Industry Association of America over alleged copyright infringement. By lining up licensing partners in advance, ElevenLabs is positioning its new tool as a lower-risk option for filmmakers, advertisers, game developers and other creators seeking custom music on demand.