ElevenLabs launched Eleven Music with prompt-led song controls
ElevenLabs launched Eleven Music on August 5, 2025, offering prompt-led music generation with vocal or instrumental output and section-level sound or lyric editing.
Historical reporting: Source 0 is ElevenLabs’ dated Eleven Music launch. Source 1 is current official music documentation, read for present model-workflow context and not used to prove August 2025 entitlement, API behavior or rights terms.
Key takeaways
- ElevenLabs announced Eleven Music on August 5, 2025 and said it was available on its website; the same announcement contains conflicting statements about public API access.
- The launch described natural-language control of genre, style and structure, with vocal or instrumental options, multilingual output and section-level edits.
- ElevenLabs described broad commercial clearance subject to plan-specific usage support; that did not remove the need to check the then-current terms and project-specific rights.
A text-to-music model became a new audio surface[source 1]
ElevenLabs announced Eleven Music on August 5, 2025. The company said users could generate studio-grade music from natural-language prompts and listed control over genre, style and structure, a choice of vocals or instrumental output, multiple languages, and edits to sound and lyrics for a section or a whole song. It said the product was available on its website. The same sentence first referred readers to public API access, then said public API access and Conversational AI integration were coming soon; the source does not resolve that internal conflict. This record therefore establishes product availability and describes API wording as ambiguous, rather than asserting API availability at launch. The event is distinct from text-to-speech or transcription: the requested object is a musical work whose arrangement, lyrics, performance-like vocal output and structure can all matter to a downstream use. The announcement establishes the company’s product claim and listed controls. It does not prove that any prompt will produce a usable composition, that a named genre will be represented faithfully, or that generated lyrics will fit a particular campaign, performer, market or editorial requirement.
Prompting covered song structure as well as sound[source 1][source 2]
The launch’s feature list indicates a control surface broader than a simple background-audio generator. Genre and style concern musical character, while structure concerns how material is organized across a track. Vocals versus instrumental changes whether a user is asking for sung material at all. Section-level sound and lyric edits suggest that a revision can target part of a composition rather than force a full restart. ElevenLabs’ current music documentation similarly frames the product as a Text to Music model that understands natural language and musical terminology, and it lists genre, style, structure, vocal or instrumental output, multilingual generation and section edits. That later page helps interpret the work category, but it is current documentation. It cannot establish that every current setting, model version, duration, concurrency rule or editing behavior was part of the August 2025 release.
Multilingual generation did not eliminate review of words or delivery[source 1][source 2]
ElevenLabs named English, Spanish, German and Japanese among the multilingual options in the announcement. Language coverage is meaningful for music because lyrics, pronunciation, idiom, vocal delivery and cultural references can each affect a result. The source did not publish a language-by-language evaluation, a guarantee that a lyric will be grammatically or culturally appropriate, or a method for validating a generated vocal against a script. A music workflow still needs review of words, pronunciation, unintended references, musical transitions and the relationship between the requested and generated material. This is especially important when a product brief supplies regulated claims, names, contact information or a campaign-specific script. The launch makes a dated availability claim for a multilingual prompt-to-music offering. It does not turn generated output into reviewed copy, cleared performance, or a factual statement about a market without separate checks.
Broad commercial language still had plan and rights boundaries[source 1][source 2]
ElevenLabs said Eleven Music was created with labels, publishers and artists and was cleared for nearly all commercial uses, then directed readers to supported usage across plans. That wording is a material part of the release, but it should be read exactly. Nearly all is not every use, and a link to plan-specific support means account level can matter. The announcement did not publish a complete rights analysis for a user’s source prompt, lyrics, distribution territory, trademark use, performer reference or final edit. Nor does a general commercial-use statement establish that a particular generated song is safe for a particular media buy. The current music documentation is valuable for operational understanding, while rights and access need the applicable dated terms and project facts. The archive records a product launch and a supplier clearance statement, not a universal legal conclusion.
Sources
- Official Eleven Music release (August 5, 2025)
- ElevenLabs music documentation
Continue reading
