SoundCloud and ElevenLabs Turn AI Music Publishing Into a Controlled Release Path
SoundCloud is partnering with ElevenLabs to let users publish music generated in ElevenLabs directly to SoundCloud. The November 5 update removes the familiar export, download and re-upload routine, and tracks sent through the integration will carry a label identifying their AI origin.
That is a small workflow change with a larger consequence: creation and release are being joined inside the same company-controlled path. AI music is no longer only a file made in one place and uploaded somewhere else. The release handoff itself is becoming part of the product.
A label is useful, but it is not the whole record
AI-origin labeling gives listeners a clearer signal about how a track entered SoundCloud. It also shows that disclosure can be built into publishing rather than left entirely to a manual choice after a file is exported.
But AI-era rights questions do not end with a label. A release may need context around who is connected to the work, what permissions apply, what inputs were involved, and which version was published. As generative tools move closer to distribution, those questions arrive earlier—at creation and release, not only after a dispute or takedown.
The convenience comes with a control boundary
Direct publishing makes ElevenLabs an input to SoundCloud’s creator workflow. SoundCloud’s earlier Moises integration similarly connected stem separation, remixing and republishing.
The larger shift is that more of a creator’s operation—creation, publishing and the release signal attached to a work—sits inside SoundCloud’s environment. The creator gains speed, but SoundCloud controls the infrastructure, rules and relationships through which that release reaches the public. The question is not whether a centralized service can make this easier. It is why the work’s underlying record and release context should have to begin inside that service.
Creator-operated records change the starting point
When AI-assisted music moves quickly from prompt to public release, a creator needs somewhere they operate to keep the work and its release context connected from the start. Without that infrastructure, the release path, public disclosure and the records surrounding a work can remain fragmented across tools controlled by other companies.
Certifyd Core takes the opposite architectural direction. Currently in beta, Core runs locally with SQLite-backed work and state storage. Its schema includes work, media, profile and split records, while its publication helpers compute canonical manifest hashes and receipt bindings. Those capabilities give a creator-operated Core a place to originate and maintain records around a work rather than leaving the publishing service as the sole operating environment for them.
Core also separates private dashboard and API routes from an allowlisted public creator server, with consent built into its public-sharing lifecycle. That does not make Core a replacement for SoundCloud’s current publishing function. It establishes a different control relationship: a creator can operate part of the infrastructure carrying work and publication records, while Certifyd is building toward a network that moves supported creator functions away from centralized intermediaries.
The next release workflow needs more than speed
SoundCloud and ElevenLabs are making AI music publishing faster, and the AI-origin label is a meaningful part of that handoff. Yet as creation tools, release tools and audience systems become one connected workflow, ownership of the operating environment becomes harder to ignore.
For creators, partners and investors, the opportunity is not simply smoother AI uploads. It is creator-operated infrastructure that keeps the work and its release record from existing only inside a single intermediary’s system. In the AI era, clearer rights decisions will require usable context from the moment a work is created—not disclosure added after the fact.
Originally published on Certifyd.
Top comments (1)
Good point that a label is a signal, not a record. Once creation and release share one pipeline, the cheap thing to keep is a signed note of which version was published and with what inputs, written at release time. Rebuilding that after a takedown request is where it gets expensive.