Skip to content

AI Content Disclosure

Last updated 12 October 2026

Every track HeySong renders is synthetic audio produced by a third-party generative model from a prompt you supply. This page says which models, who owns the result, and exactly where our providers' own terms limit what you may do with it.

1.What you are listening to

HeySong does not record, sample or license a catalogue. Nothing you download is a performance by a human musician, and nothing is a copy of a recording we licensed. Your prompt is sent to a generative model, which synthesises new audio from it.

Because the audio is generated, two tracks from the same prompt will not be identical, and the same prompt can produce materially different output at different times. That is a property of the technology, not a defect we can engineer away.

2.Who makes the audio

  • Google (Gemini) serves several of our engines, including the vocal and lyric models.
  • ElevenLabs serves several of our engines, including the music and voice models.

The engine you pick is the engine that receives your prompt. We do not send your prompt to every provider; it goes to the one you selected.

When Bring Your Own Key is active, your prompt is sent to the provider using your credentials and is governed by your own agreement with that provider.

3.Who owns it

You retain the rights you hold in your prompts, your lyrics and the audio you generate. We claim no ownership of your content. Our providers agree to the same position between themselves and their customers: ElevenLabs states that you retain all rights in its output, and Google states that it will not claim ownership over content its services generate.

That is a contractual position, not a copyright registration. A generative model cannot assign you a copyright, and no provider warrants that one will vest in any given output. Whether a particular output is protected is decided by the law of your jurisdiction and judged on its facts. If your use depends on copyright vesting: a client work-for-hire deliverable, a library sync, a catalogue registration. Clear it before you ship.

4.What our providers permit, and where that limits you

We are a reseller of our providers’ services. Their terms bind our use of them, and where they are narrower than our own marketing, they win. Two limits matter enough to state here rather than in a footnote.

Age. Google's Gemini API terms require that its APIs not be used in a service “directed towards or likely to be accessed by individuals under the age of 18”. This constrains who may generate on HeySong: see the eligibility clause in our Terms.

Vendor confirmation required: age threshold

Our Terms currently set the minimum account age at 13, which is the threshold we chose for the product. Google's Gemini API terms prohibit serving a service likely to be accessed by under-18s through those APIs. Until the owner rules on whether HeySong is age-gated at 18, or which engines an under-18 account may use, this clause states the conflict rather than resolving it. This must be settled before launch.

Commercial use. ElevenLabs distinguishes free from paid use: on a free plan its output is for non-commercial purposes only, and commercial use requires a paid plan. Its music terms go further, and prohibit using generated music for streaming media content, film or television, broadcast, advertising and commercial video games on self-serve plans without ElevenLabs' prior written approval.

Vendor confirmation required: commercial scope of ElevenLabs-served engines

Our Terms grant a commercial licence that includes monetised video and streaming distribution. That grant covers HeySong's own contribution and is accurate for the engines we can render ourselves. It cannot extend further than our provider's own rights allow, and for ElevenLabs-served engines that scope is narrower on self-serve plans. The owner must confirm which ElevenLabs plan and which written agreement HeySong holds before this licence is published as unconditional. Until then, read the commercial clause in our Terms together with this page; where they differ, your provider agreement governs your use.

5.Labelling and disclosure

Some jurisdictions and some platforms require synthetic media to be disclosed as synthetic. Where the law requires it, we label it; where a distributor requires it, that obligation is yours to meet on upload. Do not present generated audio as a recording of a real performance, and do not use it to misrepresent a person's words.

Impersonating a living artist or performer is prohibited outright: see the Acceptable Use Policy. We screen prompts before dispatch and rewrite or refuse requests that target a named artist or copyrighted lyric; a refusal does not consume credits.

6.Training

HeySong does not train generative models on your prompts or your generated audio. We have not retained any right to do so.

Our providers apply their own rules, which differ by plan and by region. Google states that on paid Gemini services it does not use your prompts or responses to improve its products, and applies the same protections to unpaid quota for users in the EEA, Switzerland and the United Kingdom; on other unpaid usage the rules Google publishes for unpaid services apply. If that matters to you, read our privacy policy's processor section or bring your own key, which moves that decision to your own account.