What you can do with what you generate
The short version: you own the audio, we do not, and the thing that changes with the plan is whether you may sell it. The longer version is the table — and every row points at the document that actually governs it, because a rights answer you cannot trace is not an answer.
| DIMENSION | Without a subscription | With VoiceLabs Pro |
|---|---|---|
| Who owns the audio | You do. We claim no ownership of the audio you upload or the audio you generate, on any plan (Terms §10). | You do — the same answer. We take only a limited licence to host, process and store your inputs and outputs in order to run the service for you. |
| Downloading the file | Included. There is no preview-only mode, no watermark on the audio you hear, and no upgrade required to obtain the file. | Included, on the same terms. |
| Personal and non-commercial use | Yes — hobby projects, learning, accessibility tooling for yourself, prototypes you are not selling. | Yes. |
| Commercial use | No. Selling the audio, putting it in client work, monetised video, advertising, or a product you charge for is what the paid plan is for. | Yes — commercial rights to everything you generate are part of VoiceLabs Pro. This is the industry-standard gate: ElevenLabs, Murf, Speechify Studio and Cartesia all place commercial use behind payment too. |
| Saved cloned voices | One saved custom voice. Cloning itself is not withheld — the capability is real on every account, which is deliberately unlike most of this market. | Unlimited saved cloned voices. |
| API and agent access | No. The HTTP API and the MCP server are part of the paid plan. | Yes — API keys with per-key scopes, and the remote MCP server for AI agents, on the same subscription. |
| Rights in the VOICE itself | None of this grants you rights in a person's voice. Owning a generated file is not the same as being allowed to imitate someone — that permission comes from the speaker, and it is required on every plan. | Identical. Paying more does not buy anyone else's voice. |
The part people get wrong
Owning the file and being allowed to use the voice are two different questions, and only the first one is about your plan. Whatever you pay us, you still need the right to the voice you cloned: your own, the speaker’s explicit permission, or material that is otherwise legally authorised. There is no tier that changes that, because it is not ours to sell — it belongs to the person whose voice it is. Our Responsible Use policy sets out what that means in practice, and you accept it before you use the studio.
Disclosure is not optional either
Audio made here is synthetic, and where law or a platform’s own policy requires it, you have to say so. To support that, generated audio carries an inaudible, machine-readable AI watermark, consistent with EU AI Act Article 50 on synthetic-media disclosure. The watermark is a marker, not a licence — it does not disclose anything on your behalf.
Where each answer comes from
- Ownership and the licence you give us — Terms of Service, sections 10 and 11.
- Consent, prohibited uses and disclosure — Responsible Use policy.
- What each plan includes — Pricing.
- How a misuse report is handled, and how to appeal one — Moderation and appeals.
VoiceLabs Pro is $8 a month, with a 7-day free trial, and carries commercial rights to everything you generate.