ElevenLabs voice cloning licensing covers four things that are easy to confuse: the technical right to create a voice on a plan, the legal right to clone a particular person, the rights a customer receives when using a voice that somebody else shared, and the arrangement under which a contributor shares a voice for rewards. A software asset manager sees them as entitlements (slots by plan), compliance obligations (consents and records) and third-party content rights (Voice Library use). The commercial overview is in ElevenLabs licensing, and the underlying contract terms are in ElevenLabs Terms of Service, commercial rights and prohibited use.
Editions
Voice features are bundled into the plan ladder. The pricing page lists voice cloning and voice slots as plan entitlements, and the cloning documentation sets the number of Professional Voice Clone slots as follows.[1][2]
| Plan | Instant Voice Cloning | Professional Voice Clone slots | Other voice allowance |
|---|---|---|---|
| Free plan | No | None | 3 custom voice slots |
| Starter plan | Yes | None | Commercial licence begins |
| Creator plan | Yes | 1 | Music streaming distribution begins |
| Pro plan | Yes | 1 | Extra Full Seats purchasable |
| Scale plan | Yes | 3 | 3 workspace seats |
| Business plan | Yes | 10 | 10 workspace seats |
| Enterprise plan | Yes | Custom | Custom |
The Professional Voice Cloning documentation adds that a Professional Voice Clone that already exists remains in the account but cannot be used below the Creator tier.[2] The price and tier details come from the plan table in ElevenLabs plans, credits and seats.
Metrics
A Professional Voice Clone slot is a slot in which a fine-tuned clone of the account holder’s own verified voice is held. A Custom voice slot is used when a voice is created with Voice Design or Voice Cloning. Voices saved from the Voice Library do not use custom voice slots.[3] The slot is the counted unit for what the customer stores. Generation with a voice is charged in ordinary credits, so voice choice does not change credit consumption except through the model and product used.
Counting / floors
Types of voice
Instant Voice Cloning. The customer uploads a sample and, before saving the voice, must confirm that it has the right and consent to clone it.[4] The confirmation is an attestation by the customer, and the Prohibited Use Policy sets the underlying rule: cloning, synthesising or impersonating the voice, identity or likeness of a living individual without explicit and informed consent or other valid authorisation is prohibited, and so is doing so to a deceased individual in a manner reasonably likely to cause harm or that infringes post-mortem rights.[5]
Professional Voice Cloning. A fine-tuned clone of a voice that goes through a verification process. The documentation states that a Professional Voice Clone may be made only of the creator’s own voice, even where another person consents, and that every clone goes through verification.[2] The licensing consequence is that an agency or a studio cannot buy one Professional Voice Clone slot and fill it with a client’s voice. The own-voice rule is stated for Professional Voice Cloning. Instant Voice Cloning instead asks the creator to confirm the right and consent to clone, and that confirmation puts the burden of proof on the customer.
Voice Design and library voices. Voices created with Voice Design use a custom voice slot. Voices saved from the Voice Library use none.[3]
Consents and records
The ElevenCreative Terms require the customer, where it submits Input containing the voice, image or other visual likeness of an identifiable natural person, to represent and warrant that it has given all notices and obtained all consents, releases and authorisations required by applicable law, including biometric-data, right-of-publicity and recording-consent laws. The customer must keep records sufficient to demonstrate those consents for the longer of the legal period or five years from the date obtained, and give ElevenLabs a copy of, or reasonable access to, them on reasonable written request where needed to respond to a claim, complaint or inquiry.[6] The Prohibited Use Policy adds that voice recordings of anyone under 18 may not be submitted to create or train a voice model, and that a voice or avatar may not use the likeness of a specific real minor.[5]
For an organisation, the practical control is a consent register: who was cloned, what authorisation exists, the scope of the permission (language, channel, duration, commercial use) and the retention date. Five years is a floor and the law of the speaker’s country can require more.
Rights in the resulting model
The customer’s licence to ElevenLabs extends to User Voice Models, to the extent the customer owns rights in them. It is perpetual, irrevocable, non-exclusive, royalty-free, worldwide and sub-licensable, and ElevenLabs states that it will not commercialise the voice on a standalone basis without the customer’s permission.[7] The customer may opt out of the use of its Content for training through the “Data use” setting.[7]
The Voice Library
The Voice Library is described as a marketplace where the community can share Professional Voice Clones and earn rewards when others use them. Only Professional Voice Clones can be shared. Instant Voice Clones and voices created with Voice Design cannot be, and Voice Library voices are not available through the API to free-tier users.[8]
Using a library voice
A customer who uses a shared voice does so on the customer’s own plan, and generation is subject to the Terms and the Prohibited Use Policy, so the general rules on commercial use apply: a free-plan user needs attribution and has no commercial licence, and a paid plan carries a commercial licence.[9] Library voices do not occupy a custom voice slot.[3] The Voice Library Addendum notes that others may generate Output with the shared voice and that ElevenLabs may charge fees for access to User Voice Models as set out in the Addendum and through the Services.[10] The addendum gives the contributor, not the library user, the Notice Period: if a contributor removes a voice, users who added it keep access for the Notice Period, and Output generated before the end of that period continues to exist and remain available for use.[10]
Sharing a voice
A contributor who shares a User Voice Model confirms several conditions. The contributor is not a resident of or located in the State of Illinois, and must tell ElevenLabs and remove the voice if the contributor becomes a resident or relocates there. The voice must be the contributor’s own, created and shared with ElevenLabs’ technology, and may not be another person’s voice or an altered version of the contributor’s voice.[10] The contributor must also hold appropriate rights and permissions and consent to the processing of voice data under the Privacy Policy and Terms.[10]
Rewards
Financial Rewards are available only for certain types of voice models, are payable in US dollars or, for users in the EEA, the UK and Switzerland, local currency, and require a Stripe Connect account. Participation and payment are at ElevenLabs’ sole discretion, and ElevenLabs may suspend or end it.[10] The default rate depends on factors ElevenLabs determines, such as the pricing tier of the users of the voice, the number of characters they generate and the Notice Period selected. In no case are Rewards provided for use of a voice by a Free User, and the earning rate for ElevenAgents is capped at 0.01 US dollars per minute.[10] The Notice Period may be increased but not reduced once the voice is shared.[10] ElevenLabs’ sole obligation for permitted use of a shared voice is to pay the Rewards earned, and the contributor otherwise waives royalties for that use.[10] The Prohibited Use Policy separately bars artificially inflating Voice Library rewards.[5] Together these form the Voice Library rewards programme.
Virtualization & partitioning
Not applicable. Voices live in a workspace, and slots are counted per workspace according to the plan. Moving a clone between unrelated workspaces is not a licensed operation, and the Terms bar sharing individual account credentials.[7]
Cloud / BYOL
Voice models are stored and run in the ElevenLabs cloud. Enterprise customers can choose data residency, described in the companion article on enterprise terms. Under the ElevenCreative Terms, Enterprise customers are not eligible for Community Sharing, so the sharing of generated Output with other users is not an Enterprise feature.[6]
Programs
- Voice Library rewards. Contributors share a Professional Voice Clone and earn rewards when paid users generate with it.[10]
- Notice Period. A contributor-set period during which existing users keep access after removal. It can be extended and cannot be shortened.[10]
- Live Moderation. A contributor setting under which tools check requests for prohibited text categories. It is not a guarantee and applies only to users who add the voice after the setting is changed.[10]
- Reader App and ElevenAgents rewards. Rewards can also arise when users select the voice in the ElevenReader app, per 1,000 characters, or in ElevenAgents conversations, per minute.[10]
Out of scope
- Biometric, publicity and recording-consent laws in each jurisdiction, which govern the customer’s own consents.
- Contracts between a customer and the speaker, such as talent agreements and union terms.
- Enterprise custom voice agreements, which are not public.
- Litigation about voice rights. No court case about ElevenLabs software licensing or licence scope was identified with a primary court record, so no case article is included.