Acceptable use policy

1. Acceptable use

You may not use this Service to:

• Clone the voice of any person without their explicit, documented consent.
• Impersonate any real person, public figure, or organisation.
• Generate content that is illegal, harassing, defamatory, or deceptive.
• Generate content intended to defraud, mislead, or manipulate elections.
• Circumvent rate limits, usage caps, or access controls.
• Use the Service in a way that could harm minors.
• Resell or redistribute access without prior written agreement.

We reserve the right to suspend or terminate accounts that violate these terms and to cooperate with law-enforcement requests where legally required.

2. Voice cloning & biometric data

If you upload a reference voice to clone, you represent and warrant that either the voice is your own, or you have obtained the express written consent of the voice owner for this use. You must retain proof of such consent for the duration of your use plus two years. We may request proof of consent at any time. Content found to violate this clause will be removed and the originating account may be terminated.

Biometric data. Voice recordings may qualify as biometric information under the Illinois Biometric Information Privacy Act (BIPA), the Texas Capture or Use of Biometric Identifier Act (CUBI), Washington HB 1493, and comparable laws. By uploading a reference voice, you affirm that you have given or obtained the written release required by those laws from every living person whose voice appears in the recording. See the Biometric Information Notice in our Privacy Policy for the retention schedule.

Indemnity. You agree to indemnify and hold us harmless from any claim arising from voice material you upload, including claims under BIPA, CUBI, the Tennessee ELVIS Act, California Civil Code §§3344/3344.1 (AB 2602 and AB 1836), the right of publicity, or similar statutes.

3. Prohibited uses — deepfakes, impersonation, robocalls

You may not use the Service to:

• Impersonate any real person, living or deceased, without their (or their estate's) express authorization, including political candidates and officials — regardless of consent.
• Create digital replicas of a performer's voice without a written agreement that satisfies California AB 2602's description and representation requirements.
• Create synthetic audio of a deceased personality in violation of California AB 1836 or equivalent post-mortem publicity statutes.
• Conduct robocalls, outbound telemarketing, or any call using AI-generated voices without prior express written consent from every called party, as required by the FCC's February 2024 TCPA ruling.
• Evade voice-verification, KYC, or authentication systems.
• Use the Service to influence or mislead voters, or to impersonate government, election, or emergency officials.
• Train competing speech-synthesis or speech-recognition models on the Service's outputs, embeddings, or model behavior.
• Remove, alter, or suppress any watermark, provenance metadata, or C2PA Content Credential applied to Service outputs.

4. AI-disclosure obligations (EU AI Act Art. 50)

If you deploy Service outputs as part of an AI system that interacts with natural persons (e.g. voice agents, interactive voice response, automated callers), you must clearly disclose to those persons that they are interacting with AI unless it is obvious from context. For AI-generated audio distributed publicly, you must clearly label the content as AI-generated. We may apply machine-readable markers (watermarks or C2PA Content Credentials) to Service outputs to support these obligations.

5. Reporting violations

If you believe content generated with our Service violates this policy, report it to our abuse team. Include the date, the affected party (if known), and any evidence you can share. We review every report and respond as quickly as possible.