Ethics, consent and takedowns

Creating a custom voice requires two things: a typed attestation that the voice is yours or that you hold written permission from the speaker, and a consent phrase naming you and this account. We store the phrase text, the audio, and a timestamp alongside the voice.

If consent is skipped or incomplete, the voice is not created. There is no way to shortcut this in the product.

• No celebrity, politician or public-figure voice gallery.
• No public marketplace of other people's cloned voices.
• No singing covers, no video dubbing of third-party footage.
• No custom voice uploads on the free Demo tier.

• Impersonating a real person to deceive, defraud or extract money or credentials.
• Non-consensual intimate audio of any kind.
• Political deepfakes, fabricated statements from officials, election interference.
• Bypassing voice-based identity verification.
• Harassment, defamation or synthetic evidence.

Accounts found doing any of the above are suspended and their voices removed. Flagged generations are reviewed by a human.

Demo and Starter output carries a watermark flag in the file metadata. Demo output also begins with a short audible marker announcing that it was generated with Tonekeep Demo. Removing or obscuring these markers is a breach of the terms.

If you believe a Tonekeep voice was created from your speech without your permission, email takedown@tonekeep.app with a description of the voice and any audio you have. We disable the voice while we investigate — access first, questions after.

We store your account details, your reference recordings, your consent records, the scripts you generate and the resulting audio. Reference audio is used only to create and run your own voices. We do not train shared models on your recordings and we do not sell or share your audio.

Deleting your account removes your voices, reference audio and generated files. Billing records are kept as long as tax law requires.

Plans are billed monthly in USD through Square and can be cancelled at any time; access continues until the end of the paid period. You are responsible for holding the rights to every voice you clone and for how you use the audio you generate.

This is an early MVP. Limits, engines and export formats may change, and we will tell you before anything you rely on is removed.