Sign in Start free

Acceptable use and voice consent

The lines that get an account closed - and the consent rule that makes voice cloning legitimate rather than a lawsuit.

Last updated 14 August 2026 · plain English, and short on purpose.

Not allowed

  • Impersonating a real person to deceive - fake endorsements, fake statements by public figures, voice-based fraud, "grandparent" scams, defeating voice authentication.
  • Sexual content involving real people's voices, or any sexual content involving minors, ever.
  • Harassment, threats, hate speech directed at a person or a protected group.
  • Election material presented as a real person's speech, and any synthetic political audio that hides that it is synthetic.
  • Medical, legal or financial instructions presented as coming from a professional who did not say them.
  • Removing or defeating the watermark and the synthetic-media disclosure, or stripping provenance from exports.
  • Reselling raw engine access outside the reseller plan, scraping the catalogue, or working around metering.

Disclosure and watermarking

Every generated file carries an inaudible AudioSeal watermark and a disclosure header saying it is synthetic, and exports keep their provenance record. That is a compliance feature and it stays on. Audio you upload for cleanup is your own recording and comes back unmarked.

Where you publish synthetic speech that a listener could mistake for a real person, say so. The EU AI Act requires it from August 2026 and platforms increasingly require it too.

If you are the person in the recording

Write to hello@naadly.com with a link or a copy of the audio. A watermarked file can be traced back to the account that made it. We remove the voice, keep the consent record as evidence and close the account where consent was faked - and we do this before asking you for anything notarised.

How we enforce

Proportionately: a warning for a mistake, an immediate close for impersonation or sexual content involving a real voice. Any generated audio can be checked against its watermark, and administrator actions are written to an audit log. There is no automated content scanning of your scripts - we act on reports and on evidence.

Questions about any of this go to hello@naadly.com and are answered by the person who wrote it.