Legal

Privacy notice

A voice is personal data. This page says what we hold, why we hold it, and what a speaker can ask us to do about it.

Section 01

What we hold about speakers

  • Name or stage name, and any notes recorded during onboarding.
  • Languages spoken and whether the speaker reads and writes Yorùbá.
  • Audio recordings and their transcripts, stored with tone marks intact.
  • Consent status, scope, dates, expiry and the signed consent document, including the timestamp of the electronic signature.
  • Session details such as date, location, equipment and audio settings.

We hold this to run the registry, to prove that consent exists, and to build datasets only from material that is consented and verified.

Section 02

What we hold about customers

Account email and name, plan and subscription status, access pass usage, credit consumption, generation provenance records, and support messages. Card details are handled by our payment processor, not by us.

Section 03

Who can see it

Registry staff who verify clips and manage consent, and the service providers that host the database and file storage, process payments, and run the speech and language models used for narration, cloning and transcription.

We do not sell personal data, and we do not hand recordings to third parties outside the consent scope the speaker signed.

Section 04

Support chat and the assistant

Messages you send in the chat widget or the contact form are stored so the team can reply. An assistant may answer first using published facts about plans and consent, and a person can take the thread over at any point. Do not put payment details in chat.

Section 05

How long we keep things

Recordings, transcripts and consent documents are kept while the corpus is in use, because the consent record has to outlive the recording it authorises. Rejected clips are kept but excluded from exports. Support threads and provenance records are kept for accountability. Ask us and we will tell you what exists about you.

Section 06

Withdrawing consent

A speaker can withdraw at any time through the contact page. On withdrawal the speaker is flagged everywhere they appear and their clips are excluded from every future export. Datasets already delivered are recalled by written request to the holder.

Section 07

Your requests

You can ask for a copy of what we hold about you, ask us to correct it, or ask us to delete it where we are not required to keep it. Write to us and say which speaker or account you mean.

This notice describes current practice and is updated when practice changes. It is not legal advice.