Docs / Voice

Cloning your voice

Train a clone of your own voice so your lens speaks like you, or pick a ready-made voice from a curated list.

A lens can read its answers out loud and take voice calls. By default it uses a clean platform voice. Train a voice clone and it speaks in your voice instead; combined with the “as me” point of view, the lens both sounds and answers like you. Speech-to-text and text-to-speech both run through ElevenLabs, so visitors hear the same voice in every browser, with no device speech engines in the loop.

Setting it up

Open your lens’s page and find the Voice card, then upload a sample of yourself speaking. Mindola trains a clone from that sample, and from then on every spoken reply on that lens uses your voice.

Some plans include a pro mode that trains on several longer clips for a closer match. What your plan includes is shown in the app.

Before the Record button unlocks, you see a one-time consent dialog with three required confirmations:

  1. This is my own voice. You are the person speaking; you are not cloning someone else.
  2. I consent to biometric processing. Explicit consent under GDPR Article 9(2)(a) to process your voice sample and send it to the voice provider (ElevenLabs) to derive the clone.
  3. I understand how it will be used and how to revoke. The clone speaks lens replies, you can delete it at any time from the Voice settings, and Mindola will not use it for anything else.

Tick all three and the recorder unlocks. Mindola emails a signed PDF record of your consent to your account email; keep it for your files. The consent version is stored on your account, so you only see this dialog once unless the legal text changes.

Revoking the clone

Delete the clone from the Voice settings whenever you want. The voice model is removed from ElevenLabs within 30 days. If you have questions about consent or processing, reply to the consent email or write to security@mindola.ai.

Ready-made voices

If you would rather not clone your own voice, or you are building a lens for someone who cannot record one, pick a ready-made voice from the curated list in the same Voice card. Each voice has a short description of its accent and tone, and the one you pick is what visitors hear in spoken replies and calls on that lens.

Picking a ready-made voice replaces a recorded clone if you had one, and you can switch back to the Mindola default at any time. Ready-made voices need no consent step, since nothing of your own voice is recorded.

Availability

Voice cloning is part of higher plans; ready-made voices are available on every plan. The app shows exactly what your plan includes.

Good to know

  • Only upload your own voice. It is your likeness; keep it that way.
  • The clone is tied to the lens. Other lenses keep their own voice settings.
  • A lens with a cloned voice shows visitors an extra notice that the voice is AI-cloned. See AI disclosure.