Voice consent

How we handle your voice

What we make

When you record a sample, we create a temporary voice model so the demo can speak your translated words in your own voice.

How long it lives

• The voice model is deleted within one hour of being created.
• The audio clip and its translated text are kept so the share link keeps working. Ask us and we will delete them.

Who can hear it

Nothing is public until you press Share clip. A shared clip is reachable by anyone holding the link.

What you promise

• The voice you record is your own.
• You will not use Mimir to impersonate another person or to record someone without their knowledge.

Who processes it

Voice cloning and speech synthesis run on ElevenLabs. Translation runs on Google Gemini. Audio files are stored on Vercel. We do not sell your recordings.