Biometric Data Policy
Version 1.0 — Effective July 1, 2026
Our written retention and destruction schedule for voice models.
Draft v1.0 — the legal entity name is a placeholder pending finalization before public launch.
1. Why this page exists
Creating a persona involves building a computer model of a person's voice — a "voice model" or "voiceprint." Some laws, including the Illinois Biometric Information Privacy Act (BIPA), treat a voiceprint as biometric information and require any company holding it to publish a written schedule saying how long it is kept and when it is destroyed. This page is that schedule. It applies to everyone, not only residents of states with such a law.
2. What we treat as biometric data
Voice models. The synthetic voice built from submitted recordings and used to speak a persona's replies. This is the biometric identifier.
Source recordings. The audio submitted to build or improve a voice model.
We do not use voice data to identify anyone. We never run speaker identification, voice matching, or diarization, and we do not use a voice model to verify identity or to unlock anything.
3. Retention schedule
Voice models are destroyed at the earliest of these events:
On request. When you delete a persona, delete your account, or ask us to remove a voice, we destroy the voice model within 30 days and instruct our voice provider to delete their copy. Account deletion has a 30-day grace period during which you can restore it; destruction happens when that period ends.
When the purpose ends. If a persona is deleted or a voice is replaced, the old model is destroyed — we do not keep superseded models.
After 3 years of inactivity. If nobody has interacted with a persona for 3 years, its voice model is destroyed automatically, whether or not anyone asks. A scheduled job runs daily and enforces this. The persona's written stories, photos, and timeline are not deleted by this rule — only the voice model.
Source recordings are deleted with the persona or account they belong to, on the same schedule.
4. How destruction works
Voice models are hosted by our voice provider, ElevenLabs. Destroying a model means deleting it from our systems and issuing a deletion request to that provider for their copy. Deletion is permanent and cannot be reversed. Limited copies may persist briefly in encrypted backups, which expire on their own cycle, and we may retain records we are legally required to keep.
5. Consent
We ask for a voice recording only from someone who tells us they have the right to provide it. If you are recording another living person, you must have their permission before you submit their voice, including any written consent your state requires for biometric data.
If your voice was used to build a persona and you did not agree to it, contact legal@heylegado.com. We will act on valid requests to destroy the voice model and remove the persona.
6. Sharing
We do not sell, lease, or trade voice models or biometric data, and we do not profit from them apart from providing the Service you signed up for. The only third party that receives voice data is our voice provider, for the sole purpose of creating and speaking the persona's voice.
7. Your choices
You can delete a persona, replace a voice, download your data, or delete your account at any time from Settings, or by writing to legal@heylegado.com. Deleting a persona or account destroys the associated voice model on the schedule in Section 3.
8. Contact
[Company Legal Name] — legal@heylegado.com.