Voice changes with age, illness, and cognitive decline. The window for capturing voice at full strength is finite. Heirloom partners with ElevenLabs for high-fidelity voice cloning (with explicit recorded consent). The voice clone is created from 30 minutes of clear audio captured during normal interview sessions. Future generations can interact with the voice — even decades after the original speaker is gone.
Frequently Asked Questions
How much voice audio is needed for a clone?
30 minutes of clear, consented audio. Captured naturally during weekly interview sessions over 4-6 weeks.
What if voice has already changed?
We use the best available recordings. If pre-illness recordings exist (voicemails, family videos), those produce higher-fidelity clones than current-state audio.
Voice clone ethics — what's Heirloom's policy?
Explicit, recorded, revocable consent in the speaker's own voice — captured before clone creation. Heirloom never trains external models on customer voice data. Full revocation at any time.
Can the clone be used for new content?
Only with prior recorded consent. The clone speaks the words our retrieval system grounds in actual captured material — never invents new statements the speaker didn't make.
Give them a book of their life — in their own voice.
Family Council captures the whole family's voices in concert. Try the free demo at theheirloom.ai/try first — no account, no card. First session within 24 hours. The gift that future generations will thank you for.
Give them the Keepsake →