Your Voice, Your Data
Your voice is the most personal thing you could hand us. This page explains — in plain language — who we are, what happens to your recording, and how we protect it.
Last updated: 12 September 2026
Who we are
Classics in Your Voice (voiceami.com) is operated by 33Ventures Pty Ltd, an Australian company — ACN 701 999 304, trading as Voiceami. You can verify our registration on the Australian Government's ASIC companies register. We are bound by the Privacy Act 1988 (Cth) and the Australian Privacy Principles, and a voice recording used to create an AI voice model is treated as biometric information — the most protected category of personal information under that law. The full legal detail lives in our Privacy Policy; this page is the plain-language version.
The journey of your recording
- You record in your browser — only after you tick an express consent box. The Record button stays off until you do, so nothing is captured before that tick.
- It travels encrypted — every connection to us uses HTTPS, enforced with HSTS.
- We create your voice simulation on servers we control — our own speech model runs on GPU servers we rent from RunPod. Your recording is never handed to a third-party voice-AI company, and it is never used to train any AI model.
- A person may listen first — before narration begins, a member of our team may play your recording to check that it is clear and matches the text you read. That is the only reason anyone on our team hears it.
- Your recordings are delivered privately — generated audio is stored in a private bucket (Cloudflare R2), is never publicly listable, and is served only through per-customer, token-protected links. Your listening page stays live for 180 days.
- Then it is deleted — order recordings are erased automatically about 7 months after the order, and the erasure date is printed in your order emails. A voice you chose to save to your account is kept until you delete it — which you can do yourself, any time, on your account page.
How we protect it
- Encrypted in transit — HTTPS everywhere, with HSTS so browsers refuse insecure connections.
- Private storage — voice files and generated audio are never publicly listable; every download link is tied to your order, and removing someone you shared a book with re-issues every link at once.
- Encrypted backups — our off-site backups are encrypted with a key that is kept offline, away from our servers. A stolen backup is unreadable.
- Delete means delete — backups included — when a recording is erased, its backup copies are queued for erasure too and purged after a 7-day undo window. Nothing lingers in a backup indefinitely.
- Restricted, logged staff access — staff sign in with expiring one-time links (no shared passwords) and see only the areas their role needs. A support view of your account lasts at most 30 minutes, is read-only by construction — it cannot delete, replace or re-record your voice — and is permanently logged.
- A consent audit record — we store exactly which consent wording you agreed to, when, and in which language, so your consent is verifiable, not assumed.
- Every sentence is checked against your voice — as each sentence is generated it is compared with your recording by a speaker-verification check, and a sentence that drifts toward a generic voice is generated again.
- We do not record what you listen to — which chapters you play, and how far you got, stays on your own device. Our usage statistics are anonymous counts with no name, account or order attached.
- Payments by Stripe — your card details go directly to Stripe and never touch our servers.
What we will never do
- Use your voice to train AI models.
- Sell or share your recording or voice model with anyone. The providers who host and deliver the service for us are named in the Privacy Policy and process it only on our instructions.
- Generate anything in your voice that you did not order.
- Let someone else order a book in your voice — even someone holding a link to your free chapter must be you (or sign in as you) to order with your recording.
- Publish a sample of your voice without your separate, explicit consent — and any sample you do choose to share lives at an unlisted link, expires after 30 days, and can be removed by you at any time.
- Ask for your phone number or send you a text message. We ask for an email address and nothing more.
Your controls
- Delete your saved voice yourself — one button on your account page. This also queues our backup copies for erasure.
- Ask us to delete everything — use the contact form and we will respond within 30 days, as our Privacy Policy commits.
- Complain — if you believe we have mishandled your information, contact us first; you can also complain to the Office of the Australian Information Commissioner (OAIC). Details are in the Privacy Policy.
Report a voice you believe was misused
If you believe a voice was simulated without its owner's consent, or content on this service infringes your rights, tell us here. We investigate every report and take down confirmed misuse. You do not need to be a customer.
For security researchers
We welcome good-faith security reports. Our disclosure details are published at /.well-known/security.txt; please report anything you find through the contact form and we will respond promptly. Please don't access other people's data while testing — a proof of concept against your own account is always enough.