Timastra Voice · Development pilot
Timastra Voice — Privacy notice
Timastra Voice is provided by Timastra LLC. This notice covers the Android development pilot and its voice-processing service. General registration, public voice enrollment and paid purchases are not open. A published notice does not give permission to use another person's voice.
Last updated: .
Who is responsible
Timastra LLC, 5830 E 2nd St, Ste 7000 #38687, Casper, WY 82609, US, is responsible for the Voice information described here. Contact support@timastra.com for privacy questions or to exercise your rights. The website and shared Timastra sign-in service have separate privacy information linked below.
Account and device information
Google sign-in goes through the existing Timastra sign-in service in your browser. Voice receives an account identifier and email address; it does not receive your Google password. The Android app keeps sign-in tokens and its creator or speaker-access choice in protected device storage. Microphone permission is used when you choose to record. Android app backup and device transfer exclude app-owned data; files you intentionally share are separate copies.
Recordings and voice checks
The service receives enrollment and spoken-permission recordings, transcribes them and checks audio quality and speaker consistency. It stores recordings, transcripts, quality results and a model representation used to generate speech. Speaker comparisons can be wrong: they do not establish identity, age, liveness or a guardian's authority. OpenVoice, transcription and speaker-comparison models run in the operator's service environment; recordings are not sent to those models' publishers for inference. The pilot uses an operator-controlled local service and GPU; no public Voice hosting region is represented as already deployed.
Family, clients and children
The account holder submits permission evidence for a named adult speaker or a child aged 6–17 and their guardian. Records can include names, birth dates, declared country or region, adult contact details, relationship, requested use and expiry, documents, privacy-notice choices and spoken evidence. The holder must have authority to submit this information. Children do not need their own account. A child's authorized use is personal and does not include a downloadable voice model. No real-country enrollment is currently enabled; a country selection or document upload alone does not authorize a voice.
Generation, sharing and purpose
Voice processes the text you submit, the selected voice and usage information to generate audio, deliver results, enforce permission and allowance limits, and investigate errors or abuse. An 80-character text preview and text hash can remain with the job record after its full text is cleared. The service keeps consent and security history, usage and recovery records, and reports you submit for these purposes. It does not use your recordings to train a general-purpose voice model. Generated audio is synthetic. Anyone receiving a downloaded recording or permitted export may keep or redistribute a copy; withdrawal cannot recall those copies.
Providers and access
Authorized Timastra operators can inspect submitted permission evidence and reports when reviewing permission or abuse. Private database and object storage hold Voice records; temporary download links expire but a recipient can retain a downloaded file. Google and the shared Timastra sign-in service process authentication. Support uses the existing Timastra support services. Their provider and location information is in the linked website notice. No Voice checkout, SMS notices or external error-reporting destination is enabled in the current pilot. If purchases are later offered, Google Play will process payment and Voice will receive purchase-verification records, not your complete payment-card details. New hosting or data-sharing arrangements must be disclosed before they apply to participants.
Retention and deletion
Accepted enrollment recordings and voice representations remain until the voice is deleted; rejected previews and enrollments abandoned for 30 days become eligible for cleanup. Generated audio expires after 30 days and prepared exports after two days, or earlier when access is withdrawn or data is deleted. Full generation text is cleared on completion or cancellation; failed-job text is kept for seven days for retry. Requests without a current protected grant become eligible for inactivity cleanup after 30 days. Cleanup retries failures, so an accepted deletion request or expired link does not prove that every storage copy has already been erased. Copies you downloaded or shared need separate deletion.
Records that remain
Deleting an account removes access and queues removal of its voice content and email. Pseudonymous identity-deletion records remain to reject old sessions. Permission, security, abuse and financial records needed for disputes, legal obligations, safe recovery or fraud prevention can remain after content deletion; an actual legal hold can delay evidence removal. The pilot has no automatic final purge period for these retained records or verified backup-erasure deadline. Contact support for the applicable record and reason. We do not promise immediate erasure from every backup or third-party copy.
Your choices and rights
You can request access, correction, export, deletion or information about retained records through support. Account export and account deletion require a fresh sign-in. A named speaker or guardian can use My voice consent with their own confirmed contact to inspect or withdraw a request, without controlling the creator's account. Withdrawal stops future authorized use; it does not undo earlier processing or erase copies already downloaded. You can revoke microphone access in Android settings. We verify authority before disclosing or deleting private information, and explain any exception. You may also contact your applicable data-protection authority.
Read the website and shared sign-in privacy notice, general deletion instructions and company details.