Apps HubTim Wu Haotian
KinVoice app icon

KinVoice

Preserve a loved one's voice and hear their stories again.

PubliciOS
View on App Store (opens in a new tab)

KinVoice Privacy Policy

Last updated: 2026-09-27 · Legal owner: WU HAOTIAN · Support: gosingk@gmail.com

Information Used

KinVoice may use voice recordings of a family member, captured with their consent (the raw reference sample is stored only in the app's private storage on your device and re-sent to the cloud voice service on each synthesis — see Third Parties), the questions you speak during voice chat (transmitted to the cloud AI service for speech-to-text and not retained on our servers), your chat messages and the family member's name and relationship you provide as context (sent to the cloud AI service to generate replies and stories), the words you type or paste for a line you want to hear in the cloned voice (sent to the cloud AI service to screen them for scam and abuse patterns and then to synthesize them; not retained on our servers — the kept line lives only on your device), your Apple account identifier and email address from Sign in with Apple (the email may be a private Apple relay address; used to create and identify your account), voice-profile metadata you create — family member name, title, generated voice label, and recording duration (stored on our server so your cloned voices list across devices), usage counters and subscription state — your monthly message count, how many lines were voiced or declined (counts only, never their words), plan tier, and Apple transaction identifiers — managed through Apple In-App Purchase and stored on our server to provide the app features described on this page.

Storage

KinVoice keeps your raw voice recordings — the reference audio used for cloning — in the app's private storage on your device. That on-device audio is the single source of truth: it is re-sent to the cloud voice service each time speech is generated, rather than stored on our servers. Our backend (a SQLite database on a private server) stores only your account (Apple user identifier, the email returned by Sign in with Apple, and an optional display name), voice-profile metadata (family member name, title, generated voice label, and duration), usage counters (including counts of lines voiced and declined, never their words), and Apple subscription state. Lines you ask a cloned voice to say, their audio, and scheduled voice notes are kept only on your device. We do not retain raw or generated audio on our servers beyond the moment needed to relay a single request to the cloud voice provider. You can delete your account and all server-side data from within the app at any time, which removes your account, voice-profile metadata, usage, and subscription records from our server; deleting the app removes the on-device library. Standard iOS device backups may include the on-device library depending on your backup settings.

Permissions

Microphone access records the voice samples you capture and the questions you speak during voice chat. Sign in with Apple creates and identifies your account; the email may be a private relay address you can disable in your Apple ID settings. Network access is required to clone voices, transcribe speech, generate chat replies, stories, and voice notes, screen and voice lines you write, synthesize cloned-voice audio, and validate subscriptions. Notifications (optional, asked when you turn on voice notes) deliver scheduled voice notes; the lock screen shows only that a note arrived, never its words.

Sharing

KinVoice does not sell your data. Your recordings, voice library, and chat content are never used for advertising, and the app contains no analytics or tracking SDK. To clone a voice, transcribe what you say, generate a reply, story, or voice note, screen the words you ask a cloned voice to say, and speak it back in the cloned voice, KinVoice transmits the relevant audio and text over HTTPS to Xiaomi MiMo, a cloud AI service operated by Xiaomi Corporation in the People's Republic of China; because this provider processes requests in China, that audio and text leave your country during processing. These requests carry only an opaque account identifier used for usage metering — your name, email, and Apple identifier are not sent with them. Subscription receipts are sent to Apple for validation. See the Third Parties section below for details.

Third Parties We Share Data With

Xiaomi MiMo (Xiaomi Corporation, People's Republic of China) — The cloud AI provider that powers voice cloning, speech-to-text, chat replies, and cloned-voice synthesis. It receives the reference voice sample (to clone a family member's voice), the audio you record during voice chat (for transcription), the text of your messages and the family member's name you provide (to generate a reply, story, or voice note), the words you type for a line (to screen them for scam and abuse patterns, then to synthesize them), and the reply or line text (to synthesize speech in the cloned voice). Audio and text are sent over HTTPS for each request and are not retained on KinVoice's own servers; no name, email, or Apple identifier accompanies these requests, only an opaque account identifier for metering. Because Xiaomi MiMo processes requests in China, this data leaves your country during processing. Xiaomi MiMo privacy policy: https://privacy.mi.com/all/en_US/ Apple (Apple Inc., United States) — Provides Sign in with Apple, which returns your Apple user identifier and, on first sign-in, your email address or a private Apple relay address, and validates In-App Purchase subscription receipts. You can manage or stop forwarding from a private relay address in your Apple ID settings. Apple privacy policy: https://www.apple.com/legal/privacy/

Support

For privacy questions, contact gosingk@gmail.com. This policy is maintained by WU HAOTIAN.