DualVox Privacy Policy

中文 · EN · 日本語 · 한국어 · DE · FR

Last updated: October 20, 2026
Effective: September 19, 2026

“DualVox” (the “App” or the “Service”) is provided by Zhongke Youyun (Beijing) Information Technology Co., LTD (“Zhongke Youyun”, “we”, or “us”). We understand how important personal information is and take reasonable steps to protect it, including under the Personal Information Protection Law of the People’s Republic of China, the Cybersecurity Law of the People’s Republic of China, Canada’s PIPEDA, and other applicable laws.

This Policy explains how we collect, use, store, share, transfer, and protect your personal information, and how you can exercise your rights. Please read it carefully before using the App. We present this Policy through in-app prompts and Settings links. After you agree and start using the App, we process information as described here. Refusing optional information or permissions does not stop you from learning about basic features, but the related feature may be unavailable.

Privacy Policy summary

1. Controller: Zhongke Youyun (Beijing) Information Technology Co., LTD
2. Core features and data:
   • Talk-through / single-cam scenes / follow-record (necessary when you use them): after you grant permission, camera video, microphone audio, or photos/videos you import are processed on device to produce a video;
   • Looks, effects, and cartoon avatars (necessary when you use them): filters or cartoon presenter processing on device;
   • Info watermark (necessary when you use it): date and time overlaid on device; if you turn on location, while-in-use location is reverse-geocoded by the system into a place name written on the watermark and into the video. The place name cannot be edited by hand. If you choose the Weather card, the current coordinates are sent to Open-Meteo for temperature and conditions (see below); weather cannot be edited by hand;
   • Captions (necessary when you use them): speech in a finished clip is recognized on device and turned into editable captions, then exported as a new video. Audio is not uploaded to our servers. On iOS, if a language pack is missing, the system may download an Apple speech model once. On Android, the first time you use a language, the app may download an on-device model (model files only, not your audio);
   • Teleprompter (necessary when you use it): scripts you type or apply from a scene template are stored on device (up to 10) for on-screen cues only; they are not burned into the video and are not uploaded;
   • Save video / take photo / import media (necessary when you use them): after permission, write or read videos and photos in the system library;
   • Account sign-in (necessary when you sign in): on iOS, Sign in with Apple only (Apple user identifier, identity token, plus email and name if you choose to share them, used for a nickname). On Android, a mainland China mobile number and SMS code are available only in Simplified Chinese. Used to identify the account, membership, and deletion. iOS does not offer phone sign-in;
   • Membership subscription (necessary when you pay): Apple transaction ID, original transaction ID, product ID, transaction JWS, and expiry, used to unlock, verify, and restore DualVox Membership. Apple processes payment; we do not collect card numbers;
   • AI script (necessary when you use it): if you generate, polish, or expand a script, your topic, notes, or draft is sent through our servers in mainland China to Volcengine Ark (Doubao; Beijing Volcano Engine Technology Co., Ltd.) to return generated or rewritten text. We store the daily count, not the script body;
   • Cloud voiceover and personal voices (necessary when you use them): if you choose a cloud voice, the script is sent through our servers to MiniMax for synthesis and the audio is returned to the device; we store usage counts, not the script text. If you record “My voice”, a sample is sent through our servers to MiniMax to create that voice; we store the voice ID, name, and created/deleted times. The clone validation prompt is stored briefly. The default on-device voice does not upload script or audio;
   • Add-on purchases (necessary when you buy): Apple transaction ID, product ID, and transaction JWS, used to verify one-time items such as a personal voice or extra cloud characters;
   • Safety and basic operation (necessary): app version, client type (iOS/Android), language preference; our servers may log IP, time, and path when you call our APIs. On first open we collect a random install ID, OS type, language, region, and app version to count installs (once per install; no video, audio, Photos, or account). If the app crashes or a save fails, and Diagnostics is not turned off in Settings, we may collect device model, OS version, app version, failure reason, whether headphones are connected, and an operation summary without video, to fix bugs. We do not use a third-party crash or advertising-tracking SDK; this data goes only to our servers in mainland China. Install counting is necessary after you start using the App. Settings → Diagnostics only controls crash and save-fail reports.
3. Important: this version does not store your finished videos in the cloud as a product purpose. Camera, microphone, imported library media, looks/cartoon, watermark, stabilization, and caption recognition run on the device by default. Videos and photos are saved to your system library by default. Location coordinates are not uploaded to our servers; the place name is resolved on device. If you choose the Weather card, coordinates go to Open-Meteo for weather and do not pass through our servers. If you sign in, account data (Apple user identifier on iOS; mobile number for Android Chinese phone sign-in; plus account ID, nickname, membership status, cloud-narration usage, daily AI-script counts, and personal-voice records) is stored on servers in mainland China. We do not use your recordings, scripts, or generated text to train AI models. On-device captions are recognized locally; the audio stays on the device.
4. Contact: [email protected]; mailing address at the end of this Policy.

1. How we collect and use personal information

We follow lawful, legitimate, necessary, and good-faith principles, and process personal information only as needed for product features. “Collect” includes information you provide, information generated while using a feature, and information obtained from system APIs after you authorize it.

1.1 Features and processing

1.2 System permissions

We request a permission only when the related feature needs it, and we explain the purpose in the system prompt or Settings. You can turn permissions off in system settings at any time. Related features may stop working; other features that do not need that permission are unaffected.

• Camera: on-camera presenting, dual or single camera, and still photos on the capture screen;
• Microphone: talk-through audio, cartoon lip-sync, and teleprompter pace-follow. iOS captions need Speech Recognition separately; Android does not request that permission. Basic recording does not upload audio to us. Only if you record “My voice” is a sample sent through our servers to MiniMax;
• Speech recognition / on-device models: recognize talk-through on device and generate captions; audio is not uploaded to our servers. iOS uses system Speech Recognition and may download an Apple model once; Android does not request that permission and may download an on-device model the first time you use a language;
• Photos / storage: save videos and photos, and import scene media;
• Location (while in use): only when you turn on the place watermark or the Weather card, to resolve a place name or query Open-Meteo. Turning this off hides place and weather details and does not affect other recording.

1.3 Information we do not collect for basic recording

We do not require ID numbers, contacts, SMS content, or upload of precise coordinates to our servers for basic recording, and we do not use IDFA / ATT advertising tracking. You can record without signing in. Mobile number is collected only on Android in Simplified Chinese when you choose mobile sign-in; iOS does not offer phone sign-in. With Sign in with Apple we receive your Apple user identifier and identity token; email and name are provided only if Apple’s first authorization includes them and you choose to share. You can refuse name/email and still sign in. Location is used only when you turn on the place watermark or the Weather card: the place name is resolved on device; a weather query sends coordinates to Open-Meteo. iOS membership and add-ons are purchased through Apple In-App Purchase; Apple processes payment, and we only receive transaction JWS, subscription status, and one-time item verification needed to fulfill the purchase. We do not currently use third-party analytics or push SDKs. If we add them, we will update this Policy first and obtain consent in the relevant flow where required.

1.4 Processing without additional consent where the law allows

Where laws require or allow it—including national security, public security, criminal justice, emergency rescue, or other legal duties—we may process related information without additional consent.

2. How we store personal information

1. Where: audio/video, looks/cartoon, watermark, stabilization, and caption recognition during use run on your device by default. Finished videos and photos are saved to your system library by default. Login tokens, teleprompter scripts, caption caches, and recording preferences stay on the device. After sign-in, mobile number (Android Chinese phone sign-in only) or Apple user identifier, account ID, nickname, membership status, cloud-narration usage and extra packs, and personal-voice IDs and names are stored on servers in the mainland of the People’s Republic of China (koubo-api.mediapea.com). Sign in with Apple and In-App Purchase credentials are processed in Apple’s systems. First-open install counts (random ID, OS, language, region, and version) are also stored on those servers. Cross-border transfers will be disclosed and consented to as required by law.
2. How long: you manage and delete local videos yourself. Install counts are kept, de-duplicated by install ID, until the statistic is no longer needed or you ask us to delete that ID; this is independent of deleting an account. Extra cloud-character packs are kept for their validity period and then stop counting toward remaining quota. After you delete a personal voice in the App, it is no longer usable, but the ID and timestamps are kept until you delete the account or the law requires otherwise, for support checks. We keep other server-side information only as long as needed for the stated purpose, or longer if the law requires it.
3. If we shut down: we will notify you as required and delete or anonymize personal information under our control, unless the law requires otherwise.

3. Sharing, transfer, and public disclosure

1. We do not sell your personal information.
2. We do not share personally identifiable information with third parties without your consent, except: your explicit consent; as needed to enter into or perform a contract you are a party to; as required by law or a competent authority; or in a merger, acquisition, or asset transfer with legal protections.
3. Basic recording does not depend on uploading your videos to third-party cloud storage. iOS sign-in uses Sign in with Apple only; payment is completed through Apple In-App Purchase. SMS codes go through the UECloud platform to Alibaba Cloud SMS only for Android Simplified Chinese mobile sign-in. Android membership in mainland China is paid through Alipay. If you choose the Weather card, coordinates are sent to Open-Meteo; see the Third-Party Sharing List. On-device captions are recognized locally and are not sent to a third-party transcription service. iOS may download an Apple language pack; Android may download an on-device model the first time you use a language (model files only). If you use AI script, your topic, notes, or draft is sent through our servers to Volcengine Ark (Doubao) to generate or rewrite text, and the result returns to the device. If you choose a cloud voice for narration, your script is sent through our servers to MiniMax for speech synthesis and the audio is returned to the device. If you record “My voice”, a sample of your speech is sent only to create that voice. The default on-device voice does not use MiniMax. If we later add analytics or push SDKs, we will list names, purposes, and data types in this Policy or the related schedule, and obtain consent where required.
4. Information that has been de-identified so a person cannot be re-identified may be shared, transferred, or disclosed as permitted by law.

4. How we protect personal information

We use reasonable technical and organizational measures, including least-privilege permissions, on-device processing, and access control, to reduce unauthorized access, disclosure, use, alteration, damage, or loss. Please keep your device and unlock method secure. If a security incident occurs, we will follow the law, take emergency steps, and notify you as required.

5. Your rights

1. Access, copy, and correction: you can open this Policy from Settings. Local videos can be managed in the system library. If account profile pages are added later, you can review and correct them there.
2. Deletion: delete videos in the system library, or uninstall the App and clear app data as the OS allows. If we hold personal information on our servers, you may request deletion via the contacts below (except where the law requires retention).
3. Withdraw consent: turn off permissions in system settings, or stop using the App. Diagnostics can be turned off in Settings; that does not stop the one-time install count. You can ask us to delete the install ID using the contacts below. Withdrawal does not affect processing that already happened based on prior consent.
4. Delete your account: after signing in, tap “Delete account” at the bottom of Settings. We then delete account and related server data under our control (except where the law requires retention). This cannot be undone.
5. After we verify your identity, we will respond to reasonable requests within fifteen days or the period required by law.

6. Children

We take children’s personal information seriously. If you are under 18, please read this Policy and use the Service with a parent or guardian. If you are under 13 (or under 14 in mainland China), you may use the Service only with a guardian’s consent. Guardians may contact us about a child’s personal information using the details below.

7. Scope

This Policy applies to the DualVox client (Android / iOS) and related pages on our documentation site. It does not apply to third-party apps, sites, or platforms you use after recording in DualVox; those parties have their own policies.

8. Changes

We may update this Policy from time to time. For material changes (purposes, types of data, sharing parties, or how you exercise rights), we will notify you prominently, such as with a dialog or notice. If you do not agree, please stop using the App. Continued use means you accept the updated Policy.

9. Contact us

For questions, complaints, or suggestions about this Policy or personal information:

• Email: [email protected]
• Mail: Room 21082, 22F, Building 1, Greenland Plaza, Plot IV-07, Canal Core Area, Tongzhou District, Beijing, China
• Website: https://dualvox.mediapea.com/

We will review and respond within a reasonable time.

Appendix: terms

• Personal information: information recorded electronically or otherwise that relates to an identified or identifiable natural person, excluding information after anonymization.
• Sensitive personal information: personal information that, if leaked or used unlawfully, could easily harm dignity or personal or property safety, including biometrics and tracking information, and personal information of children under 13 (under 14 in mainland China).
• Anonymization / de-identification: processing so a specific person cannot be identified and cannot be restored, or cannot be identified without additional information.