Privacy Policy

Effective Date: July 7, 2026

This Privacy Policy explains how Zhongyun Hengke Information Technology Co., Ltd. ("we", "us", or "our") collects, uses, and shares information through the Voxa AI Screen Translator mobile application.

Core Privacy Principle: Voxa is designed with an "On-Device First" architecture. All screen captures, image OCR, and voice recognition are processed locally on your device. We do NOT store or upload your raw media files (images, audio, video) to any server.

1. Information We Collect and How We Use It

A. Content Data (Processed Locally)

To provide translation, Voxa captures your screen using the Media Projection API. This data is processed in real-time within your device's volatile memory. We do not store, record, or share your screen content, photos, or voice recordings.

B. Information You Provide

If you contact our support team, we collect your email address and any information included in your message to resolve your request.

C. Automatically Collected Data

We use third-party SDKs to collect technical information for app stability, performance optimization, and advertising attribution:

Device Identifiers: Advertising ID (GAID), Android ID.
Technical Info: Device model, OS version, IP address, battery level, and screen resolution.
Usage Data: Feature clicks, crash logs, and session duration.

D. Analytics and Diagnostic Events

To improve product quality, detect failures, measure feature adoption, and understand advertising performance, we may collect analytics and diagnostic events through services such as Firebase Analytics and Firebase Crashlytics. These events may include:

Feature usage, such as text translation, screen translation, game translation, manga translation, voice translation, subtitle translation, and widget or shortcut usage.
Translation workflow status, such as translation started, succeeded, failed, cache hit, quota limit, OCR result count, layout mode, language pair, and duration buckets.
Model download status, such as download prompt, start, progress bucket, success, failure, retry, cancel, file verification, extraction, and load result.
Advertising events, such as ad request, load, show, impression, click, dismiss, reward earned, and quota continuation after rewarded ads.
Crash and non-fatal diagnostic information, such as app version, Android version, device model, feature context, error type, stack trace, and service startup failures.

We do not send raw translation content through analytics events. Analytics and diagnostic events do not include your original text, translated text, screenshots, screen recordings, photos, audio recordings, video files, or OCR images. Where possible, we use aggregated or bucketed values, such as text length ranges, duration ranges, progress ranges, and result categories.

2. Third-Party SDKs and Data Sharing

We integrate the following SDKs to provide core services. These partners may collect and process data according to their own policies:

Partner / SDK	Data Collected	Purpose	Policy Link
Google AdMob / Firebase	Advertising ID, app usage events, crash logs, device and OS information, app version, approximate diagnostics, and ad interaction data	Ads, analytics, crash reporting, diagnostics, attribution, and product improvement	Google Privacy Policy Firebase Privacy and Security
TradPlus / Pangle	Device Identifiers, Ad Data	Ad Mediation	View
Meta (Facebook) Ads	Advertising ID, App Events	Marketing & Ads	View
GameAnalytics	Remote configuration data, OS information, app version, and technical metadata	Remote configuration and product operation	View
Google Play Billing	Transaction ID, Purchase Token	Payments	View

API Interfaces for Translation:

When you use the translation feature, the extracted text strings are securely transmitted via HTTPS to:

Google Cloud Translation API
DeepSeek API
Cloudflare Workers / AI

These services receive only text, not images or audio. They process data as "Processors" to return translation results.

For offline translation, offline OCR, local speech recognition, or on-device subtitle recognition features, supported processing is performed locally on your device and is not uploaded to our servers for translation.

3. Permissions Disclosures

Voxa requests the following sensitive permissions:

Screen Capture (Media Projection): Essential for recognizing text on other apps. Used only on-demand.
Camera: Used only for AR/Image translation features.
Microphone: Used only for Voice translation features.
Storage (Read/Write): Used to import/export local media files for translation.

4. GDPR & CCPA Compliance (User Rights)

If you reside in the European Economic Area (EEA) or California, you have the right to:

Access, update, or delete your data.
Opt-out of "Personalized Advertising" via Android System Settings.
Withdraw consent at any time for data processing.
Request information about the categories of personal information we process and the third-party services involved.

5. Data Retention & Security

Since we do not store content data, retention is minimal. Analytics data is retained according to our third-party partners' standard retention periods (typically 14-26 months). We use TLS/SSL encryption for all data in transit.

Crash reports and diagnostic events are retained only as long as reasonably necessary to identify, diagnose, and fix product issues. We do not use crash logs or analytics events to reconstruct your translated content.

6. Contact Us

Zhongyun Hengke Information Technology Co., Ltd.
Email: [email protected]
Address: Room 06298, 18th Floor, Building B, Bailong Plaza, Xiaozhai East Road, Yanta District, Xi'an City, Shaanxi Province