Privacy Policy

Mongol Talk — AI Translator | Effective Date: 1 May 2026 | Terms of Service / 服務條款

This Privacy Policy describes how Mongol Talk — AI Translator ("the App", "we", "our") collects, uses, and protects information when you use the App. We are committed to safeguarding your privacy and being transparent about our data practices.

1. Information We Collect

1.1 Voice Recordings

When you use the speech-to-text feature, your voice is captured by your device's microphone and sent via our backend to Google Cloud Speech-to-Text (Google) for processing. We do not use voice recordings as permanent, user-facing “history” on our servers. How audio is handled depends on the speech pipeline in use:

Typical (shorter) requests: Audio is relayed for recognition and is not kept as a long-lived audio file on our systems after the request completes.
Longer recordings (Speech-to-Text V2 batch path): For some long clips, our backend may place encrypted audio temporarily in Google Cloud Storage within our Google Cloud project solely to complete batch recognition with Google, then delete it where technically possible; any remaining temporary objects may be removed by automated lifecycle rules. Google still processes audio under Google's own terms and policies.

Optional optimisation (server-side): To reduce repeated cloud processing when you retry the same audio with the same recognition settings, our backend may store only the recognised text (not your raw audio) as a small JSON object in Google Cloud Storage, referenced by a cryptographic fingerprint of the audio and settings. This cache can be disabled in our server configuration.

1.2 Text Input & Translated Text

Text you type or text derived from speech recognition is sent via our backend to Google Cloud Translation API (Google) for translation. We do not store input or translated text on our own servers.

1.3 Synthesised Speech (Text-to-Speech)

Translated text is sent via our backend to Microsoft Azure Text-to-Speech (TTS) to generate audio playback. The resulting audio may be temporarily cached on your device for playback and is not uploaded to our servers.

1.4 Translation History & Favourites

Your translation history and saved favourites are stored locally on your device only using on-device storage (AsyncStorage). This data is never uploaded to any server. You can clear this data at any time from the App's Settings.

1.5 Daily Usage Count

The App displays and tracks a daily usage indicator locally on your device to help manage the free usage quota (5 free uses per day). Separately, when you call our authenticated APIs, our backend may record anonymous usage counters (for example per-day call counts) for security, fair use enforcement, and cost stability; see Section 1.9. These server records are not your translation history from the App (which remains on-device per Section 1.4).

1.6 Device & Technical Information

We may collect limited technical information for error diagnostics and service stability, including:

Device type, operating system, and app version
IP address (transmitted to third-party services as part of standard network requests)
Anonymous user identifier generated by Firebase Authentication
Crash reports and error logs (via Sentry, on native devices only)

1.7 Subscription & Purchase Information

If you subscribe to Mongol Talk Pro, your purchase is processed by the Apple App Store or Google Play Store. Subscription verification is handled by RevenueCat, which receives only your anonymous user ID and purchase receipt. We do not receive or store your payment card details.

1.8 Language Preference

Your selected interface language is stored locally on your device to provide a consistent experience across sessions.

1.9 Server-side operational metrics (Firebase / Google Cloud)

For users who authenticate (including anonymous Firebase Authentication), our backend may write non-content operational records to Google Cloud Firestore in our project. Examples include aggregated API usage counters and speech-recognition quality statistics (such as duration, latency, error codes, and model identifiers). These records do not include the full text of your translations as stored in the App's local history (Section 1.4), and do not store full transcripts or raw audio payloads in Firestore. Structured server logs may be emitted for reliability (for example a single-line JSON metric); they are not designed to retain your transcript text.

2. How We Use Your Information

Core Features: To provide real-time speech recognition, translation, and text-to-speech functionality.
Error Diagnostics: To monitor app stability, diagnose crashes, and improve performance via Sentry (native devices only).
Subscription Management: To verify your subscription status and manage access to Pro features via RevenueCat.
API Security: To authenticate API requests using anonymous Firebase Authentication to prevent abuse.
Free Quota Management: To help enforce free-tier and fair-use limits, including on-device indicators and server-side anonymous counters where applicable.
Service reliability: To monitor speech-recognition performance and stability using privacy-safe operational statistics (Section 1.9).

3. Third-Party Services

The App relies on the following third-party services. Each service processes data according to its own privacy policy:

Service	Purpose	Data Shared	Privacy Policy
Google Cloud Speech-to-Text	Voice recognition	Voice audio stream (via our backend)	Google Privacy Policy
Google Cloud Storage	Temporary storage for some long speech-to-text jobs; optional recognised-text cache (JSON)	Encrypted objects in our Google Cloud project (temporary audio where used; optional text-only cache)	Google Cloud Privacy Notice
Google Cloud Translation API	Text translation	Input text (via our backend)	Google Privacy Policy
Microsoft Azure TTS	Text-to-speech synthesis	Translated text	Microsoft Privacy Statement
Firebase Authentication	Anonymous user authentication	Anonymous user ID	Firebase Privacy
RevenueCat	Subscription management	Anonymous user ID, purchase receipt	RevenueCat Privacy Policy
Sentry	Crash reporting & error monitoring (native devices only)	Device info, OS version, app version, anonymised error/crash logs	Sentry Privacy Policy

Note: API requests are routed through our secure backend for authentication and rate limiting. Your translation history and favourites remain on your device (Section 1.4). For speech recognition, audio is relayed to Google Cloud Speech-to-Text; in some configurations, temporary encrypted audio or a text-only recognised transcript cache may reside in our Google Cloud Storage as described in Section 1.1. Text you send for translation is relayed to Google Cloud Translation API and is not retained by us as long-term content storage. Always refer to the third-party policies linked in the table above (and the in-app Settings → Third-party services list where shown).

4. Data Storage & Retention

Data Type	Where Stored	Retention
Voice recordings	Processed by Google Cloud Speech-to-Text; may pass through our backend; optional temporary encrypted objects in our Google Cloud Storage for some long requests (Section 1.1)	No long-term audio retention by us as user history; temporary objects deleted where possible and otherwise subject to automated lifecycle rules
Optional speech transcript cache (text only)	Google Cloud Storage in our project (hashed object names)	Short retention via automated lifecycle rules; can be disabled in server configuration
Input & translated text	Processed in real-time by Google Cloud Translation API; not stored by us	Not retained
TTS audio cache	Temporarily on your device	Automatically cleared
Translation history & favourites	Locally on your device only	Until you clear it in Settings
Daily usage count	Locally on your device only	Resets daily; clearable in Settings
Language preference	Locally on your device only	Until changed by you
Crash reports & error logs	Sentry servers (native devices only)	Subject to Sentry's retention policy
Subscription status	RevenueCat servers	Subject to RevenueCat's retention policy
Anonymous user ID	Firebase servers	Subject to Firebase's retention policy
Server-side usage & speech quality statistics	Google Cloud Firestore in our project (non-content aggregates)	Operational retention consistent with our backend configuration (see support documentation on metrics retention)

5. Data Sharing

We do not sell, rent, or share your personal data with any third party for marketing or advertising purposes. Data is shared with the third-party services listed in Section 3 solely to provide the App's core functionality.

6. User Consent & Revocation

6.1 Consent Before Data Sharing

Before any voice, text, or other user data is sent to third-party AI services (Google Cloud Speech-to-Text, Google Cloud Translation API, Microsoft Azure Text-to-Speech), the App displays a Data Usage Notice on first launch. This notice clearly identifies:

The specific third-party services your data will be sent to (as listed in the App, including under Settings);
The exact types of data shared (voice recordings, text input, translated text);
The purpose of sharing the data (speech recognition, translation, text-to-speech).

You must explicitly tap "Accept & Enable AI Features" to grant consent. If you decline, no data will be sent to any third-party AI service, and translation features will remain disabled.

6.2 Revoking Consent

You may withdraw your consent at any time by navigating to Settings → Permissions & Data → Cloud AI Features → Withdraw Consent. Upon revocation:

All AI-powered features (translation, speech recognition, text-to-speech) will be immediately disabled.
No further voice or text will be sent via our backend to Google Cloud Speech-to-Text, Google Cloud Translation API, or Microsoft Azure Text-to-Speech for AI processing.
Your local translation history and favourites will not be affected.
You may re-enable AI features at any time by re-granting consent through the Data Usage Notice.

7. Your Rights & Choices

Consent Management: You can grant or withdraw AI data consent at any time from Settings within the App (see Section 6).
Clear Local Data: You can delete all translation history, favourites, and usage counts from Settings within the App.
Manage Subscriptions: You can manage or cancel your subscription through the Apple App Store or Google Play Store.
Data Enquiries: You may contact us to enquire about, request correction of, or request deletion of any data associated with your anonymous identifier.
Opt Out of Error Reporting: Sentry crash reporting is active only on native (iOS/Android) devices. Using the App via Expo Go or web does not send data to Sentry.

8. Children's Privacy

The App is not designed for or directed at children under the age of 13. We do not knowingly collect personal information from children under 13. If you believe a child under 13 has provided us with personal data, please contact us and we will take steps to delete such information.

9. Security

Translation, speech recognition, and text-to-speech credentials are kept on our backend and are not exposed in the app; where feasible we inject secrets using Google Secret Manager (or equivalent) rather than embedding them in application source code.
Firebase Authentication uses a public API key that is designed to be included in client-side code and is protected by Firebase Security Rules.
Sensitive local data (such as usage counts and subscription status) is stored using encrypted on-device storage (Secure Store).
All network communications between the App and our servers (and third-party services) are encrypted via HTTPS.

No method of electronic transmission or storage is 100% secure, and we cannot guarantee absolute security.

10. Disclaimer

Mongol Talk — AI Translator is an AI-powered tool and may not always produce perfectly accurate translations. The App is provided "as is" without warranties of any kind. We are not liable for any damages, misunderstandings, or losses arising from the use of translations provided by the App.

10.1 Flags and regional symbols

Any national or regional flags (or similar symbols) shown in the App are used only as visual aids to help you identify languages, locales, regional voice options, or interface display settings. They are not intended to express any political opinion, endorsement, territorial claim, or affiliation with any government or political entity. Nothing in the App should be interpreted as taking a political position. We reject misuse of these neutral UI elements for political pressure, harassment, or bad-faith complaints.

11. Changes to This Policy

We may update this Privacy Policy from time to time. Any changes will be posted on this page with an updated effective date. We encourage you to review this page periodically. Continued use of the App after changes constitutes acceptance of the updated policy.

12. Support & FAQ

12.1 Frequently Asked Questions

Q: How many free translations do I get per day?
A: Free users receive 5 translations per day. The count resets at midnight (local time). Mongol Talk Pro removes the daily free cap on translations, subject to our Terms of Service and fair use policy (abnormally high or automated use may be limited).

Q: Which languages are supported?
A: Mongol Talk supports 100+ languages for text translation via Google Cloud Translation API, and many languages for voice input (Google Cloud Speech-to-Text) and voice output (Microsoft Azure Text-to-Speech). The full list is available inside the app.

Q: Is my voice data stored on your servers?
A: We do not keep voice audio as a user-facing history on our servers. Audio is sent via our backend to Google Cloud Speech-to-Text. For some long recordings, encrypted audio may be held temporarily in our Google Cloud Storage solely to complete recognition, then removed where possible (see Section 1.1). An optional text-only cache of recognised words may also exist in Google Cloud Storage to reduce repeated processing. Your translation history and favourites remain on your device (Section 1.4).

Q: How do I manage or cancel my subscription?
A: iOS: Settings → Apple ID → Subscriptions → Mongol Talk Pro.
Android: Google Play Store → Menu → Subscriptions → Mongol Talk Pro.
You can also restore purchases from the app's Settings page.

Q: Why does the app ask for microphone permission?
A: Microphone access is required for the speech-to-text feature. You can grant or deny permission at any time from your device settings. The app also works with text input only.

Q: How do I withdraw my data consent?
A: Go to Settings → Permissions & Data in the app. Tap "Withdraw Consent" to immediately stop all data transmission to third-party AI services. You can re-enable consent at any time.

Q: Can I use the app offline?
A: Translation requires an internet connection because the app uses cloud-based AI services. However, your saved translation history and favourites are available offline.

12.2 Troubleshooting

Translation not working: Check your internet connection. Ensure you have granted data consent (Settings → Permissions & Data). If you see "Daily limit reached", wait until midnight or upgrade to Pro.

Voice input not responding: Ensure microphone permission is granted in your device settings. Try restarting the app.

Subscription not recognised: Go to Settings → Restore Purchases in the app. Make sure you are signed in with the same Apple ID / Google account used for the purchase.

13. Contact Us

If you have any questions, concerns, or requests regarding this Privacy Policy or need support, please contact us at:

Email: dracocheng2024@gmail.com