Privacy Policy

Javier Pozueco Perez (“we,” “our,” or “us”) operates the mobile application WeLang (the “App”). This Privacy Policy explains how we collect, use, disclose, and protect information when you use the App.

1. Scope

This Privacy Policy applies to all users of WeLang in all regions where the App is available.

2. Information We Collect

WeLang does not require account registration or login.

Depending on the features you use, we may process:

• User-provided learning content (for example: words, definitions, examples, notes, prompts, and practice text)
• Voice recordings when you use pronunciation/speaking features (microphone access is requested only for this purpose)
• Backend usage data (for example: requested backend endpoint, HTTP method, request timestamps, request counters, and Gemini token-usage statistics)
• Installation identifier: a randomly generated app-specific client_id stored locally on your device and sent with requests to help operate and protect the service

WeLang does not use advertising SDKs and does not collect data for advertising purposes.

3. How We Use Information

• Provide core app functionality
• Generate language-learning responses and feedback
• Evaluate pronunciation and speaking exercises
• Maintain reliability, security, and abuse prevention
• Improve app quality and performance
• Provide user support

4. Legal Bases for Processing

Where UK or European data-protection law applies, we rely on the following legal bases:

• Performance of a contract or steps requested by you: to provide the App features you choose to use, including generating learning content and pronunciation feedback
• Legitimate interests: to operate and improve the App, maintain reliability and security, prevent abuse, and enforce reasonable usage limits
• Legal obligations: where processing is necessary to comply with applicable law or valid legal requests
• Consent: where we specifically ask for consent and applicable law requires it; consent may be withdrawn at any time without affecting processing already carried out

5. Data Processing and Third Parties

WeLang communicates with our backend service to process feature requests (including AI-powered features).

We use Google Cloud Run to operate our backend, Cloud Firestore for temporary usage records and predefined vocabulary managed by WeLang and used to provide suggestions, Google Cloud Storage for a shared cache of synthetic pronunciation assets for texts included in WeLang’s predefined vocabulary catalogue, and Cloud Logging for security and operational logs.

Cloud Logging may process request metadata such as IP address, user agent, request path, HTTP method, response status, request and response size, and latency. These platform logs are retained for 30 days. Request bodies and the X-Client-Id header are not intentionally recorded in these platform request logs.

Our backend uses the paid Google Gemini API to generate requested language-learning content and feedback. Information sent to Gemini may include text you provide and contextual information required to produce the requested response. Google does not use prompts or responses submitted through paid Gemini API services to improve its products. Google currently retains prompts, contextual information, and generated responses for 55 days for abuse detection and prevention, service security, and required legal or regulatory disclosures. This monitoring uses automated systems, and content flagged as potentially violating Google’s policies may be reviewed by authorized Google personnel.

For speaking, pronunciation, and speech-synthesis features, our backend uses Microsoft Azure Speech. The Azure Speech resource used by WeLang is configured in the West Europe region.

Data is shared only as necessary to provide requested functionality and operate the service.

6. Audio and Microphone Data

If you use speaking/pronunciation features:

• We request microphone permission
• Your audio recording and the reference text you are practising are transmitted to our backend
• Our backend sends the audio and reference text to Microsoft Azure Speech in the West Europe region to recognize speech and calculate pronunciation feedback, including scores at sentence, word, and phoneme level
• The temporary audio file created by our backend is deleted after the pronunciation request has been processed
• According to Microsoft documentation, Azure Speech does not retain or store customer data provided for real-time speech recognition and pronunciation assessment

For pronunciation feedback, the voice recording itself is not sent to Gemini. After Microsoft Azure Speech evaluates the recording, WeLang sends derived assessment information to the paid Google Gemini API to generate personalized feedback. This information may include weak or incorrectly pronounced words, weak phonemes, pronunciation, accuracy, fluency, completeness and prosody scores, the overall performance band, and suggested focus areas.

Synthetic pronunciation audio and associated timing data for texts included in WeLang’s predefined vocabulary catalogue may be stored in Google Cloud Storage as part of a shared cache. This allows WeLang to reuse these assets instead of generating them repeatedly. These assets may be retained without a fixed expiration period and are not recordings of the user’s voice.

Synthetic pronunciation assets for other texts are delivered to the App without being retained in our cloud storage. The App may store them locally on the user’s device.

If you do not use these features, microphone data is not collected.

7. Local Storage on Device

WeLang stores certain data locally on your device, including:

• App preferences and settings
• Learning/session state
• App-generated installation identifier (client_id)
• User-created learning content (as part of app functionality)
• Locally downloaded synthetic pronunciation audio and associated timing data
• Your voice recordings used for pronunciation assessment, when you save the associated practice results

Locally stored synthetic pronunciation audio and associated timing data remain on the device until the associated learning item is deleted, the App’s data is cleared, or the App is uninstalled.

Voice recordings saved with pronunciation-assessment results remain on the device until the associated practice result or learning item is deleted, the App’s data is cleared, or the App is uninstalled.

8. Data Retention

When backend features are used, we temporarily record the installation identifier together with the requested endpoint, HTTP method, request timestamps, request counters, and Gemini token-usage statistics. These usage records do not contain the user’s prompts, voice recordings, or AI responses. They are assigned an expiration time of 48 hours. Cloud Firestore normally completes TTL deletion within approximately 24 hours after that expiration time.

Google Cloud platform request logs are retained for 30 days. Synthetic pronunciation assets for texts included in WeLang’s predefined vocabulary catalogue may be retained without a fixed expiration period while they remain part of the shared cache.

If you contact us for support, we may retain your email address, Installation ID, message content, attachments, and our responses for up to 12 months after the support request is closed. We use this information to respond to and manage support requests. These communications are processed using Gmail provided by Google.

We retain records of privacy and deletion requests, identity or Installation ID information provided for the request, our response, and evidence of the action taken for three years after the request is closed. This allows us to respond to requests and demonstrate compliance with applicable data-protection obligations. These records may be retained for longer only when required by applicable law or necessary for an ongoing legal claim or regulatory matter.

Other information is retained only for as long as needed to provide and secure the service, improve functionality, and comply with legal obligations. Where feasible, data is deleted or anonymized when no longer needed.

9. Children’s Privacy

WeLang is intended only for users aged 18 and over. Before the App loads its main features, users must confirm that they are at least 18.

We do not knowingly collect personal information from anyone under 18. If we learn that an under-18 user has bypassed the age screen, we will take reasonable steps to block access and delete any associated personal information within our control.

A parent or guardian who believes that someone under 18 has used WeLang may contact us using the details below to request review and deletion.

10. Data Sharing

We do not sell personal data. We may share data with:

• Infrastructure and processing providers needed to run the App
• Google Gemini APIs, to generate requested educational responses and feedback
• Microsoft Azure Speech, to provide requested speech synthesis, speech recognition, and pronunciation assessment
• Authorities if legally required
• Successors in the event of a merger, acquisition, or asset transfer

11. Security

We implement reasonable technical and organizational safeguards to protect data. However, no transmission or storage system can be guaranteed to be 100% secure.

12. International Processing

Because WeLang is available globally, information may be processed in countries other than the country where you live. Where personal data is transferred outside the United Kingdom or European Economic Area, we rely, as applicable, on adequacy regulations or decisions, approved contractual safeguards, or another lawful transfer mechanism. You may contact us to request further information about the safeguards that apply.

13. Your Rights and Deletion Requests

Depending on your location and the legal basis used, you may have the right to request access to, correction of, deletion of, restriction of, or portability of your personal data, and to object to certain processing. Where processing is based on consent, you may withdraw that consent at any time.

You may also lodge a complaint with the data-protection supervisory authority responsible for your country or place of residence.

When contacting us about server usage records, you can include the Installation ID shown under Profile → Your installation ID in the App so that we can identify the relevant records.

To exercise a privacy right or request deletion, contact:

• welang.support@gmail.com (deletion/support requests)
• jpozueco@gmail.com (privacy contact)

We will review and respond within a reasonable timeframe and according to applicable law.

14. Changes to This Policy

We may update this Privacy Policy from time to time. Any updates will be posted with a revised “Last updated” date.

15. Contact

• Developer: Javier Pozueco Perez
• Country of establishment: United Kingdom
• Privacy email: jpozueco@gmail.com
• Support/deletion email: welang.support@gmail.com