1. Voice Data & Recordings

When you record raw audio to evaluate your English pronunciation:

• Your voice input is processed on-the-fly to perform speech-to-text conversion and phonetic grading.
• Your voice and recordings are never recorded, saved, or sold to third parties.
• The audio processing is transaction-based, meaning voice segments are discarded promptly after analysis.

2. Client Storage & scoring History

Any score history or pronunciation levels you see on your dashboard are stored entirely inside your local browser storage. Clearing your browser data, cookie caches, or logging out will discard all local session statistics immediately.

3. Service Evaluation

To transcribe and score your voice clips, our application executes secure, direct API calls to language models (such as Google’s Gemini platform). This standard processing is encrypted and adheres strictly to corporate data trust rules.

4. Browser Diagnostics

We collect safe diagnostics like system screen width, general click counts, and mic input status alerts to debug and stabilize our media record controls.

5. Queries and Revisions

We may improve this document as the system expands. If you have concerns about client-side permissions or microphone capture, reach out to us inside the project space.