Privacy Policy — AuraLang

What AuraLang accesses

• Tab audio: when you click “Start translation,” the extension captures the audio stream of your active tab (tabCapture permission) in order to transcribe and translate it. The original tab audio is muted while capturing.
• Settings: your source/target language, interface language, and theme (dark/light) are stored locally via chrome.storage.local.

What AuraLang does NOT do

• It does not collect personal identification data, health data, financial data, authentication credentials, personal communications, location, browsing history, or user activity (clicks, keystrokes).
• It does not have a backend server. There is no account, no login, no analytics, no telemetry.
• It does not sell or share user data with third parties for advertising or any purpose unrelated to providing the translation feature.

How audio is processed

1. Transcription: audio is transcribed locally, on your own device, using a Whisper model (@huggingface/transformers, ONNX Runtime WASM). The model weights (~40–150 MB depending on the variant) are downloaded once from Hugging Face (huggingface.co) and cached locally. Audio never leaves your device for transcription.
2. Translation: the transcribed text is sent to Google’s public translation endpoint (translate.googleapis.com) solely to obtain the translated text. No other data accompanies this request.
3. Speech: the translated text is read aloud locally using the browser’s built-in Web Speech API (speechSynthesis). This does not leave your device.

Data retention

Nothing is stored outside your browser. Settings persist in chrome.storage.local until you uninstall the extension or clear it manually. Audio and transcribed text are processed in memory and are not saved anywhere.

Contact

Questions about this policy can be raised via GitHub Issues.