Video Translator does not collect, store, or transmit any personal data.

WHAT STAYS ON YOUR DEVICE

Video frames sampled from the page, the subtitle text recognized from them, and the translations produced from that text are all processed locally, inside the extension, and are never sent anywhere. They are held in memory for the duration of playback and discarded.

Your settings (target language, source language, OCR quality, overlay size, opacity and position) are stored using the browser's own extension storage. They stay in your Firefox profile. If you have add-on syncing enabled in your Firefox Sync settings, Firefox itself may sync them between your own devices; the add-on neither controls nor reads that process.

NETWORK ACCESS

The extension makes exactly one kind of outbound request: downloading neural translation models from Mozilla's public translation model storage on storage.googleapis.com, the first time you use a given language pair. These are plain GET requests for model files. They carry no video data, no recognized text, no page addresses, and no identifier of any kind. Downloaded models are cached locally so the request is not repeated.

There is no analytics, no telemetry, no error reporting, no advertising, and no third-party service of any kind.

WHY THE PERMISSIONS ARE NEEDED

The add-on requests three things, and nothing else:

• Access to youtube.com: to read video frames and draw the translation overlay on the page.
• Access to storage.googleapis.com: solely to download the translation models described above.
• Storage: to remember your settings between sessions.