How can we help?

Make sure you've granted microphone permission to the extension. Click the SafeDictate icon in the toolbar → open Settings → confirm the microphone is enabled in Chrome's site permissions. Also try refreshing the page after enabling dictation for the first time.

The Whisper model (~90MB) is cached in your browser's IndexedDB storage. If Chrome clears storage on exit (a setting some users enable for privacy), the model is re-downloaded each session. Go to chrome://settings/cookies and add chrome-extension://[your-extension-id] to the "Sites that can always use cookies" list to persist the cache.

SafeDictate inserts text into the element that had focus when you activated dictation. Click directly inside the target input field before pressing the shortcut key. Google Docs requires a separate content-script integration — if you're having issues there specifically, let us know via the form below.

After checkout, SafeDictate activates Pro automatically using the email from your purchase — there's no key to enter. If Pro isn't recognized, make sure you're online so the extension can verify your subscription with Paddle, and that you paid with the same email. If it still doesn't activate, email us with your Paddle order ID and we'll sort it out.

Yes — once the Whisper model is downloaded and cached, transcription works fully offline. The only network calls are the one-time model download and, for Pro users, a periodic subscription check with Paddle — neither involves your audio.

SafeDictate supports all ~99 languages that Whisper supports — English, Spanish, French, German, Russian, Japanese, Chinese, Portuguese, and more. You can set your preferred language in the extension Settings.