SafeDictate — Private Voice Typing

Voice dictation that runs Whisper AI locally — your audio never leaves your device. Built for privacy-sensitive work.

🎙️ SafeDictate — Voice Typing That Physically Can't Leak The Chrome extension that runs Whisper AI directly in your browser. No server. No cloud. No way for your audio to go anywhere. Press Alt+Shift+V → speak → release. Your words appear in the active field. That's it. Nothing is sent anywhere. Ever. ━━━━━━━━━━━━━━━━━━━━━━━━━ 🔒 THE ZERO-LEAK ARCHITECTURE Other tools say "we don't store your audio." SafeDictate makes it physically impossible to send it in the first place. Whisper AI runs entirely inside your browser using WebAssembly. The microphone data never touches a network socket. ✅ Built for HIPAA-conscious work — your audio never leaves your device ✅ GDPR-friendly for EU users ✅ Zero telemetry, zero tracking, zero sign-up required ✅ Works offline after the first model download ━━━━━━━━━━━━━━━━━━━━━━━━━ 🩺 BUILT FOR PEOPLE WHO HANDLE SENSITIVE DATA → Doctors & therapists: dictate clinical notes in your EHR (SimplePractice, Jane App, Athena) while keeping audio on your device → Lawyers: draft documents with NDA content and zero cloud exposure → Privacy-first professionals: because "trust us" is not a compliance strategy ━━━━━━━━━━━━━━━━━━━━━━━━━ ⚡ WORKS EVERYWHERE YOU TYPE SafeDictate injects your transcription directly into the active text field — no copy-paste required: ✉️ Gmail — compose full emails hands-free 🤖 ChatGPT — prompt engineering at speaking speed (150 WPM vs 40 WPM typing) 📄 Google Docs — seamless voice editing with cursor awareness 📊 Google Sheets — cell-level dictation 📋 Any input, textarea, or contenteditable field on any website ━━━━━━━━━━━━━━━━━━━━━━━━━ 🚀 HOW IT WORKS 1. Install → quick 4-step onboarding (2 minutes) 2. Choose language: English or Auto-detect (50+ languages supported) 3. Whisper model downloads once (~40 MB) and is cached locally 4. Click any text field on any page → press Alt+Shift+V → speak → press again to stop 5. Text appears. Done. Or right-click any editable field and choose "Start voice dictation" from the context menu. ━━━━━━━━━━━━━━━━━━━━━━━━━ 🧠 POWERED BY OPENAI WHISPER SafeDictate uses the same Whisper model that powers leading transcription APIs — running entirely on your machine via WebAssembly. • Whisper Tiny (~40 MB) — fast, great for most use cases • Whisper Base (~80 MB) — higher accuracy for complex terminology (Pro plan) • Multilingual — automatically detects your language or use manual selection • No GPU required — runs on any modern laptop or desktop (Windows & Mac) ━━━━━━━━━━━━━━━━━━━━━━━━━ ❓ FREQUENTLY ASKED QUESTIONS Q: Does it really work with no internet? A: After the one-time model download — yes, completely offline. Q: What languages are supported? A: Whisper supports 50+ languages. English and Russian have the best accuracy with the base model. Q: Does it work on Windows? A: Yes! Unlike Voibe or Superwhisper (Mac-only), SafeDictate runs in Chrome on any OS. Q: Do I need to create an account? A: No account, no email, no sign-up. Install and go. Q: Is my audio stored anywhere? A: No. Audio is processed in-memory inside your browser tab and immediately discarded. Q: What permissions does the extension need? A: Microphone (to record your voice) and access to the current tab (to inject text). That's it. ━━━━━━━━━━━━━━━━━━━━━━━━━ 🔗 Links • Website: https://safedictate.softlab.cc • Privacy Policy: https://safedictate.softlab.cc/privacy • Support: https://safedictate.softlab.cc/support