Private Guard — PII Redaction

About

🛡️ Private Guard — Your PII Never Leaves Your Browser
Private Guard automatically detects and redacts Personally Identifiable Information before it's sent to ChatGPT, Gemini, Claude, or Copilot — and seamlessly restores the original values in the AI's responses.
🔍 3-LAYER DETECTION PIPELINE:
• 30+ regex patterns — emails, phone numbers, credit cards, IBANs, SSNs, API keys, JWTs, crypto wallets, and more
• AI-powered NER — on-device Named Entity Recognition detects names, organizations, and locations
• Active Learning MLP — a neural network that learns YOUR company-specific jargon and project names in real-time
🔄 AUTOMATIC RESTORATION:
When the AI responds with tags like [PERSON_1], Private Guard instantly restores them to the original values — you see the real answer, the AI never saw the real data.
👁️ LIVE PREVIEW:
Before sending, a floating panel shows all detected PII with type icons, editable labels, and one-click dismiss for false positives.
🔐 PRIVACY-FIRST:
• 100% on-device — zero data sent to external servers
• No tracking, no telemetry, no analytics
• Per-tab isolation — each tab has its own redaction state
• Open source — inspect the code yourself
✅ SUPPORTED SITES:
• ChatGPT (chatgpt.com)
• Google Gemini (gemini.google.com)
• Claude (claude.ai)
• Microsoft Copilot (copilot.microsoft.com)
⚡ CUSTOM PROTECTION:
• Add your own watchlist terms with custom labels
• Mark any word as sensitive via right-click context menu
• The built-in neural network learns and remembers your custom terms
Built with ❤️ for your privacy. Free and open source.