PrivatePrompt Guard

Private second look before you send prompts to ChatGPT, Claude, or Gemini. Flags emails, IBANs, keys—on your device.

PrivatePrompt Guard gives you a private second look before you send a prompt to a cloud AI chat in the browser. How it works 1. Write as usual on a supported AI chat site. 2. When you press Send, the extension checks the text in the message box on your device. 3. If it finds patterns that often mean sensitive data—ID numbers, IBANs, card-like numbers, common API-key shapes, JWTs, private-key blocks, or password-style assignments—it pauses and shows a warning. 4. You can review the draft, anonymize matches with placeholders, send anyway once, or cancel. Why install it People paste client details, keys, and billing data into AI chats by accident. PrivatePrompt Guard is a local reminder, not a cloud scanner. Prompt text is not uploaded to AI4Context servers. Sensitivity - Balanced (default): medium and high-risk patterns - Strict: also emails and phone numbers - Relaxed: high-risk only - Custom: choose which types to flag Optional on-device AI If Chrome’s Gemini Nano (Prompt API) is available, it can help flag extra fragments. If it is not available, rule-based detection still works. Privacy Checks run on your device. The extension only acts on allowlisted AI chat sites and only when you try to send. It does not watch your other tabs in the background. Not enterprise DLP. It may miss secrets. You decide whether to send. Interface: English and Spanish. By AI4Context.