AI Promptwall

AI Promptwall stops personal details and credentials from reaching an AI chatbot by mistake. It watches the message box on ChatGPT, Claude, Gemini, Perplexity and Microsoft Copilot. The moment you press Enter or click send, it scans what you wrote and masks anything sensitive, before the message leaves your browser. No account. No API key. No sign-up. Nothing is ever sent anywhere. WHAT IT CATCHES Twenty-four detection rules covering personal contact details, payment card numbers, Australian and United States government identifiers, bank account details, home addresses, and the credentials people paste without thinking: cloud provider keys, access tokens, database logins and private keys. Checksum validation keeps an order number, a year or a version string from being mistaken for a card number or a tax file number. That matters more than it sounds: a tool that constantly false-alarms gets switched off, and a switched-off tool protects nobody. MASKED OR BLOCKED Most items are partially masked, so your message still reads sensibly. An email address becomes nxxxxxxx@example.com. A card number becomes xxxx followed by its last four digits. A few things stop the send completely: private keys, database logins, cloud secret keys and Tax File Numbers. A leaked credential needs rotating, and quietly hiding it would keep that from the one person who can act on it. You get a button to strip the credential and send the rest of your message. FILE ATTACHMENTS TOO Attach a spreadsheet or a document and AI Promptwall cleans a copy before it uploads. The original bytes never reach the site. Plain text files and Office documents are scanned and rebuilt. Anything that cannot be inspected safely inside a browser, such as a PDF or an image, is blocked rather than passed through unchecked. HOW IT HANDLES YOUR DATA Everything runs inside your own browser. There is no server, no account and no network request of any kind. The extension keeps a short local history of what it caught, recording category labels only, such as "Email address", and never the values themselves. That history is capped, expires after thirty days, and you can clear it whenever you like. No third-party libraries are bundled, which removes an entire supply-chain surface. WHAT IT WILL NOT DO This is a safety net, not a compliance control. It catches accidental disclosure by well-meaning people. It will not catch sensitive information written as ordinary prose, text inside images, or values that have been deliberately obscured, and anyone can switch it off. Do not rely on it as your only safeguard. Click the toolbar icon for a single on and off toggle.