PromptVeil – AI Prompt Privacy Guard

Stop sensitive data before it reaches an AI chat. PromptVeil reviews your prompt locally when you try to send it on an AI website you have enabled. If it detects something sensitive, it pauses the send and lets you redact selected findings, edit the prompt, or send it unchanged. WHAT PROMPTVEIL CAN FLAG • API keys, access tokens, passwords, and private keys • Service-account and database credentials • Payment-card and financial identifiers • Email addresses and phone numbers • IP addresses and other network identifiers • Customer names, project names, and your own confidential terms HOW IT WORKS 1. Install PromptVeil and open its settings. 2. Enable only the AI websites where you want protection. 3. Write your prompt normally. 4. When you try to send, review any detected sensitive text. 5. Redact selected findings, cancel and edit, or send unchanged. YOU STAY IN CONTROL PromptVeil does not silently modify or block your prompt. Every warning gives you the final decision. You can enable or disable individual detectors and add confidential terms that matter to your work. PRIVACY BY DESIGN • Prompt inspection and redaction run locally in your browser. • Prompt text and matched values are not uploaded by PromptVeil. • No account, publisher backend, analytics, advertisements, or external AI API. • Website access is optional and requested separately for each site you enable. • PromptVeil does not request access to all websites, cookies, or clipboard contents. • Temporary redaction mappings remain only in the current tab’s memory. • Optional activity counts are off by default and contain counts only. SUPPORTED AI WEBSITES ChatGPT, Claude, Gemini, Microsoft Copilot, and Perplexity. IMPORTANT Detection is heuristic and may occasionally miss sensitive text or flag harmless text. PromptVeil is a review aid, not guaranteed protection or a compliance control. The AI website receives any final text you choose to submit and processes it under its own privacy policy. Add a local review step before sensitive information reaches your AI prompts.