ShadowAI Guard

Checks your prompt for sensitive data before it is sent to an AI tool. Runs fully on your device.

ShadowAI Guard checks a prompt for common sensitive information when you try to send it on a supported AI chat website. It can flag text such as email addresses, payment card numbers, and access keys. If it finds a possible match, you can redact the detected text, cancel to review the prompt, or send it unchanged when that option is available. Redaction replaces detected text with placeholders. After sending a redacted prompt, Restore values can replace placeholders in visible page text, including replies. Scanning runs in your browser. The extension does not upload your prompt to the developer. The chat website may handle drafts and sent prompts under its own privacy policy. Pattern matching can miss sensitive text or flag harmless text, so review your prompt before sending.