Privent

Silent thin-client extension for Privent prompt detection.

Privent analyzes text entered into AI tools such as ChatGPT, Claude, and Gemini to identify potentially sensitive information at the time of submission. The extension captures prompt content at submit-time and sends relevant data to a configured backend service for analysis. Based on the response, it provides risk evaluation signals. Features: - Captures prompt input at submission time - Sends data to backend for security analysis - Detects common sensitive data patterns (API keys, personal data, tokens) - Returns risk evaluation results Important disclosures: - This extension reads user input on supported AI websites - Prompt data may be transmitted to a backend service for analysis - The extension does not provide blocking by itself; it only enables analysis and signaling - Prompt content is not stored permanently by the extension