PasteGuard
PasteGuard is a privacy agent for large model calls. It will automatically detect and desensitize personal information and key content before requests are sent to the OpenAI/Anthropic compatible interface, reducing the risk of sensitive data leakage.
use It is suitable for use in daily AI conversations, code assistants, and corporate intranet call scenarios. It focuses on protecting sensitive information such as name, mailbox, mobile phone number, geographical location, and private key/Token, while retaining normal Q & A and development processes. use method 1. Fill in baseurl and key (upstream model service address and key) when installing the application. 2. Change your client API address to the PasteGuard address (such as https://your domain name/openai/v1 or.../ anthropic）。 3. When the model request is initiated in the original way, PasteGuard will automatically perform detection and desensitization. 4. Open/dashboard to view statistics such as request volume, desensitization hit rate, and scanning time consumption. Backend address: https://Your domain name/dashboard
