Humane Proxy

Humane Proxy is a lightweight, plug-and-play AI safety middleware designed to intercept and respond to messages containing self-harm or criminal intent before they reach your underlying LLM. It alerts operators via channels like Slack, Discord, PagerDuty, or email, while responding empathetically to users and providing crisis resources where appropriate. Humane Proxy protects human users from harmful interactions with AI, supports robust privacy practices, and is highly configurable as a reverse proxy server, Python library, or Model Context Protocol (MCP) server, making it accessible for AI and platform developers wanting to add a human safety layer to their AI applications.