Humane Proxy
A safety middleware that sits between your users and any LLM, scanning prompts for self-harm ideation and criminal intent before they reach the model. Exposes three MCP tools: check_message_safety for realtime scanning, get_session_risk for per-user scoring, and list_recent_escalations for audit trails. Works as a standalone reverse proxy, Python library, or stdio MCP server. Uses a three-stage cascade starting with regex and keyword heuristics, escalating to semantic embeddings (sentence-transformers), and finally LlamaGuard or OpenAI's moderation API when ambiguous. Blocks flagged messages and returns empathetic crisis resources instead of letting them through. Supports bearer token auth when exposing the HTTP MCP endpoint beyond localhost. Install the ml extra for Stage 2 embeddings or provide GROQ_API_KEY for the reasoning layer.
- GitHub stars
- 28
- tools
- 3
- auth
- Secret needed
- transport
- STDIO
- env vars
- 2
- last updated
- Jul 6, 2026
Tools
3 toolsPublic tool metadata for what this MCP can expose to an agent.
check_message_safetyCheck a message for self-harm or criminal intent.
Check a message for self-harm or criminal intent.
No parameter schema in public metadata yet.
get_session_riskGet the risk trajectory for a session.
Get the risk trajectory for a session.
No parameter schema in public metadata yet.
list_recent_escalationsList recent escalation events.
List recent escalation events.
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageConfiguration
OPENAI_API_KEYsecretOpenAI API key for Stage 3 reasoning (optional)
GROQ_API_KEYsecretGroq API key for LlamaGuard Stage 3 (optional)
Readme
Lightweight, plug-and-play AI safety middleware that protects humans.
HumaneProxy sits between your users and any LLM. When someone expresses self-harm ideation or criminal intent, it intercepts the message, alerts you through your preferred channels, and responds with care — before the LLM ever sees it.
Details
- Repository
- github.com/vishisht16/humane-proxy
- Category
- AI & LLM Tools
- Package
humane-proxy- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active
Related AI & LLM Tools MCP servers
View all
Ai Cli Mcpmkxultra/ai-cli-mcpRun Claude, Codex, Gemini, Forge, and OpenCode CLIs through MCP with background jobs27stars
You.com Web Access & AIyoudotcom-oss/mcpWeb search, AI agent, and content extraction via You.com APIs0stars
Olostep MCP Serverolostep/olostep-mcp-serverLive web search, scraping and crawling API for AI. Get structured, LLM-ready data from any site.25stars
Chart Library — Market-state researchgrahammccain/chart-library-mcpMarket memory for AI: historical states, published studies and source evidence.23stars
Unleash Mcpunleash/unleash-mcpMCP server for Unleash: an AI-agent-friendly interface to Unleash21stars
Codex Mcp Toolx51xxx/codex-mcp-toolMCP server bridging AI assistants to OpenAI Codex CLI for code analysis and review21stars