FHIR standardized how health data is structured. MCP standardized how AI connects to tools. Nobody standardized the guardrails in between. This project does.
Quick Start · MCP Tools ·
MCP.Pizza Chef: aks129
HealthClawGuardrails acts as a security layer between AI assistants and your clinical health data. It automatically removes sensitive information, keeps detailed audit logs, requires extra approvals for sensitive actions, and isolates data per user. It works with apps like Claude Desktop and ChatGPT and can be set up by copying a small config snippet. You can try a public demo with fake data without any account or keys.
Other
Use This MCP server To
Search my health records securely with AI Redact sensitive info from clinical data automatically Audit all AI access to my health data Require extra approval for sensitive data access Isolate data access per user or tenant Connect AI agents safely to FHIR health servers
README
HealthClawGuardrails FAQ
Can I use this to safely access and query my clinical health records?
Yes — HealthClawGuardrails adds security layers like PHI redaction and audit to protect your data when AI agents access FHIR health records.
Can I use this with Claude Desktop or ChatGPT?
Yes — it supports both Claude Desktop and ChatGPT clients for secure healthcare data access.
Do I need an API key or special account to try the demo?
No — there is a public demo server with synthetic data you can use without any keys or accounts.
How hard is it to set up HealthClawGuardrails?
Setup is simple — you can add it by copying and pasting a small config snippet into your app settings.
Does it work with other AI models besides OpenAI?
Yes — it supports OpenAI, Gemini, and Claude among others for AI interactions.
Is this a standalone product I can use as is?
It's an open-source server you configure to add security guardrails between AI and clinical data, not a finished app itself.