redacta

A privacy boundary for clinical text. Pseudonymises patient identifiers before the AI sees them, keeps the reversal map out of the model's context, then checks the output for leaks. MCP server, Agent Skill, CLI and self-hosted Kubernetes service.

Installation

MCP client config (paste into mcp.json)

{
  "mcpServers": {
    "redacta-mcp": {
      "command": "npx",
      "args": [
        "redacta-mcp"
      ]
    }
  }
}

Agent install prompt

Copy and paste into Cursor / Claude / other agents to install this MCP for you

Please read this MCP repository and install it into my current environment:
https://github.com/nickjlamb/redacta

Known working client config (can write into mcp.json):
{
  "mcpServers": {
    "redacta-mcp": {
      "command": "npx",
      "args": [
        "redacta-mcp"
      ]
    }
  }
}

Please: 1) check README for deps and env vars; 2) write config for my client; 3) if secrets are missing, tell me what to fill — do not invent values.

GitHub

Related servers