Skip to content
Conduit

mcp server

pharmaguard

A Model Context Protocol (MCP) server providing real-time openFDA drug adverse event analytics, active recall tracking, and label warnings.

0xh1tesh/pharmaguardcommunitylistedunreviewed

Set up

Runs on your computer. Pick your agent and operating system.

Before you start

  • Node.js 20 or newer
~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "pharmaguard": {
      "command": "npx",
      "args": [
        "-y",
        "pharmaguard-mcp"
      ]
    }
  }
}

same format Cursor, GitHub Copilot CLI, Junie, JetBrains AI Assistant, Augment (Auggie CLI), Factory Droid, Kiro, Devin CLI +12

Values in angle brackets are placeholders. Restart Claude Desktop after editing its config.Claude Desktop docs

README

synced Oct 26, 2026

PharmaGuard MCP

Project Overview

PharmaGuard MCP is an advanced Model Context Protocol (MCP) server built with NitroStack. It provides AI agents (like Claude Desktop or Cursor IDE) with direct access to official FDA open data. The server enables AI to look up drug details, retrieve adverse event reports (FAERS), check active FDA enforcement recalls, and compare drug safety profiles side-by-side.

Features

  • NDC Lookup: Retrieve detailed National Drug Code (NDC) information for generic and brand name drugs.
  • Adverse Events Search: Analyze top adverse reactions (FAERS) associated with a specific medication.
  • FDA Label Cross-Reference: Automatically compares reported adverse reactions against the official FDA Drug Label to determine if a reaction is a known warning or potentially unlisted.
  • Historical Trends & Seriousness: Visualizes the timeline of adverse event reports and provides a clinical seriousness breakdown (Death, Hospitalization, etc.).
  • Enforcement Recalls: Checks the FDA enforcement database for active, ongoing recalls related to a drug.
  • Drug Comparison: Compare two medications side-by-side for their adverse event profiles and recall statuses.
  • Dynamic Widgets: Renders interactive, highly-visual React widgets directly within supported MCP clients (like NitroStudio), featuring Recharts data visualizations.

Pipeline Diagram

Installation

  1. Clone the repository:

    git clone <your-repo-url>
    cd pharmaguard-mcp
    
  2. Install dependencies:

    npm install
    

Environment Variables

Copy the provided .env.example file to create your own .env:

cp .env.example .env

Ensure the following variables are set in your .env file:

  • OPENFDA_API_KEY: Your openFDA API key. Get one at https://open.fda.gov/apis/authentication/ to increase your rate limits from 240/min to 240,000/day.
  • PORT: The port the server will run on (default is 3000).

Note: Do not commit your .env file to version control.

Build Commands

To build the MCP server and compile the React widgets for production:

npm run build

(This compiles the TypeScript code into dist/ and bundles the widgets into src/widgets/out/)

For development with hot-reloading:

npm run dev

To run smoke tests to verify the tools are functioning correctly:

npm run test:smoke

Deployment Instructions

Running the Server

You can start the production server using:

npm start

Integrating with MCP Clients

For NitroStudio:

  1. Open NitroStudio.
  2. Select "Add Project" and point it to the pharmaguard-mcp directory.
  3. The server will automatically start and render the custom React widgets.

For Claude Desktop: Update your claude_desktop_config.json with the following configuration:

{
  "mcpServers": {
    "pharmaguard": {
      "command": "node",
      "args": ["/absolute/path/to/pharmaguard-mcp/dist/index.js"],
      "env": {
        "OPENFDA_API_KEY": "your_api_key_here"
      }
    }
  }
}

For Cursor IDE:

  1. Navigate to Settings -> MCP Settings.
  2. Add a new MCP server.
  3. Select "Command" and use: node /absolute/path/to/pharmaguard-mcp/dist/index.js.
  4. Ensure your environment variables are configured within the Cursor environment if necessary.
↑↓ navigate openesc close