zesty-genius128/Certus_server

Certus Drug Information MCP Server

A comprehensive Model Context Protocol (MCP) server providing real-time FDA drug information including shortages, recalls, labels, and market analysis. This server enables AI assistants and other MCP clients to access comprehensive pharmaceutical data using official FDA sources.

Try the Live Chatbot

Test the Certus system immediately: https://certus-chat.opensource.mieweb.org

Access real-time FDA drug information through a ChatGPT-like interface. No setup required - just visit the link and start asking about drug shortages, recalls, and medication information.

Note: If your search does not return results right away, try asking the chatbot to check the identifier type again. This is a known limitation of the current chatbot implementation, not a server issue.

Key Features for Healthcare Professionals

Real-Time Drug Shortage Information

• Search current FDA drug shortage database by generic or brand name
• Multiple search strategies to handle name variations and misspellings
• Raw FDA shortage data with minimal processing for accuracy
• Returns shortage reasons, resolution estimates, and affected product details when available

Comprehensive FDA Drug Label Data

• Complete FDA-approved prescribing information and structured product labeling
• Generic and brand name cross-referencing through openFDA identifiers
• Dosage forms, strengths, routes of administration, and manufacturer information
• Active ingredients, indications, contraindications, warnings, and adverse reactions

Drug Recall and Safety Information

• Real-time search of FDA enforcement database for drug recalls
• Recall classification (Class I, II, III), reason for recall, and affected lot information
• Product descriptions, distribution information, and recall initiation dates
• Raw FDA enforcement data for comprehensive recall analysis

Historical Data Analysis

• Drug shortage pattern analysis using historical FDA shortage records
• Configurable time periods for trend analysis (1-60 months)
• Calculated trend metrics including shortage frequency, duration, and patterns
• Real-time trend analysis with meaningful insights from FDA historical data

Batch Processing Capabilities

• Simultaneous analysis of up to 25 medications in a single request
• Combined shortage, recall, and optional trend data for each drug
• Efficient formulary-wide assessment with single API call
• Structured results for each drug in the batch with error handling

Clinical Information Integration

• Combined medication profiles with both FDA label and current shortage data
• Intelligent search across multiple FDA identifier types (generic name, brand name, openFDA fields)
• Cross-referenced data from FDA drug label and shortage databases
• Streamlined access to comprehensive FDA medication information

FDA Adverse Event Reporting (FAERS)

• Search comprehensive FDA adverse event database for reported side effects and reactions
• Filter for serious adverse events only (hospitalization, death, disability, life-threatening conditions)
• Raw FAERS data with patient demographics, reaction terms, and safety report details
• Both general adverse events and serious events filtering with medical disclaimers

Note: Adverse events queries may return large datasets. For comparative queries, use individual drug searches rather than asking to compare multiple drugs simultaneously to avoid token limits.

How to Use the Chatbot

Visit https://certus-chat.opensource.mieweb.org and ask questions like:

Drug Shortage Queries:

• "Check current shortage status for insulin"
• "Search for metformin shortages"
• "Show shortage information for lisinopril"
• "Find shortage data for acetaminophen"

Medication Label Information:

• "Get FDA label information for metformin"
• "Show prescribing information for atorvastatin"
• "What are the approved uses for hydrochlorothiazide"
• "Get dosage forms available for ibuprofen"

Drug Recall Searches:

• "Search for recalls involving insulin products"
• "Check for any metformin recalls"
• "Find recent recalls for blood pressure medications"
• "Search recall database for acetaminophen"

Complete Medication Profiles:

• "Get complete medication profile for lisinopril"
• "Show combined label and shortage data for insulin"
• "Comprehensive information for metformin"

Trend Analysis:

• "Analyze shortage patterns for insulin over 12 months"
• "Show shortage trends for metformin in the past 6 months"
• "Historical shortage data for aspirin"

Batch Analysis:

• "Analyze these drugs for shortages: insulin, metformin, lisinopril"
• "Check shortage status for multiple diabetes medications"
• "Batch analysis of cardiovascular drugs with trend data"

Adverse Events Queries:

• "What adverse events have been reported for aspirin?"
• "Show me serious adverse events for warfarin"
• "Search for side effects of metformin in the FDA database"
• "Has the FDA received reports of serious reactions to atorvastatin?"

Integration Options for Healthcare Systems

Web-Based Chatbot Interface

Access the live chatbot at https://certus-chat.opensource.mieweb.org for immediate drug information queries without any setup required.

Claude Desktop Integration

Add Certus tools directly to Claude Desktop for integrated AI assistant workflow with FDA drug data access.

Custom MCP Client Development

Deploy the Certus server with any MCP-compatible client including:

• LibreChat for custom healthcare chatbots
• VS Code with MCP extensions for development environments
• Custom applications using MCP client libraries
• Third-party MCP-compatible tools and platforms

Direct API Integration

Use REST API endpoints for custom healthcare applications:

• HTTP POST to /mcp endpoint for JSON-RPC tool calls
• Direct tool endpoints at /tools/{tool_name} for simple HTTP integration
• Health monitoring via /health endpoint
• Real-time data integration with existing systems

Universal MCP Client Compatibility

Why Certus Works Everywhere

Certus uses a universal HTTP-based architecture with stdio transport bridges, making it compatible with all major MCP clients including Claude Desktop, VS Code, Cursor, Visual Studio, Windsurf, and LibreChat.

Key Advantages:

• Universal Compatibility - Works with any MCP client that supports stdio transport
• Production-Ready Hosting - Always-on server with HTTPS, health monitoring, and scaling
• No Client Dependencies - Clients don't need to manage server lifecycle or installations
• Easy Debugging - Standard HTTP endpoints for testing with curl, Postman, or browsers
• Simple Updates - Server updates are deployed centrally without client-side changes

Quick Start - Add to Any MCP Client

Step 1: Universal Configuration (Recommended)

This configuration works with all MCP clients (Claude Desktop, VS Code, Cursor, Visual Studio, Windsurf):

Configuration file locations:

• Claude Desktop (macOS): ~/Library/Application Support/Claude/claude_desktop_config.json
• Claude Desktop (Windows): %APPDATA%\Claude\claude_desktop_config.json
• VS Code: .vscode/mcp.json (workspace) or global settings
• Cursor: ~/.cursor/mcp.json or .cursor/mcp.json (project)
• Visual Studio: <SOLUTIONDIR>\.mcp.json

{
  "mcpServers": {
    "Certus": {
      "command": "npx",
      "args": ["mcp-remote", "https://certus.opensource.mieweb.org/mcp"]
    }
  }
}

Step 2: Backup Option (If npx mcp-remote doesn't work)

If the npx mcp-remote method doesn't work with your specific MCP client setup, use our stdio wrapper:

1. Download the stdio wrapper: Get stdio-wrapper.js from our repository
2. Update your config to point to the wrapper file:

{
  "mcpServers": {
    "Certus": {
      "command": "node",
      "args": ["/path/to/your/stdio-wrapper.js"]
    }
  }
}

Note: Replace /path/to/your/stdio-wrapper.js with the actual path where you saved the wrapper file.

Alternative Configuration (Backup Server)

If the main server is unavailable, use the Railway backup:

{
  "mcpServers": {
    "Certus": {
      "command": "npx",
      "args": ["mcp-remote", "https://certus-server-production.up.railway.app/mcp"]
    }
  }
}

Step 2: Restart the App

Close and reopen Claude Desktop/the MCP client completely. The drug information tools should now be available.

Step 3: Test the Integration

In Claude Desktop, try asking:

• "Check for insulin shortages in the FDA database"
• "Get complete medication profile for metformin"
• "Search for drug recalls involving acetaminophen"
• "Analyze shortage trends for lisinopril"
• "Show me FDA label information for atorvastatin"

System Architecture

Sequence Diagram

sequenceDiagram
    participant User
    participant ClaudeDesktop["Claude Desktop"]
    participant LibreChat
    participant mcpRemote1["npx mcp-remote"]
    participant StdioWrapper["stdio-wrapper.js"]
    participant CertusMCP["Certus MCP Server"]
    participant FDAAPIs["FDA APIs"]

    Note over User,FDAAPIs: Healthcare-Focused MCP Architecture

    %% Claude Desktop Flow
    rect rgb(240, 248, 255)
        Note over ClaudeDesktop,mcpRemote1: Claude Desktop Integration
        User->>ClaudeDesktop: "Check insulin shortage status"
        ClaudeDesktop->>mcpRemote1: stdio JSON-RPC tool call
        mcpRemote1->>CertusMCP: HTTP POST /mcp
        Note over CertusMCP: search_drug_shortages tool
        CertusMCP->>FDAAPIs: Query FDA shortage database
        FDAAPIs-->>CertusMCP: Raw FDA shortage data
        CertusMCP-->>mcpRemote1: MCP JSON-RPC response
        mcpRemote1-->>ClaudeDesktop: stdio response
        ClaudeDesktop-->>User: "No current insulin shortages found"
    end

    %% LibreChat Flow
    rect rgb(255, 248, 248)
        Note over LibreChat,StdioWrapper: LibreChat Integration
        User->>LibreChat: "What recalls for metformin?"
        LibreChat->>StdioWrapper: stdio JSON-RPC tool call
        StdioWrapper->>CertusMCP: HTTP POST /mcp
        Note over CertusMCP: search_drug_recalls tool
        CertusMCP->>FDAAPIs: Query FDA enforcement database
        FDAAPIs-->>CertusMCP: Raw FDA recall data
        CertusMCP-->>StdioWrapper: MCP JSON-RPC response
        StdioWrapper-->>LibreChat: stdio response
        LibreChat-->>User: "3 metformin recalls found..."
    end

Architecture Components

Frontend Layer:

• Claude Desktop - Primary MCP client for healthcare professionals and patients seeking drug information
• LibreChat Interface - ChatGPT-like web interface at https://certus-chat.opensource.mieweb.org for public access

Integration Layer:

• npx mcp-remote - Universal stdio-to-HTTP bridge for Claude Desktop client integration
• stdio-wrapper.js - Custom stdio bridge specifically for LibreChat integration
• Transport Compatibility - Handles protocol translation between stdio and HTTP MCP transports

Backend Layer:

• Certus MCP Server - Express.js server implementing MCP 2024-11-05 protocol with HTTP JSON-RPC
• OpenFDA Client - Intelligent API client with multiple search strategies and error handling
• FDA Data Sources - Drug Shortages, Labels, Enforcement, and Adverse Events (FAERS) databases

Healthcare-Focused Data Flow:

1. Users ask healthcare-related questions through Claude Desktop or LibreChat interface
2. Tool calls are routed through appropriate transport bridges (mcp-remote or stdio-wrapper) to HTTP endpoint
3. MCP server executes FDA API calls with intelligent fallback strategies and multiple search methods
4. Raw FDA data is returned with minimal processing to preserve medical accuracy and regulatory compliance
5. AI clients analyze and present medical information with appropriate disclaimers and safety warnings

Available Tools

Core Drug Information Tools

search_drug_shortages

Search the FDA drug shortages database using intelligent matching strategies.

Parameters:

• drug_name (string, required): Name of the drug (generic or brand name)
• limit (integer, optional): Maximum results to return (1-50, default: 10)

Returns: Raw FDA shortage data including shortage reasons, estimated resolution dates, and affected products.

get_medication_profile

Get comprehensive drug information combining FDA label data and current shortage status.

Parameters:

• drug_identifier (string, required): Drug name or identifier
• identifier_type (string, optional): Type of identifier (default: "openfda.generic_name")

Returns: Combined medication profile with both FDA label information and shortage data.

search_drug_recalls

Search the FDA enforcement database for drug recalls using multiple search strategies.

Parameters:

• drug_name (string, required): Drug name to search for recalls
• limit (integer, optional): Maximum results (1-50, default: 10)

Returns: Raw FDA enforcement data including recall classifications, reasons, and affected products.

get_drug_label_info

Retrieve FDA-approved drug label information and prescribing data.

Parameters:

• drug_identifier (string, required): Drug identifier
• identifier_type (string, optional): Type of identifier (default: "openfda.generic_name")

Returns: Complete FDA structured product labeling data.

Advanced Analysis Tools

analyze_drug_market_trends

Analyze drug shortage patterns using historical FDA shortage data.

Parameters:

• drug_name (string, required): Drug name to analyze
• months_back (integer, optional): Analysis period in months (1-60, default: 12)

Returns: Comprehensive trend analysis with calculated metrics including shortage frequency, duration patterns, and historical timeline analysis using FDA shortage data.

search_adverse_events

Search the FDA adverse event reporting system (FAERS) for reported side effects and reactions.

Parameters:

• drug_name (string, required): Name of the drug to search for adverse events
• limit (integer, optional): Maximum adverse event reports to return (1-50, default: 5)
• detailed (boolean, optional): Return full raw FDA data (true) or summarized data (false). Default false for better performance.

Returns: Raw FDA FAERS data including patient demographics, reaction terms, seriousness indicators, and safety report details.

search_serious_adverse_events

Search for serious adverse events only (hospitalization, death, disability, life-threatening conditions).

Parameters:

• drug_name (string, required): Name of the drug to search for serious adverse events
• limit (integer, optional): Maximum serious adverse event reports to return (1-50, default: 5)
• detailed (boolean, optional): Return full raw FDA data (true) or summarized data (false). Default false for better performance.

Returns: Raw FDA FAERS data filtered for serious outcomes with safety warnings and report classifications.

batch_drug_analysis

Simultaneously analyze multiple drugs for shortages, recalls, and optionally trends.

Parameters:

• drug_list (array, required): List of drug names (max 25 drugs)
• include_trends (boolean, optional): Include trend analysis (default: false)

Returns: Array of analysis results containing shortage data, recall data, and optionally trend data for each drug.

Example Usage in Claude

Search for current shortages of insulin and show me the raw FDA data

Get the complete FDA medication profile for metformin including label and shortage information

Search the FDA enforcement database for any recalls involving acetaminophen

Analyze shortage patterns for lisinopril using historical FDA data from the past 6 months

Perform batch analysis on these drugs: insulin, metformin, lisinopril, aspirin - include trend data

Get FDA label information for atorvastatin and check current shortage status

Search for adverse events reported to FDA for aspirin

Show me only serious adverse events for warfarin that resulted in hospitalization or death

Get FDA adverse event data for metformin and check for any lactic acidosis reports

Server Infrastructure

Primary Deployment (Proxmox)

• Main Server: https://certus.opensource.mieweb.org/mcp
• Status Check: https://certus.opensource.mieweb.org/health
• Host: certus.opensource.mieweb.org
• Service: OpenFDA MCP Server
• Tools Available: 8 FDA drug information tools
• Protocol: HTTPS with CORS enabled (MCP Streamable HTTP 2024-11-05)

Backup Deployment (Railway)

• Backup Server: https://certus-server-production.up.railway.app/mcp
• Status Check: https://certus-server-production.up.railway.app/health
• Host: Railway cloud platform

API Health Status

Current FDA API endpoint availability:

• Drug Labels Database: Operational (HTTP 200)
• Drug Shortages Database: Operational (HTTP 404 responses indicate no current shortages, which is normal)
• Drug Enforcement Database: Operational (HTTP 200)
• API Key Status: Not configured (using public rate limits of 1,000 requests/day)

Deploy Your Own Server

Prerequisites

• Node.js 18+
• Railway account (free) or your own hosting
• Git

Step 1: Clone and Setup

git clone https://github.com/zesty-genius128/Certus_server.git
cd Certus_server
npm install

Step 2: Test Locally

# Start the server
npm start

# Test in another terminal
curl http://localhost:3000/health

# Test drug search
curl -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "search_drug_shortages",
      "arguments": {"drug_name": "insulin", "limit": 5}
    }
  }'

Step 3: Deploy to Railway (Backup Option)

# Install Railway CLI
npm install -g @railway/cli

# Login to Railway
railway login

# Initialize project
railway init

# Deploy
railway up

# Get your URL
railway domain

Step 4: Deploy to Your Own Infrastructure

For Proxmox or other self-hosted environments:

# Copy files to your server
scp -r . user@your-server:/path/to/certus-server/

# On your server
cd /path/to/certus-server
npm install
npm start

# For production with PM2
npm install -g pm2
pm2 start official-mcp-server.js --name certus-server
pm2 save
pm2 startup

Step 5: Update Claude Config

Replace the URL in your Claude config with your deployed URL:

{
  "mcpServers": {
    "Certus": {
      "command": "npx",
      "args": ["mcp-remote", "http://your-server.com:3000/mcp", "--allow-http"]
    }
  }
}

Note: The --allow-http flag is only required if you are using an http URL. If your server URL starts with https, you do not need to include this flag.

Configuration

Environment Variables (Optional)

Create a .env file for enhanced API performance:

OPENFDA_API_KEY=your_fda_api_key_here
PORT=3000

Get a free FDA API key at: https://open.fda.gov/apis/authentication/

Rate Limits

• Without API Key: 1,000 requests/day
• With API Key: 120,000 requests/day
• Intelligent Caching: Reduces API calls through smart result caching

Testing and Debugging

Test with MCP Inspector

# Test main server
npx @modelcontextprotocol/inspector https://certus.opensource.mieweb.org/mcp

# Test backup server
npx @modelcontextprotocol/inspector https://certus-server-production.up.railway.app/mcp

# Test local development
npx @modelcontextprotocol/inspector http://localhost:3000/mcp

Health Check Commands

# Check main server status
curl https://certus.opensource.mieweb.org/health

# Check backup server status
curl https://certus-server-production.up.railway.app/health

# Get available tools
curl https://certus.opensource.mieweb.org/tools

Direct API Testing

# Test drug shortage search
curl -X POST https://certus.opensource.mieweb.org/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "search_drug_shortages", 
      "arguments": {"drug_name": "insulin", "limit": 3}
    }
  }'

# Test medication profile
curl -X POST https://certus.opensource.mieweb.org/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 2,
    "method": "tools/call",
    "params": {
      "name": "get_medication_profile",
      "arguments": {"drug_identifier": "metformin"}
    }
  }'

Project Structure

Certus_server/
├── official-mcp-server.js    # Main MCP server implementation
├── openfda-client.js         # FDA API integration and data processing
├── stdio-wrapper.js          # Local development and testing support
├── package.json              # Dependencies and scripts
├── claude_desktop_config.json # Example Claude configuration
└── README.md                 # This documentation

API Endpoints

OpenFDA API Endpoints Used

Certus integrates with the following official FDA openFDA API endpoints to provide comprehensive drug information:

Base URL: https://api.fda.gov

Authentication: Optional API key for higher rate limits (1,000 requests/day without key, 120,000 requests/day with key)

Rate Limiting: Automatically managed by Certus server with intelligent request strategies

Data Format: All endpoints return raw FDA JSON data with minimal processing to preserve accuracy

Official Documentation: FDA openFDA APIs

Advanced Features

Intelligent Drug Matching

• Multiple search strategies for drug name variations (generic, brand, openFDA fields)
• Automatic fallback search methods when initial queries return no results
• Handles common misspellings and name variations automatically
• Search strategy reporting shows which method successfully found results

Comprehensive Data Integration

• FDA Drug Shortages Database: Real-time shortage information with reasons and estimated resolution
• FDA Drug Labels Database: Complete structured product labeling and prescribing information
• FDA Enforcement Database: Drug recall and safety information with classification details
• FDA Adverse Event Reporting System (FAERS): Comprehensive adverse event and safety data with reaction classifications

Minimal Processing Architecture

• Raw FDA API responses preserved with minimal post-processing
• Original openFDA JSON structure maintained for accurate data representation
• Reduced data transformation overhead for better performance
• Enhanced for AI analysis with structured but unmodified FDA data

Batch Processing

• Analyze up to 25 drugs simultaneously in a single request
• Combined shortage, recall, and optional trend analysis
• Efficient API usage with intelligent request batching
• Formulary-wide assessment capabilities

Troubleshooting

Common Issues

• Tool not appearing in Claude:
  • Restart Claude Desktop completely
  • Verify config file syntax is correct
  • Check that the server URL is accessible
• Connection errors:
  • Test server health: curl https://certus.opensource.mieweb.org/health
  • Try backup server if main is down
  • Check firewall/network connectivity
• No results found:
  • Try different drug name variations (generic vs brand)
  • Check spelling of drug names
  • Verify drug exists in FDA database
• Rate limit issues:
  • Add OpenFDA API key to environment variables
  • Reduce request frequency
  • Use batch operations for multiple drugs

Debug Commands

# Check server status and capabilities
curl https://certus.opensource.mieweb.org/health

# View available tools
curl https://certus.opensource.mieweb.org/tools

# Test specific tool functionality
npm run test

# MCP protocol testing
npx @modelcontextprotocol/inspector https://certus.opensource.mieweb.org/mcp

Comprehensive Test Suite

The project includes a comprehensive test suite with 63 test cases covering all FDA tools:

• File: tests/comprehensive-test.js
• Coverage: All 8 FDA drug information tools
• Test Categories: Trend analysis, core tools, batch analysis, performance, error handling
• Framework: Custom assertion-based testing with error tracking
• Results: 63 passed, 0 failed (verified July 15, 2025)

# Run the comprehensive test suite
node tests/comprehensive-test.js

Use Cases

Healthcare Applications

• Real-time drug shortage monitoring and information retrieval
• FDA drug label data integration for clinical systems
• Drug recall notification and safety monitoring
• Medication information lookup for clinical decision support

AI Assistant Integration

• Enhanced medical chatbots with FDA drug data access
• Clinical workflow automation with drug information tools
• Healthcare assistant applications with real-time drug data
• Patient information systems with shortage and recall alerts

Research and Analytics

• Pharmaceutical shortage pattern analysis using historical FDA data
• Drug recall trend research and safety monitoring
• Supply chain analysis using FDA shortage data
• Regulatory compliance monitoring with automated FDA data access

Technical Specifications

• Protocol: Model Context Protocol (MCP) 2024-11-05
• Data Sources: FDA openFDA APIs (Drug Shortages, Labels, Enforcement, Adverse Events)
• Server: OpenFDA MCP Server v2.0.0
• Node.js: 18+ required
• Dependencies: Express, CORS, Helmet, Compression, MCP SDK
• Response Format: JSON-RPC 2.0 with raw FDA data
• Transport: HTTP POST (JSON-RPC) with stdio bridge compatibility
• Rate Limiting: FDA API public limits (1,000 requests/day without API key)
• Tools Available: 8 FDA drug information tools
• CORS: Enabled for cross-origin requests

Resources

• Model Context Protocol Documentation
• OpenFDA API Documentation
• Certus Server Repository
• Railway Deployment Guide

License

MIT License

Live Chatbot: https://certus-chat.opensource.mieweb.org
Primary Server: https://certus.opensource.mieweb.org/mcp
Backup Server: https://certus-server-production.up.railway.app/mcp
Status: Production Ready
Protocol: MCP 2024-11-05
Data Sources: FDA Drug Shortages, Labels, Enforcement, and Adverse Events (FAERS) Databases