MCP Server Overview

​

🤖 What is the MCP Server?

​

🎯 Key Features

​

MCP Protocol Support

• Full Model Context Protocol implementation
• Compatible with Claude, Cursor, and other MCP clients
• Standardized tool interface for AI agents

​

Customer Insight Tools

• list_insights - Retrieve insights with filtering and pagination
• get_insight_detail - Get detailed analysis of specific insights
• search_insights - Full-text search across all insight fields
• Rich metadata and AI analysis results

​

🔧 Technical Specifications

​

Server Details

• URL: https://mcp.closedloop.sh
• Protocol: Model Context Protocol (MCP)
• Authentication: Bearer token (API key)
• Data Format: JSON responses

​

Dependencies

• Node.js 18+
• PostgreSQL database
• MCP SDK (@modelcontextprotocol/sdk)
• Express.js for HTTP endpoints

​

🚀 Quick Start

​

1. Get Your API Key

# API keys are managed through the ClosedLoop web interface
# Visit: https://app.closedloop.sh/api-keys

​

2. Connect Your AI Client

{
  "mcpServers": {
    "closedloop": {
      "command": "npx",
      "args": ["@closedloop-ai/mcp-client"],
      "env": {
        "CLOSEDLOOP_API_KEY": "your-api-key-here"
      }
    }
  }
}

​

3. Use the Tools

{
  "name": "list_insights",
  "arguments": {
    "limit": 10,
    "date_from": "2024-01-01"
  }
}

​

📊 Available Tools

​

list_insights

• date_from (optional): Start date (YYYY-MM-DD)
• date_to (optional): End date (YYYY-MM-DD)
• page (optional): Page number (default: 1)
• limit (optional): Items per page (default: 20, max: 100)

​

get_insight_detail

• insight_id (required): UUID of the insight item

​

search_insights

• Full-text search across 8+ insight fields
• Relevance ranking - most relevant results first
• Phrase search and fuzzy matching
• Field-specific search - search specific fields or all fields
• Advanced filtering - category, severity, date range, source
• Language agnostic - works with any language

• query (required): Search query text
• fields (optional): Specific fields to search
• category (optional): Filter by insight category
• severity (optional): Filter by severity level
• date_from (optional): Start date filter
• date_to (optional): End date filter
• source (optional): Filter by data source
• page (optional): Page number (default: 1)
• limit (optional): Items per page (default: 20, max: 100)

​

🎯 Use Cases

​

AI-Powered Analysis

• Claude can analyze customer insight patterns
• Cursor can provide insights during development
• Custom AI agents can process insight data

​

Development Integration

• Access insight data in your development environment
• Get customer insights while coding
• Integrate insights into your workflow

​

Automated Reporting

• Generate insight summaries
• Track sentiment trends
• Monitor customer satisfaction

​

🔗 Integration Examples

​

Claude Desktop

{
  "mcpServers": {
    "closedloop": {
      "command": "npx",
      "args": ["@closedloop-ai/mcp-client"],
      "env": {
        "CLOSEDLOOP_API_KEY": "your-api-key"
      }
    }
  }
}

​

Cursor IDE

{
  "mcpServers": {
    "closedloop": {
      "command": "npx", 
      "args": ["@closedloop-ai/mcp-client"],
      "env": {
        "CLOSEDLOOP_API_KEY": "your-api-key"
      }
    }
  }
}

​

🛠️ Development

​

Local Setup

# Clone the repository
git clone https://github.com/closed-loop-ai/mcp-server

# Install dependencies
npm install

# Set environment variables
cp env.example .env

# Start development server
npm run dev

​

Testing

# Run tests
npm test

# Test MCP connection
npm run test:mcp

​

📞 Support

​

Documentation

• Tool Reference - Complete tool documentation
• Integration Guide - Setup and configuration
• API Reference - REST API documentation

​

Contact

• Technical Support: support@closedloop.sh
• Security Issues: security@closedloop.sh
• Health Check: https://mcp.closedloop.sh/health