kongyo2/MCP-Server-Search

MCP Server Search

A comprehensive MCP (Model Context Protocol) server that provides extensive search capabilities for discovering MCP servers across GitHub repositories and the Glama.ai curated directory.

🚀 Features

• Comprehensive Search: Searches both GitHub repositories and Glama.ai curated MCP servers
• Maximum Coverage: Optimized to retrieve up to 3,500+ MCP servers within rate limits
• Smart Analysis: Automatically detects and analyzes 21+ MCP frameworks and SDKs
• Confidence Scoring: Provides high/medium/low confidence ratings for server relevance
• No Authentication Required: Works without API keys or authentication tokens

🔍 Search Capabilities

GitHub Repository Search

• 10 comprehensive search queries covering all major MCP frameworks
• Up to 300 results per query (100 per page × 3 pages)
• Theoretical maximum: 3,000 GitHub repositories
• SDK Analysis: Analyzes package.json files for MCP framework usage

Glama.ai Directory Search

• Curated MCP server directory with high-quality servers
• Pagination support for up to 500 servers (100 per page × 5 pages)
• Rich metadata including tools, attributes, and licensing information

Supported MCP Frameworks

The server detects and analyzes the following MCP frameworks:

TypeScript/JavaScript

• FastMCP, LiteMCP, MCP Framework, ModelFetch, Simple MCP, MCPcat

Python

• EasyMCP, FastAPI MCP, MXCP

Java

• Quarkus MCP Server, Spring AI MCP

Go

• Foxy Contexts, MCP-Go, MCP-Golang

Other Languages

• MCP-RS (Rust), Sakura MCP (Scala), MCP Dart, Tidewave Rails/Phoenix, MCPC (C)

🛠 Development

Installation

Installing via Smithery

To install mcp-server-search for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @kongyo2/mcp-server-search --client claude

Installing Manually

git clone <repository-url>
cd mcp-server-search
npm install

Start the server

npm run start

Development with hot reload

npm run dev

Note: The dev command provides an interactive CLI interface for testing the MCP server.

🔧 Available Tools

search-mcp-servers

Searches extensively for MCP servers from both GitHub repositories and Glama.ai curated directory.

Parameters:

• query (string, required): Specific description for the MCP Server needed

Query Guidelines:

• ✅ GOOD Examples:

  • "MCP Server for AWS Lambda Python3.9 deployment"
  • "MCP Server for United Airlines booking API"
  • "MCP Server for Stripe refund webhook handling"
  • "MCP Server for MongoDB database operations"
  • "MCP Server for GitHub repository management"

• ❌ BAD Examples:

  • "MCP Server for cloud" (too vague)
  • "MCP Server for booking" (which booking system?)
  • "MCP Server for payment" (which payment provider?)

Query Requirements:

1. Target platform/vendor (e.g., AWS, Stripe, MongoDB, GitHub)
2. Exact operation/service (e.g., Lambda deployment, webhook handling, database operations)
3. Additional context if applicable (e.g., Python version, specific events)

Example Usage:

{
  "name": "search-mcp-servers",
  "arguments": {
    "query": "MCP Server for AWS Lambda Python3.9 deployment"
  }
}

Response Format:

Found 150 MCP servers total:
- 120 from GitHub repositories
- 30 from Glama.ai curated directory
- 45 high confidence, 80 medium confidence

Server 1:
Name: kongyo2/evetycoon-mcp-server
Description: EVE Online Tycoon MCP Server for market data and trading
Source: GITHUB (github-repo)
GitHub URL: https://github.com/kongyo2/evetycoon-mcp-server
Frameworks: fastmcp
Confidence: MEDIUM

🧪 Testing

The project includes comprehensive tests covering:

• Unit Tests: Core functionality and MCP framework detection
• Integration Tests: Real-world repository analysis
• Enhanced Search Tests: Rate limiting and pagination features

npm run test

Test Coverage:

• 27 test cases covering all major functionality
• Framework detection for 21+ MCP libraries
• GitHub API and Glama.ai API integration testing
• Rate limiting and error handling validation

📊 Performance & Rate Limits

Optimization Features

• Smart Rate Limiting: 50-100ms delays between requests to avoid API limits
• Efficient Pagination: Retrieves maximum results per API call
• Duplicate Filtering: Removes duplicate servers across sources
• Confidence-Based Sorting: Prioritizes high-quality results

Expected Performance

• Search Time: 10-30 seconds for comprehensive searches
• Results Volume: Up to 3,500+ servers depending on query
• Analysis Depth: Up to 100 repositories analyzed for SDK usage

🔧 Development Tools

Linting & Formatting

npm run lint    # Check code quality
npm run format  # Auto-format code

Build

npm run build   # Compile TypeScript

🏗 Architecture

Core Components

1. GitHub Search Engine

   • Multiple optimized search queries
   • Package.json analysis for MCP framework detection
   • Confidence scoring based on SDK usage

2. Glama.ai Integration

   • Curated server directory access
   • Rich metadata extraction
   • Pagination support

3. Unified Result Processing

   • Duplicate removal across sources
   • Confidence-based ranking
   • Framework categorization

MCP Framework Detection

The server uses sophisticated pattern matching to identify MCP servers:

• High Confidence: Official @modelcontextprotocol/sdk usage
• Medium Confidence: Popular frameworks (FastMCP, LiteMCP, etc.)
• Low Confidence: Keyword-based detection in descriptions

📝 Configuration

The server runs with STDIO transport by default, making it compatible with:

• Claude Desktop
• MCP Inspector
• Custom MCP clients

No additional configuration or API keys required.

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Ensure all tests pass: npm test
5. Submit a pull request

📄 License

MIT License - see LICENSE file for details.