Generic MCP Server

Generic MCP Server

A production-ready Model Context Protocol (MCP) server implementation built with TypeScript and Express, designed to seamlessly connect AI agents with external APIs.

Overview

This project demonstrates how to implement a Server-Sent Events (SSE) MCP server using the @modelcontextprotocol/sdk. It provides developers with a practical reference for exposing APIs to AI agents and Large Language Models (LLMs) with a focus on movie database interactions.

The implementation details can be found in src/mcp/tools.ts, src/mcp/prompts.ts, src/helpers/ directory, and related files.

What is MCP?

The Model Context Protocol (MCP) establishes a standardized interface for AI models to interact with external systems securely and efficiently. It defines a structured framework for:

• Resources: Data objects that models can access
• Tools: Executable functions that models can invoke
• Prompts: Templates that guide model interactions

This standardization simplifies the integration of AI capabilities into existing applications and services.

For comprehensive documentation, visit the Model Context Protocol repository.

Key Features

• Full implementation of the MCP specification
• Real-time communication via Server-Sent Events (SSE)
• Modular architecture for easy extension
• Built-in movie database integration for searching people and discovering films
• Type-safe interfaces for API responses
• Comprehensive error handling
• Example client for testing and reference

Prerequisites

• Node.js 18 or higher
• npm or yarn package manager

Getting Started

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/generic-mcp-server.git
   cd generic-mcp-server
   

2. Install dependencies:

   npm install
   

3. Configure your environment:

   cp .env.example .env
   

4. Add your API credentials to the .env file:

   PORT=8089
   ACCESS_TOKEN=your_movie_db_access_token
   

Running the Server

Development mode with hot reloading:

npm run dev

Production mode:

npm run build
npm start

The server will be available at http://localhost:8089 (or the port specified in your .env file).

API Reference

Endpoints

Endpoint	Method	Description
/sse	GET	Server-Sent Events endpoint for MCP communication
/messages	POST	Endpoint for clients to send messages to the MCP server

MCP Components

Resources

• greeting://welcome: Static welcome message
• users://{userId}: Dynamic user data resource

Tools

The server implements powerful movie database tools:

• movie_database_search_person: Search for actors, directors, and film professionals with detailed information about their careers, biographical data, and filmographies.
• movie_database_discover_films: Advanced film discovery with filtering by cast, genres, release dates, ratings, and more.

Prompts

• greeting: Friendly user greeting
• help: Detailed command assistance
• ask-question: Question template for guiding conversations
• error: Graceful error handling prompt

Movie Database Features

The server provides access to The Movie Database (TMDB) API with the following capabilities:

• Search for film industry professionals including actors and directors
• Discover movies with extensive filtering options:
  • By release date, popularity, or rating
  • By cast or crew members
  • By genres, keywords, or companies
  • By runtime or language
  • By monetization type (free, subscription, rental, etc.)

Environment Configuration

Variable	Description	Default
PORT	Server port number	8089
ACCESS_TOKEN	Authentication token for movie database API	required

Testing with the Example Client

The repository includes a demonstration client that shows how to connect to and interact with the MCP server:

npm run client

The client will:

• Establish a connection to the server
• Display available resources, tools, and prompts
• Read the greeting resource
• Demonstrate tool usage
• Retrieve prompt templates

Integrating with AI Models

This server can be integrated with any LLM that supports the MCP protocol. The standardized interface ensures compatibility with various AI frameworks and platforms.

Architecture

The project follows a modular architecture:

• src/mcp/: Core MCP protocol implementations
• src/helpers/: API-specific helper functions
• src/utils/: Utility functions and shared code
• src/types.ts: TypeScript interface definitions

Available Interfaces

The project includes typed interfaces for API responses:

• ApiResponse<T>: Generic API response wrapper with data, error, and status fields
• ApiConfig: Configuration for API client with baseUrl and optional headers

These interfaces help ensure type safety when working with API data.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT