MCP Agent 2p6

🚀 MCP Agent Project Documentation

MCP Agent is a multi - purpose control platform designed to offer powerful functionality by integrating various services. This document will elaborate on the project's structure, configuration, and usage.

🚀 Quick Start

1. Clone the project repository:

git clone https://github.com/yourusername/mcp-agent.git

2. Install dependencies:

cd mcp-agent/mcp-agent-app && npm install

3. Configure environment variables by filling in the .env.local file.

4. Start the application:

npm run dev

✨ Features

MCP Agent is a multi - purpose control platform that integrates multiple services to provide a wide range of capabilities.

📦 Installation

Prerequisites

• Ensure you have Node.js and npm installed on your system.

Steps

1. Clone the repository as described in the Quick Start section.
2. Navigate to the mcp-agent-app directory and install the necessary dependencies using npm install.
3. Configure the environment variables in the .env.local file.

💻 Usage Examples

Basic Usage

After starting the application, you can access the main page through your browser. Interact with the chat interface to use the integrated services.

📚 Documentation

Project Components

Core Components

• mcp-agent-app: A Next.js - based front - end application responsible for handling user interactions and API requests.
• MCP Configuration File: Defines the connection parameters of the MCP server and the tool schema.
• LLM Configuration File: Provides the definition of LLM providers and models for the user interface.

Project Structure

mcp-agent/ (Project root directory)
├── .gitignore
├── LICENSE
├── README.md (This file)
├── docs/ (Additional documentation)
├── mcp-agent-app/ (Next.js application)
│   ├── .env.local (API Key, paths, database URL)
│   ├── llm-config.json (LLM provider/model definition for UI)
│   ├── mcp-config.json (MCP server command configuration)
│   ├── next.config.ts
│   ├── package.json
│   ├── prisma/
│   │   └── schema.prisma (Database schema)
│   │   └── migrations/ (Migration files)
│   ├── src/
│   │   ├── app/
│   │   │   ├── api/
│   │   │   │   ├── chat/route.ts (Main chat API, using Vercel AI SDK)
│   │   │   │   └── chats/ (API routes for saving/loading chat records)
│   │   │   └── page.tsx (Homepage component)
│   │   ├── components/ (React components: ChatInterface, Sidebar, etc.)
│   │   ├── context/
│   │   │   └── ChatContext.tsx (Manages application state, integrates useChat hook)
│   │   └── lib/
│   │       └── prisma.ts (Prisma client instance)
│   └── ... (Other Next.js files)
├── duckduckgo-mcp-server/ (DuckDuckGo search MCP server source code)
├── github-mcp-server/ (GitHub MCP server source code, no need to run directly if using Docker image)
├── mcp-gsuite/ (GSuite MCP server source code)
└── whatsapp-mcp/
    ├── whatsapp-bridge/ (Go component - needs to be executed separately)
    └── whatsapp-mcp-server/ (Python component)

Configuration File Explanation

.env.local File

This file stores the environment variables required for the project to run, including:

• API Key: Keys for accessing various services (e.g., GitHub, GSuite).
• Path Configuration: Specifies the locations of each MCP server component.
• Database URL: Defines the database address for the application to connect to.

llm-config.json File

This file provides detailed information about LLM providers and models for the user interface, facilitating selection and configuration.

mcp-config.json File

This file defines the connection parameters of each MCP server, including:

• id: A unique identifier.
• description: A brief description.
• command: The execution command, including optional parameters and environment variable settings.

Current MCP Server Implementations

WhatsApp MCP Server

• Source Code Location: whatsapp-mcp/
• Execution Method: Run through a Go component and a Python script.
• Functionality: Handles WhatsApp messages and interactions.

GitHub MCP Server

• Source Code Location: github-mcp-server/
• Execution Method: Run based on a Docker image.
• Functionality: Integrates the GitHub API to implement automated operations.

GSuite MCP Server

• Source Code Location: mcp-gsuite/
• Execution Method: Run a Python script directly.
• Functionality: Manages Google Workspace services.

DuckDuckGo Search MCP Server

• Source Code Location: duckduckgo-mcp-server/
• Execution Method: Invoked through an HTTP request.
• Functionality: Provides web search results.

Adding a New MCP Server

1. Update the mcp-config.json file:

   • Add a new server entry with its unique identifier and description.
   • Configure the execution command, including optional parameters and environment variables.

2. Implement the server component:

   • Write code according to requirements, ensuring compatibility with the existing architecture.
   • Provide detailed documentation on the functionality and configuration method.

3. Test and integrate:

   • Ensure that the new server communicates properly with the main application.
   • Handle all potential dependencies and error situations.

🔧 Technical Details

MCP Agent Protocol (JSON - RPC)

MCP uses the JSON - RPC protocol for communication. The specific interface definitions are as follows:

• Request Format:

{
    "jsonrpc": "2.0",
    "method": "string",
    "params": "array",
    "id": "integer"
}

• Response Format:

{
    "jsonrpc": "2.0",
    "result": "object",
    "id": "integer"
}

Database Structure

• Chat Record Table: chat_records
  • id: Integer, primary key.
  • content: Text, stores chat content.
  • timestamp: Timestamp, records the creation time.

🤝 Contribution Guidelines

Community contributions are welcome! Please read the CONTRIBUTING.md file before submitting a PR to understand the code standards and submission process.

📄 License

This project is licensed under the MIT License. For detailed information, please refer to the LICENSE file.

⚠️ Important Note

This document is continuously being updated. Please stay tuned for future version releases for more details.