
MCP Host Server
A multi-tenant remote server platform that enables applications to connect to Model Context Protocol servers over WebSocket connections, allowing secure access to file management, database, and API capabilities.
README
MCP Host Server
A multi-tenant remote MCP (Model Context Protocol) server hosting platform built for DigitalOcean deployment. This server allows multiple applications to connect to and use various MCP servers remotely over WebSocket connections.
Features
- 🏗️ Multi-tenant Architecture: Isolated MCP servers for different tenants
- 🔒 Secure Authentication: JWT-based authentication with role-based access
- 🌐 Remote Access: WebSocket-based connections for remote MCP usage
- 📊 Management API: RESTful API for server management
- 🛡️ Security: Rate limiting, input validation, CORS protection
- 📈 Scalable: Designed for cloud deployment on DigitalOcean
Architecture
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ MCP Client │ │ MCP Client │ │ MCP Client │
│ (Claude, etc.) │ │ (API App) │ │ (Custom) │
└─────────┬───────┘ └─────────┬───────┘ └─────────┬───────┘
│ │ │
└──────────────────────┼──────────────────────┘
│ WebSocket Connections
┌────────────▼────────────┐
│ MCP Host Server │
│ ┌─────────────────┐ │
│ │ Tenant A │ │
│ │ ├ File Server │ │
│ │ ├ DB Server │ │
│ │ └ API Server │ │
│ ├─────────────────┤ │
│ │ Tenant B │ │
│ │ ├ Custom Server │ │
│ │ └ File Server │ │
│ └─────────────────┘ │
└─────────────────────────┘
Quick Start
1. Installation
# Clone the repository
git clone https://github.com/motionmavericks/mcp-server.git
cd mcp-server
# Install dependencies
npm install
# Copy environment file
cp .env.example .env
2. Configuration
Edit .env
file with your settings:
PORT=3000
JWT_SECRET=your_secret_here
ALLOWED_ORIGINS=https://yourdomain.com
ADMIN_PASSWORD=YourSecurePassword123!
3. Development
# Start in development mode
npm run dev
# Or start normally
npm start
4. Production Deployment
# Set environment to production
export NODE_ENV=production
# Start the server
npm start
API Usage
Authentication
# Login
curl -X POST http://localhost:3000/api/auth/login \
-H "Content-Type: application/json" \
-d '{"email": "admin@mcp.mvrx.com.au", "password": "SecureMCP2024!"}'
Create MCP Server
# Create a file manager server
curl -X POST http://localhost:3000/api/servers \
-H "Authorization: Bearer YOUR_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "My File Server",
"type": "file-manager",
"description": "File management MCP server"
}'
Get Connection URL
# Get WebSocket connection URL for MCP client
curl -X POST http://localhost:3000/api/servers/SERVER_ID/connect \
-H "Authorization: Bearer YOUR_TOKEN"
MCP Client Configuration
Claude Desktop
Add to your claude_desktop_config.json
:
{
"mcpServers": {
"remote-file-server": {
"command": "node",
"args": ["path/to/mcp-websocket-client.js"],
"env": {
"MCP_SERVER_URL": "ws://mcp.mvrx.com.au:3001/mcp-ws?tenant=TENANT_ID&server=SERVER_ID&token=CONNECTION_TOKEN"
}
}
}
}
Custom MCP Client
import { WebSocket } from 'ws';
import { Client } from '@modelcontextprotocol/sdk/client/index.js';
const ws = new WebSocket('ws://mcp.mvrx.com.au:3001/mcp-ws?tenant=TENANT_ID&server=SERVER_ID&token=CONNECTION_TOKEN');
const client = new Client({ name: 'my-app', version: '1.0.0' }, { capabilities: {} });
await client.connect(new WebSocketClientTransport(ws));
// Use MCP tools
const tools = await client.listTools();
console.log('Available tools:', tools);
Available MCP Server Types
File Manager
read_file
: Read file contentswrite_file
: Write file contentslist_files
: List directory contents
Database
query_database
: Execute SQL queries
API Client
api_request
: Make HTTP API requests
Custom
- Define your own tools and resources
Security Features
- JWT Authentication: Secure token-based authentication
- Rate Limiting: Prevents abuse and DoS attacks
- Input Validation: Sanitizes all inputs
- CORS Protection: Configurable cross-origin policies
- Security Headers: Helmet.js security headers
- Request Logging: Comprehensive audit logging
Deployment on DigitalOcean
App Platform
- Create new app from GitHub repository
- Set environment variables in App Platform dashboard
- Configure custom domain (optional)
- Deploy!
Example App Spec
name: mcp-host-server
services:
- name: api
source_dir: /
github:
repo: motionmavericks/mcp-server
branch: main
run_command: npm start
environment_slug: node-js
instance_count: 2
instance_size_slug: apps-s-1vcpu-1gb
http_port: 3000
envs:
- key: NODE_ENV
value: production
- key: JWT_SECRET
value: YOUR_SECRET
- key: ALLOWED_ORIGINS
value: https://mcp.yourdomain.com
Monitoring
Health Check
curl http://localhost:3000/health
System Status
curl -H "Authorization: Bearer TOKEN" \
http://localhost:3000/api/status
Development
Project Structure
src/
├── config/ # Configuration management
├── middleware/ # Express middleware
├── routes/ # API routes
├── services/ # Business logic
└── utils/ # Utilities
Adding New MCP Server Types
- Extend
MCPServerManager.registerServerTools()
- Add new tool handlers
- Update validation in API routes
- Add tests
Testing
npm test
Contributing
- Fork the repository
- Create feature branch (
git checkout -b feature/new-feature
) - Commit changes (
git commit -am 'Add new feature'
) - Push to branch (
git push origin feature/new-feature
) - Create Pull Request
License
MIT License - see LICENSE file for details.
Support
- GitHub Issues: https://github.com/motionmavericks/mcp-server/issues
- Documentation: https://github.com/motionmavericks/mcp-server/wiki
- Email: support@motionmavericks.com.au
推荐服务器

Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。