Clean Architecture MCP Server
Enables scaffolding and managing Clean Architecture in Next.js projects, providing tools to initialize structure, create features with all layers, and validate architecture compliance.
README
Clean Architecture MCP Server
Model Context Protocol Server for scaffolding and managing Clean Architecture in Next.js projects.
🎯 Features
This MCP Server provides 3 powerful tools:
initialize_clean_architecture- Set up complete Clean Architecture structurecreate_feature- Generate new features with all layersvalidate_architecture- Check for architecture violations
🚀 Installation
Option 1: Global Installation (Recommended)
# From the mcp-server directory
cd mcp-server
npm install
npm run build
npm link
# Now available globally as 'clean-architecture-mcp'
Option 2: Local Installation
cd mcp-server
npm install
npm run build
⚙️ Configuration
For Cursor/Claude Desktop
Add to your MCP configuration file:
Mac/Linux: ~/.config/cursor/mcp.json or ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Cursor\mcp.json or %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"clean-architecture": {
"command": "node",
"args": ["/absolute/path/to/mcp-server/dist/index.js"]
}
}
}
Or if globally installed:
{
"mcpServers": {
"clean-architecture": {
"command": "clean-architecture-mcp"
}
}
}
Restart Cursor/Claude
After configuration, restart Cursor or Claude Desktop to load the MCP server.
📚 Usage
Tool 1: Initialize Architecture
Initialize Clean Architecture in your Next.js project:
In Cursor chat:
Use the clean-architecture MCP server to initialize the project structure
With parameters:
Use clean-architecture to initialize with features: products, orders, customers
What it does:
- ✅ Creates complete directory structure
- ✅ Adds .gitkeep files
- ✅ Creates README.md in each layer
- ✅ Updates tsconfig.json paths
- ✅ Sets up domain/infra/components/tests folders
Tool 2: Create Feature
Generate a complete feature:
In Cursor chat:
Use clean-architecture to create a "products" feature
What it creates:
src/domain/products/
├── models/product.ts # Zod schema
├── business-rules/ # Pure business logic
├── services/product-service.ts # Use cases
└── ports/product-repository.ts # Interface
src/infra/adapters/
└── product-repository.prisma.ts # Implementation
src/app/api/products/
├── route.ts # GET, POST
└── [id]/route.ts # GET, DELETE
Tool 3: Validate Architecture
Check for architecture violations:
In Cursor chat:
Use clean-architecture to validate the project
What it checks:
- ❌ Next.js imports in domain layer
- ❌ React imports in domain layer
- ❌ Direct Prisma imports in domain
- ⚠️ Missing ports/interfaces
- ⚠️ Missing tests
Output:
{
"success": true,
"score": 100,
"issues": [],
"warnings": [],
"message": "✅ Architecture validation passed!"
}
🎨 Examples
Example 1: New Project Setup
User: Initialize Clean Architecture in this Next.js project with features: users, products, orders
AI: [Uses initialize_clean_architecture tool]
Result:
✅ Created 45 directories
✅ Created 6 README files
✅ Updated tsconfig.json
✅ Ready to develop!
Example 2: Create Feature
User: Create a "products" feature with fields: name, price, stock
AI: [Uses create_feature tool with featureName: "products"]
Result:
✅ Created models/product.ts
✅ Created business rules
✅ Created service layer
✅ Created repository port
✅ Created Prisma adapter
✅ Created API routes
Example 3: Validate
User: Check if my architecture follows Clean Architecture principles
AI: [Uses validate_architecture tool]
Result:
❌ Found 2 issues:
- src/domain/users/services/create-user.ts: Contains Next.js import
- src/domain/products/models/product.ts: Direct Prisma import
Score: 60/100
🛠️ Development
# Install dependencies
npm install
# Build
npm run build
# Watch mode (during development)
npm run dev
# Test the server
node dist/index.js
📋 Tool Reference
initialize_clean_architecture
Parameters:
targetDir(optional): Target directory (default: current directory)features(optional): Array of feature names (default:['users', 'auth', 'payments'])
Returns:
{
success: boolean;
message: string;
created: number;
directories: string[];
features: string[];
}
create_feature
Parameters:
featureName(required): Feature name in kebab-casetargetDir(optional): Target directory (default: current directory)fields(optional): Custom model fields
Returns:
{
success: boolean;
message: string;
feature: string;
created: string[];
}
validate_architecture
Parameters:
targetDir(optional): Directory to validate (default: current directory)
Returns:
{
success: boolean;
score: number;
issues: string[];
warnings: string[];
message: string;
}
🔧 Troubleshooting
MCP Server not appearing in Cursor
- Check configuration file path
- Verify absolute path to
dist/index.js - Restart Cursor completely
- Check Cursor logs:
Help > Show Logs
"Command not found" error
If globally installed:
npm link
# Verify
which clean-architecture-mcp
TypeScript errors
npm run build
# Check for compilation errors
Tool execution fails
Check that you're in a Next.js project directory with package.json.
📦 Publishing
To publish to npm:
npm publish
Then users can install globally:
npm install -g clean-architecture-mcp-server
🤝 Contributing
- Add new tools in
src/index.ts - Update tool schemas
- Test with
npm run build && node dist/index.js - Update this README
📄 License
MIT
🔗 Resources
Made with ❤️ for Clean Architecture enthusiasts
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。