CADCAMfun MCP Server

A Model Context Protocol (MCP) server for CAD/CAM applications that provides AI-powered design assistance, toolpath optimization, and machining parameter recommendations.

Overview

This MCP server exposes resources, tools, and prompts for CAD/CAM operations, allowing LLM applications to:

• Generate CAD components from text descriptions
• Optimize G-code for CNC machines
• Analyze CAD designs for manufacturability, structural integrity, and cost
• Get optimal cutting parameters for specific materials and tools
• Save components and toolpaths to the database
• Access CAD/CAM libraries, materials, and tools

Features

• Resources: Access component data, toolpaths, projects, materials, and machine configurations
• Tools: Generate CAD components, optimize G-code, analyze designs, calculate cutting parameters
• Prompts: Generate component designs, get machining parameters, and suggest design improvements
• Chat: Interactive assistance for CAD/CAM questions

Requirements

• Node.js 18+
• PostgreSQL database
• TypeScript 5.0+
• Prisma ORM

Installation

1. Clone the repository:

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

2. Install dependencies:

   npm install

3. Configure environment variables:

   cp .env.example .env
   # Edit the .env file with your database credentials and settings

4. Apply database migrations:

   npx prisma migrate dev

5. Build the project:

   npm run build

Running the Server

HTTP Server (Default)

To run as an HTTP server (recommended for web applications):

npm run start:http

The server will start at http://localhost:3000/mcp by default.

Stdio Server

To run as a stdio server (useful for command-line tools):

npm run start:stdio

Docker Deployment

Prerequisites

• Docker and Docker Compose
• API key for Anthropic Claude

Steps

1. Set up the environment variables:

   Create a .env file in the root directory with your API key:

   ANTHROPIC_API_KEY=your-api-key-here

2. Start the containers:

   cd docker
   docker-compose up -d

3. Initialize the database:

   docker-compose exec mcp-server npm run setup-db

4. Access the server at http://localhost:3000/mcp

API Reference

Resources

| URI | Description | |-----|-------------| | version://app | Application version information | | component://{id} | Get component by ID | | toolpath://{id} | Get toolpath by ID | | project://{id} | Get project by ID | | cam://materials | List available materials | | cam://tools | List available cutting tools | | cam://machine-configs | List available machine configurations | | cad://library-items | List public library items |

Tools

| Name | Description | |------|-------------| | generate-cad-component | Generate CAD components from text description | | optimize-gcode | Optimize G-code for CNC machines | | analyze-design | Analyze a CAD design for various aspects | | get-cutting-parameters | Get optimal cutting parameters | | save-component | Save a component to the database | | save-toolpath | Save a toolpath to the database |

Prompts

| Name | Description | |------|-------------| | generate-cad-component | Generate CAD components from description | | suggest-cad-improvements | Get suggestions for improving a CAD model | | machining-parameters-advice | Get advice on machining parameters |

Client Usage Example

import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js";

// Create client
const client = new Client({
  name: "My CADCAMfun Client",
  version: "1.0.0"
});

// Connect to server
const serverUrl = new URL("http://localhost:3000/mcp");
const transport = new StreamableHTTPClientTransport(serverUrl);
await client.connect(transport);

// Generate a CAD component
const result = await client.callTool({
  name: "generate-cad-component",
  arguments: {
    description: "A simple 10mm x 10mm x 2mm base plate with 4 mounting holes",
    complexity: "simple",
    style: "precise"
  }
});

console.log("Generated Component:", result.content[0].text);

Using with Claude and Cursor

To use the CADCAMfun MCP server with Claude and Cursor, you'll need to:

1. Start the MCP server
2. Configure Cursor to use the MCP server as a language model provider
3. Use Claude via the MCP server to assist with CAD/CAM tasks

This setup allows you to leverage Claude's capabilities through the MCP protocol, enhancing your CAD/CAM workflow with AI assistance.

Project Structure

cadcamfun-mcp-server/
├── .env                         # Environment variables
├── .env.example                 # Example environment file
├── .gitignore                   # Git ignore file
├── package.json                 # Node.js package configuration
├── tsconfig.json                # TypeScript configuration
├── prisma/                      # Prisma ORM files
│   └── schema.prisma            # Database schema
├── src/                         # Source code
│   ├── index.ts                 # Main entry point
│   ├── mcp-server.ts            # MCP server implementation
│   ├── types/                   # TypeScript type definitions
│   │   └── index.ts             # Types used in the server
│   ├── services/                # Service implementations
│   │   ├── aiService.ts         # AI service implementation
│   │   ├── cadService.ts        # CAD service implementation
│   │   └── camService.ts        # CAM service implementation
├── dist/                        # Compiled JavaScript files
├── scripts/                     # Utility scripts
│   └── setup-db.ts              # Database setup script
└── docker/                      # Docker deployment files
    ├── Dockerfile               # Docker image definition
    └── docker-compose.yml       # Docker Compose configuration

Available Environment Variables

| Variable | Description | Default | |----------|-------------|---------| | DATABASE_URL | PostgreSQL connection string | postgresql://username:password@localhost:5432/cadcamfun | | DIRECT_DATABASE_URL | Direct PostgreSQL connection | Same as DATABASE_URL | | TRANSPORT | Transport type (http or stdio) | http | | PORT | HTTP server port | 3000 | | ANTHROPIC_API_KEY | API key for Anthropic Claude | | | AI_MODEL | Default AI model | claude-3-7-sonnet-20250219 | | AI_MAX_TOKENS | Maximum tokens for AI responses | 6000 | | AI_TEMPERATURE | Temperature for AI generation | 0.7 | | AI_CACHE_ENABLED | Enable AI response caching | true | | AI_ANALYTICS_ENABLED | Enable AI analytics | true | | MCP_ENABLED | Enable MCP protocol | true | | MCP_STRATEGY | MCP caching strategy | balanced | | MCP_CACHE_LIFETIME | Cache lifetime in milliseconds | 3600000 | | CORS_ORIGINS | Allowed CORS origins | http://localhost:3000,https://cadcam.example.com | | JWT_SECRET | Secret for JWT tokens | |

Contributing

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

License

MIT