⚠️ 本项目目前维护精力有限,欢迎有兴趣的开发者参与协作!
⚠️ Currently this project is not actively maintained due to time constraints. I would be very happy to collaborate with anyone interested in co-maintaining or extending it.
基于 Model Context Protocol (MCP) 的 AutoCAD 集成服务器,允许通过 Claude 等大型语言模型 (LLM) 与 AutoCAD 进行自然语言交互。
An AutoCAD integration server based on Model Context Protocol (MCP), enabling natural language interaction with AutoCAD via large language models like Claude.
🔗 项目在 MseeP.ai 展示 / Referenced on MseeP.ai:
https://mseep.ai/app/zh19980811-easy-mcp-autocad
- 🗣️ 自然语言控制 AutoCAD 图纸 / Natural language control of AutoCAD drawings
- ✏️ 基础绘图功能(线条、圆)/ Basic drawing tools (line, circle)
- 📚 图层管理 / Layer management
- 🧠 自动生成 PMC 控制图 / Auto-generate PMC control diagrams
- 🔍 图纸元素分析 / Drawing element analysis
- 🔦 文本高亮匹配 / Highlight specific text patterns
- 💾 内嵌 SQLite 数据库存储 CAD 元素 / Integrated SQLite storage for CAD elements
- Python 3.10+
- AutoCAD 2018+ (需支持 COM 接口 / with COM interface support)
- Windows 操作系统 / Windows OS
git clone https://github.com/yourusername/autocad-mcp-server.git
cd autocad-mcp-server
| `export_to_database` | 将 CAD 元素信息存入 SQLite |
## 2. 创建虚拟环境 / Create virtual environment
**Windows:**
```bash
python -m venv .venv
.venv\Scripts\activate
macOS / Linux:
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
pyinstaller --onefile server.py
python server.py
编辑配置文件:
Windows 路径 / Config path on Windows:
%APPDATA%\Claude\claude_desktop_config.json
示例配置 / Example config:
{
"mcpServers": {
"autocad-mcp-server": {
"command": "path/to/autocad_mcp_server.exe",
"args": []
}
}
}
功能 / Function | 描述 / Description |
---|---|
create_new_drawing |
创建新的图纸 / Create a new drawing |
draw_line |
画直线 / Draw a line |
draw_circle |
画圆 / Draw a circle |
set_layer |
设置当前图层 / Set current drawing layer |
highlight_text |
高亮显示匹配文本 / Highlight matching text |
scan_elements |
扫描并解析图纸元素 / Scan and parse drawing elements |
export_to_database |
导出 CAD 元素到数据库 / Export CAD data to SQLite |
📬 联系我 / Contact: [email protected]
Made with ❤️ for open-source learning.