Jhacksman OpenSCAD MCP Server

🚀 OpenSCAD 3D Model Generation Server

This project is an OpenSCAD 3D model generation server, which supports image generation, multi - view processing, 3D model reconstruction, and exports in multiple 3D file formats, providing a convenient and efficient 3D model generation solution.

🚀 Quick Start

Prerequisites

• Operating System: Linux or Windows are supported (Ubuntu 20.04+ is recommended).
• Python Version: Python 3.8+.
• Dependencies: OpenCV, Flask, requests, jsonschema, numpy, open3d.
• CUDA Support: CUDA 11+ for accelerating 3D model reconstruction.

Installation Steps

git clone https://github.com/your-repository.git
cd openscad-server
pip install -r requirements.txt

✨ Features

Core Functions

• Image Generation: Supports Gemini and Venice.ai APIs, allowing customization of the number of views and styles.
• Multi - view Processing: Approve and manage multi - view images through the web interface.
• 3D Model Reconstruction: Utilize CUDA MVS technology for model reconstruction, supporting both local and remote server computing.
• Format Export: Supports multiple 3D file formats such as OBJ, STL, PLY, SCAD, etc.

Example Requests

Image Generation

{
    "prompt": "A cute rabbit",
    "num_views": 4,
    "model": "gemini-2.0-flash-exp-image-generation"
}

Multi - view Reconstruction

{
    "image_ids": ["view_1.jpg", "view_2.jpg", "view_3.jpg", "view_4.jpg"],
    "output_name": "Rabbit model"
}

📚 Documentation

Image Generation Options

1. Google Gemini API (Default): Uses the Gemini 2.0 Flash Experimental model to provide high - quality images.
   • Supports consistent multi - view style generation. Google API Key needs to be configured.
2. Venice.ai API (Optional): Supports multiple models such as flux - dev and fluently - xl.
   • Venice.ai API Key needs to be configured.
3. Custom Images: Upload local pictures, suitable for existing photos or renderings.

Multi - view Workflow

1. Generate multi - view images.
2. Approve the images.
3. Convert to a 3D model (supports local and remote CUDA MVS).
4. Optionally use OpenSCAD for model optimization.

Remote Processing Flow

1. Discover CUDA MVS servers on the network.
2. Upload approved multi - view images to the remote server.
3. Submit the job and track the progress.
4. Download the completed 3D model.

Supported Export Formats

• OBJ: Wavefront Object File (standard 3D model).
• STL: Standard Triangle Language (suitable for 3D printing).
• PLY: Polygon File Format.
• SCAD: OpenSCAD Source Code.
• CSG: OpenSCAD CSG Format (retains all parameter attributes).
• AMF: Additive Manufacturing File Format.
• 3MF: 3D Manufacturing Format (a modern alternative to STL).

Web Interface

Access URL: http://localhost:8000/ui/

Functions include:

• Generate and approve multi - view images.
• Preview 3D models (supports rotation and zoom).
• Manage API keys and server configurations.
• View historical job records.

📄 License

This project is licensed under the MIT License. The code repository is available at OpenSCAD Server.

🔧 Contribution Guidelines

We welcome PRs and issues:

• When submitting an issue, please include the log file and a detailed description.
• Pass unit tests before submitting a PR.
• Fork the repository and develop based on the main branch.