MCP 服务器

mcp_3d_relief

mcp\_3d\_relief

README

MCP STL 3D Relief Generator

</div>

This project provides a MCP server that converts 2D images into 3D relief models in STL format, suitable for 3D printing or rendering.

Features

Convert any image to a 3D relief model
Control model dimensions (width, thickness)
Add optional base to the 3D model
Invert depth for different relief effects
Fast processing with immediate download links

Installation

Prerequisites

Python 3.10+
uv

Option 1: Local Installation

Clone the repository:

git clone https://github.com/bigchx/mcp_3d_relief.git
cd mcp_3d_relief

Install dependencies:

uv pip sync requirements.txt

Run/Inspect the server:

mcp run server.py
mcp dev server.py

Usage

JSON Configuration

{
  "mcpServers": {
    "mcp_3d_relief": {
      "command": "uv",
      "args": ["--directory", "{fill_in_your_path_here}", "run", "server.py"]
    }
  }
}

MCP Tool Parameters

image_path: Local path or web URL to the input image file
model_width: Width of the 3D model in mm (default: 50.0)
model_thickness: Maximum thickness/height of the 3D model in mm (default: 5.0)
base_thickness: Thickness of the base in mm (default: 2.0)
skip_depth: Whether to use the image directly or generate a depth map (default: true)
invert_depth: Invert the relief (bright areas become low instead of high) (default: false)
detail_level: Controls the resolution of the processed image (default: 1.0). At detail_level = 1.0, the image is processed at 320px resolution, producing an STL file typically under 100MB. Higher values improve detail quality but significantly increase both processing time and STL file size. For example, doubling the detail_level can increase file size by 4x or more. Use with caution.

Response

The MCP Tool returns a JSON response with:

{
  "status": "success",
  "depth_map_path": "path/to/yourimage_depth_map.png",
  "stl_path": "path/to/yourimage.stl"
}

Where LLMs can access the generated files from this MCP server, using the provided URLs.

Command Line

You can also use the script directly from the command line to generate a relief model from an image:

python3 relief.py path/to/your/image.jpg

External Depth Map Generation

For higher quality depth maps, you can use external depth map generation services like Depth-Anything-V2. This service can generate more accurate depth maps that you can then use with this project:

Visit https://huggingface.co/spaces/depth-anything/Depth-Anything-V2
Upload your image to generate a depth map
Download the generated depth map
Use this depth map with our converter by setting skip_depth=false

This approach can provide better 3D relief models, especially for complex images.

How It Works

The image is processed to create a depth map (darker pixels = lower, brighter pixels = higher)
The depth map is converted to a 3D mesh with triangular facets
A base is added to the bottom of the model
The model is saved as an STL file