Trellis MCP

🚀 Trellis MCP Server

Trellis MCP provides an interface between the AI assistant and Trellis, enabling integration through the Model Context Protocol (MCP).

Disclaimer

This project demonstrates a very basic integration of MCP with Trellis, a lightweight and open - source 3DAIGC (Three - Dimensional Artificial Intelligence Generated Content) model for text - to - 3D image/image - to - 3D model conversion. Compared with other existing integrations such as Rodin in Blender - MCP and TriPo integration, it has the following advantages:

• Faster and memory - efficient: You can use locally deployed Trellis (only 8GPU+ VRAM required), and generate textured meshes from text in only ~15s (can be reduced to 10s with more VRAM).
• Free: You don't need to pay for expensive APIs like Rodin/Meshy/TriPo.

However, it has the following limitations:

• Trellis is open - source, and there is no ready - made API provider available. You need to deploy it yourself (refer to README).
• The API/Prompt has not been fully tested/tuned, and there may be stability issues.

Therefore, please use it at your own risk.

🎥 Demo

A minimal demonstration of generating a single object. The integration of more complex prompts with Blender - MCP is under optimization.

✨ Features

• [x] Use the Trellis API to generate 3D assets from natural language (text) and import them into Blender.
• [ ] Use the Trellis API to generate textures/materials for a given 3D mesh from natural language (text) and import them into Blender.

🚧 Development Plan

Prerequisites

• Python 3.10+
• Blender
• Trellis Blender Plugin
• Trellis API Backend
• Claude / Cursor (recommended) / Windsurf

📦 Installation

1. Trellis Blender Plugin

1. Download the Trellis Blender plugin from here.
2. Open Blender -> Edit -> Preferences -> Add - ons -> Install from File -> Select the downloaded plugin -> Install.
3. In the 3D view -> 3D View > Sidebar > TRELLIS -> Start the MCP server.

2. Configure the API Backend

Since Trellis is an open - source text - to - 3D model, you need to deploy the Trellis API backend yourself according to README.

# Clone a trellis branch for the API
git clone https://github.com/FishWoWater/TRELLIS
# After cd to the TRELLIS directory, execute the following command:
python manage.py runserver

3. Install Python Packages

Run the following command in the terminal to install the required Python packages:

pip install trellis-mcp

Configuration Example

Start the MCP Server

Open Blender and start the MCP server in the Trellis plugin.

💻 Usage Examples

Basic Usage

1. Enable the Trellis plugin in Blender.
2. Open the MCP server.
3. Use your AI assistant to interact with Trellis through the MCP protocol.

Advanced Usage

• Support for customizing mesh generation parameters.
• An extensible plugin architecture to support more features.

Code Examples

Create a Basic Grid

from trellis.mcp import GridGenerator

# Create a simple cube grid
grid = GridGenerator.create_cube(length=1.0, width=1.0, height=1.0)

Apply Texture

from trellis.mcp import TextureApplicator

# Apply texture to the grid
TextureApplicator.apply_texture(grid, texture_path="path/to/texture.png")

❓ Issues and Support

If you need help or want to report an issue, please visit the GitHub repository to submit an ISSUE.

🤝 Contribution

Contributions are welcome! Please refer to the CONTRIBUTING.md file to learn how to participate in this project.