Blender Open MCP

🚀 Blender MCP Open Source Project

The blender-open-mcp is an open-source project that integrates Blender with a local AI model (based on Ollama) to enable natural language control of Blender. This project utilizes the Model Context Protocol (MCP), allowing users to complete complex 3D modeling tasks through natural language instructions.

🚀 Quick Start

Start the Services

• Ollama Server: Run ollama serve.
• Blender MCP Service: Start the backend service to provide API support.

Example Command

mcp prompt "Create a cube" --host http://localhost:8000

✨ Features

• Natural Language Control: Operate Blender through natural language instructions.
• MCP Integration: Use the Model Context Protocol for efficient interaction.
• Ollama Support: Support local AI models, providing powerful inference capabilities.
• Blender Plugin: Offer a dedicated plugin to enhance the user experience.
• PolyHaven Integration: Optionally integrate PolyHaven resources to enrich modeling materials.

📦 Installation

Prerequisites

• Ensure Python 3.8 or a higher version is installed.
• Install Git and the pip package management tool.

Installation Steps

1. Clone the Repository: Use the following command to obtain the project code:

   git clone https://github.com/dhakalnirajan/blender-open-mcp.git
   

2. Create a Virtual Environment (recommended):

   cd blender-open-mcp
   python -m venv env
   source env/bin/activate  # On Linux/Mac
   # Or on Windows: .\env\Scripts\activate
   

3. Install Dependencies:

   pip install -r requirements.txt
   

4. Install the Blender Plugin:
   • Copy the plugin files to Blender's plugin directory.
   • Enable the plugin in Blender.
5. Download the Ollama Model: Use the following command to obtain the required AI model:

   ollama pull <model_name>
   

📚 Documentation

Function List

💡 Usage Tip

• Ensure that both the Ollama and MCP servers are running properly.
• Check if the Blender plugin is correctly enabled.
• Double - check the command - line parameters to avoid formatting errors.

For more help, please visit GitHub Issues.