This project connects LLMs like Claude to Blender through the Model Context Protocol, enabling prompt-assisted 3D modeling and scene manipulation. It relays commands from an AI client over TCP sockets to a Blender addon, which executes the operations and returns results or viewport screenshots.
Project overview
It bridges text prompts from MCP-compatible AI clients directly to Blender scene operations, supporting object creation, material application, and Python code execution within the 3D environment.
Project type
MCP
Deployment
Refer to project documentation
License
MIT
Best for
Creators and developers seeking to connect MCP-compatible AI clients to Blender for prompt-assisted 3D modeling and scene manipulation.
Key capabilities
Connects AI clients to Blender through a socket-based server using a simple JSON-based protocol.
Allows the AI client to create, modify, and delete 3D objects in the connected Blender scene.
Provides the ability to apply and modify materials and colors on objects within the scene.
Retrieves detailed information about the current Blender scene for the AI client to process.
Runs arbitrary Python code in Blender from an AI client, which requires caution in production environments.
Captures and provides screenshots of the Blender viewport so the AI client can better understand the scene.
Supports running the Blender MCP setup on a remote host.
Includes added support for integration with Hunyuan3D.
Limitations and risks
Complex operations might need to be broken down into smaller steps for the AI client to process and execute correctly.
The execute_blender_code tool allows running arbitrary Python code in Blender, which is potentially dangerous. Users should always save their work before using it and exercise caution in production environments.
Getting started
You must install the uv package manager and the Blender addon, then edit the JSON configuration for your MCP client. Begin by installing uv, then download and install addon.py in Blender. Next, edit your client config to add the uvx blender-mcp command, start the MCP server in a terminal, and click Connect to Claude in Blender.
Alternatives and comparisons
A fair-code platform to build and deploy AI agents and workflows with a visual canvas, custom code, and 1500+ integrations.
A lightweight, extensible AI assistant and agent harness that runs 24/7, supports multiple LLMs and messaging channels, manages long-term memory, and is easily deployable via one-line installer.
Create, modify, and enhance draw.io diagrams through natural language commands using AI.