Blender
MCP server addon for Blender - Control Blender via AI agents through 51 powerful tools. Made to be used with PolyMCP for intelligent tool orchestration. Features thread-safe execution, auto-dependency installation, and complete 3D workflow automation.
claude mcp add --transport stdio poly-mcp-blender-mcp-server python blender_mcp.py \ --env HOST="0.0.0.0" \ --env PORT="8000" \ --env ENABLE_CACHING="True" \ --env AUTO_INSTALL_PACKAGES="True" \ --env THREAD_SAFE_OPERATIONS="True"
How to use
This Blender MCP Server turns Blender into an MCP-enabled API playground. Once the addon script is running, the server exposes a FastAPI-based HTTP interface at port 8000 with more than 50 tools that let you create and manipulate objects, materials, animations, cameras, lights, physics simulations, geometry nodes, file I/O, and scene management. You can discover available tools, invoke specific tools, and inspect results via the provided endpoints. For integration with AI agents, you can connect PolyMCP and use the endpoint /mcp/invoke/{tool_name} to perform operations or chain multiple tools for multi-step tasks. The /docs endpoint provides interactive API documentation, while /mcp/list_tools enumerates all available tools, categorized by capability (Object Operations, Transformations, Materials & Shading, etc.).
How to install
Prerequisites:
- Blender 3.0+ installed on your machine
- Basic familiarity with Blender and Python
Installation steps:
- Download the Blender MCP server script: blender_mcp.py
- Open Blender and install the addon script via:
- Edit → Preferences → Add-ons
- Install from Disk… and select blender_mcp.py
- Enable the addon by checking 'MCP Complete Server for Blender'
- Place blender_mcp.py in a known location (the same directory where you will run it from is fine)
- Run the server:
- Open Blender, press N to open the sidebar, go to the MCP Server tab, and click Start Server
- By default the server will be available at http://localhost:8000
Alternatively you can run directly from a terminal (outside Blender) if the script is designed for standalone execution:
python blender_mcp.py
First-run setup will install required dependencies automatically (FastAPI, Uvicorn, Pydantic, and others) if you have network access.
Additional notes
Tips and troubleshooting:
- The server runs on port 8000 by default; if you need a different port, edit the Config in blender_mcp.py or set the PORT env var when launching.
- If the server cannot start, check Blender's System Console (Window → Toggle System Console) for errors and ensure port 8000 is not in use by another process.
- If dependencies fail to install automatically, install them manually inside Blender's Python environment: /path/to/blender/python -m pip install fastapi uvicorn pydantic docstring-parser numpy
- The API docs are available at http://localhost:8000/docs; use /mcp/list_tools to see all tools and /mcp/invoke/{tool_name} to run a tool.
- For production use, consider running Blender headlessly or in a dedicated environment; the addon is designed for thread-safe operation and queues multiple requests safely.
Related MCP Servers
bytebot
Bytebot is a self-hosted AI desktop agent that automates computer tasks through natural language commands, operating within a containerized Linux desktop environment.
cursor-talk-to-figma
TalkToFigma: MCP integration between AI Agent (Cursor, Claude Code) and Figma, allowing Agentic AI to communicate with Figma for reading designs and modifying them programmatically.
bytechef
Open-source, AI-native, low-code platform for API orchestration, workflow automation, and AI agent integration across internal systems and SaaS products.
cheatengine -bridge
Connect Cursor, Copilot & Claude directly to Cheat Engine via MCP. Automate reverse engineering, pointer scanning, and memory analysis using natural language.
robloxstudio
Create agentic AI workflows in ROBLOX Studio
automagik-genie
🧞 Automagik Genie – bootstrap, update, and roll back AI agent workspaces with a single CLI + MCP toolkit.