home / mcp / unity mcp with ollama integration mcp server

Unity MCP with Ollama Integration MCP Server

Locally run MCP server that connects Unity to Ollama-based LLMs for asset and editor automation.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "zundamonnovrchatkaisetu-unity-mcp-ollama": {
      "command": "python",
      "args": [
        "server.py"
      ]
    }
  }
}

You set up a local MCP server that lets Unity talk to your own local LLMs through Ollama, enabling asset management, scene control, material editing, and editor automation without internet access or API keys.

How to use

Launch the Unity MCP bridge from your Unity project to connect the editor with a Python MCP server and Ollama-hosted models. Start the Python server to handle requests from Unity, and run Ollama to serve local LLMs. Use the integrated chat interface to issue practical tasks such as creating or modifying assets, manipulating scenes, applying materials, generating or updating scripts, and triggering editor actions.

Typical workflows you can perform include creating a red cube at a specific position, adding a sphere with a blue material, listing objects in the current scene, or generating and attaching a simple movement script to an object. You operate everything via the Unity MCP window, typing instructions into the chat and sending them for processing.

How to install

Prerequisites
- Unity 2020.3 LTS or newer
- Python 3.10 or newer
- Ollama installed on your system
- Ollama models pulled: ollama pull deepseek-r1:14b, ollama pull gemma3:12b
- Ollama server running: ollama serve

Install and configure the Unity MCP Ollama integration
1) Prepare Unity editor scripts
2) Set up a Python MCP server
3) Configure Ollama settings in Unity MCP window
4) Use the chat interface to issue commands

The Unity side requires you to import Editor scripts, start the Unity bridge, then run a local Python server that communicates with Ollama-hosted models. You interact with the system through the Unity MCP chat interface once all pieces are running.

Additional content

Prerequisites and setup steps are described in the following flow. You will configure the Ollama host and port, select a model (deepseek-r1:14b or gemma3:12b), and ensure the Python server is available for Unity to query. The system also provides status indicators for Python Server, Unity Bridge, and Ollama to help you diagnose connectivity.

Available tools

Asset Management

Create, import, and manipulate Unity assets programmatically through conversational prompts.

Scene Control

Manage scenes, objects, and their properties via natural language instructions.

Material Editing

Modify materials and their properties from prompts.

Script Integration

View, create, and update Unity scripts using generated code.

Editor Automation

Trigger Unity Editor actions such as undo/redo, play, and build from prompts.