Model Context Protocol for Unreal Engine logo

Model Context Protocol for Unreal Engine

by chongdashu

This project enables AI assistant clients like Cursor, Windsurf and Claude Desktop to control Unreal Engine through natural language using the Model Context Protocol (MCP). It provides comprehensive tools for actor management, blueprint development, editor control, and more.

View on GitHub

Last updated: N/A

What is Model Context Protocol for Unreal Engine?

The Unreal MCP integration is a set of tools and a plugin that allows AI assistants to control Unreal Engine through natural language commands. It includes a native TCP server, integrates with Unreal Editor subsystems, and implements actor manipulation tools.

How to use Model Context Protocol for Unreal Engine?

To use Unreal MCP, you need to install the plugin in your Unreal Engine project, set up the Python server, and configure your MCP client (e.g., Cursor, Windsurf, Claude Desktop) with the correct server settings. The README provides a quick start guide with detailed instructions for each step.

Key features of Model Context Protocol for Unreal Engine

  • Actor Management (creation, deletion, transforms, querying)

  • Blueprint Development (creation, component configuration, compilation)

  • Blueprint Node Graph manipulation

  • Editor Control (viewport focus, camera control)

  • Natural Language Command Interface

Use cases of Model Context Protocol for Unreal Engine

  • Automating repetitive tasks in Unreal Engine

  • Controlling Unreal Engine with voice commands

  • Rapid prototyping of game mechanics

  • Creating AI-driven tools for level design

FAQ from Model Context Protocol for Unreal Engine

What Unreal Engine versions are supported?

Unreal Engine 5.5 and later are supported.

What Python version is required?

Python 3.12 or later is required.

Which MCP clients are compatible?

Claude Desktop, Cursor, and Windsurf are known to be compatible.

Where can I find the MCP configuration file?

The location depends on the MCP client you are using. See the README for specific locations for each client.

Is this project production-ready?

No, this project is currently in an EXPERIMENTAL state and is not recommended for production use.