Unreal_mcp
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to control Unreal Engine through the native C++ Automation Bridge plugin. Built with TypeScript, C++, and Rust (WebAssembly) for ultra-high-performance game development automation.
claude mcp add --transport stdio chir24-unreal_mcp npx -y unreal-engine-mcp-server \ --env UE_PROJECT_PATH="C:/Path/To/YourProject" \ --env MCP_AUTOMATION_PORT="8091"
How to use
This MCP server enables AI assistants to control Unreal Engine through the native C++ Automation Bridge plugin. Built with TypeScript and C++, it exposes a broad set of capabilities for asset management, actor control, editor operations, level handling, animation, physics, visual effects, sequencer, graph editing, audio, and system commands. You can orchestrate Unreal workflows by issuing MCP tool actions to remotely spawn actors, manage assets, control the editor, manipulate levels, run PIE sessions, capture screenshots, adjust cameras, and query/modify numerous Unreal systems. The server relies on the MCP Automation Bridge inside Unreal to route commands, and it supports runtime discovery of objects, safe command execution, and per-IP rate limiting for the metrics endpoint. To connect your client, configure an MCP server entry pointing to the installed server (for example via NPX usage or a built binary) and provide the Unreal project path and automation port as environment variables.
How to install
Prerequisites:
- Node.js 18+ installed on your machine
- Unreal Engine 5.0–5.7 installed and a project ready for automation
Installation steps:
- Install the MCP server locally (NPX option is recommended):
# Option A: NPX (Recommended)
npx unreal-engine-mcp-server
- Alternatively, clone and build from source:
git clone https://github.com/ChiR24/Unreal_mcp.git
cd Unreal_mcp
npm install
npm run build
node dist/cli.js
- Prepare the Unreal Engine plugin (MCP Automation Bridge) inside your Unreal project as described in the repository:
- Copy or install the plugin from Unreal_mcp/plugins/McpAutomationBridge into YourUnrealProject/Plugins/McpAutomationBridge
- Enable required plugins in Unreal Editor (MCP Automation Bridge, Editor Scripting Utilities, Niagara)
- Configure the MCP client by adding an MCP server entry, either using NPX or a built path (see the example configurations in the README):
{
"mcpServers": {
"unreal-engine": {
"command": "node",
"args": ["path/to/Unreal_mcp/dist/cli.js"],
"env": {
"UE_PROJECT_PATH": "C:/Path/To/YourProject",
"MCP_AUTOMATION_PORT": "8091"
}
}
}
}
- Start the MCP server and connect your MCP client to the configured server.
Additional notes
Environment variables:
- UE_PROJECT_PATH is required for the server/client to locate your Unreal project.
- MCP_AUTOMATION_HOST and MCP_AUTOMATION_PORT control where the Automation Bridge listens. By default, the bridge binds to 127.0.0.1:8091 for security.
LAN Access: If you enable non-loopback access, ensure you understand the security implications and restrict access with firewalls. See the configuration guidance in the README for enabling non-loopback and setting Listen Host to 0.0.0.0.
Common issues:
- If the Unreal plugin fails to load after the first project open, simply reopen the project or rebuild the plugin from source.
- Ensure required plugins are enabled in Unreal Editor before starting automation.
- Make sure your Node/NPX version matches the documented prerequisites (Node.js 18+).
Monitoring:
- Use the MCP metrics endpoint to monitor request rates and health, and adjust LOG_LEVEL as needed for debugging.
Related MCP Servers
unreal
Enable AI assistant clients like Cursor, Windsurf and Claude Desktop to control Unreal Engine through natural language using the Model Context Protocol (MCP).
UE5
MCP for Unreal Engine 5
openfoam
LLM-powered OpenFOAM MCP server for intelligent CFD education with Socratic questioning and expert error resolution
flapi
API Framework heavily relying on the power of DuckDB and DuckDB extensions. Ready to build performant and cost-efficient APIs on top of BigQuery or Snowflake for AI Agents and Data Apps
processhacker
your ai debugger, vibe hacking tool
machinepal
The Cloud-Native MCP and X402 Gateway to Run and Monetize your AI Agents and Services, as well as optimize your AI costs