db-lyon/ue-mcp

Complete Unreal Engine development toolkit exposed as MCP tools.

What it solves

UE-MCP allows AI assistants to directly interact with the Unreal Engine editor. It removes the manual effort of translating natural language requests into complex editor actions by providing a programmable interface that AI agents can use to build levels, author materials, create Blueprints, and manage project assets.

How it works

The project implements the Model Context Protocol (MCP) to create a bridge between an AI client (like Claude or Cursor) and the Unreal Editor. It consists of a a TypeScript/Node.js server that communicates via WebSocket/JSON-RPC with a C++ bridge plugin installed inside the Unreal Editor. This setup allows the agent to perform over 1,000 actions across 24 categories, including direct filesystem access to C++ headers and INI configs for project exploration without the editor even being open.

Who it’s for

Game developers and technical artists using Unreal Engine 5.4-5.8 who want to automate editor tasks, procedurally generate content via AI, or use AI agents to handle repetitive technical setup in their projects.

Highlights

  • Deep Editor Access: Controls everything from Niagara VFX and PCG graphs to Gameplay Ability System (GAS) and StateTree.
  • YAML Flows: Supports multi-step workflows defined in YAML that can be hot-reloaded and include rollback capabilities.
  • Native Toolset Integration: Wraps Epic's native AI Toolset Registry (UE 5.8+) to surface hundreds of official tools as agent actions.
  • Extensible Architecture: New actions can be added via an npm plugin system.
  • Safety Guards: Includes write guards and source control integration to prevent AI agents from accidentally overwriting critical project data.

Related

  • Project
    cardenluo/ComfyUI-Apt_Preset

    A ComfyUI plugin that simplifies complex workflows by organizing them into a streamlined Loader-Controller-Sampler framework with preset support.

  • Project
    ModelEngine-Group/unified-cache-management

    A cache management framework for LLMs that reduces inference latency and GPU memory usage by persisting KV caches and implementing sparse attention retrieval.

  • Project
    janosh/pymatviz

    A visualization toolkit for materials informatics that provides specialized plots and interactive widgets for chemical compositions, crystal structures, and vibrational properties.

  • Project
    microsoft/agent-host-protocol

    A synchronized, multi-client state protocol that allows multiple clients to maintain a consistent view of AI agent sessions via a standalone sessions server.