CoplayDev/unity-mcp

Unity MCP acts as a bridge between AI assistants and your Unity Editor. Give your LLM tools to manage assets, control scenes, edit scripts, and automate tasks within Unity

Summary

MCP for Unity is an MIT-licensed Unity package that exposes the Unity Editor to AI assistants through the Model Context Protocol, letting an LLM create scenes, edit C# scripts, manage assets, run tests, and automate game-dev tasks. It runs as a Unity Editor plugin plus a small Python bridge, and works with any MCP-capable client such as Claude Desktop, Cursor, VS Code, Windsurf, Cline, or Gemini CLI. The project is community-maintained and is not affiliated with Unity Technologies.

What it is useful for

The project is useful when you want an AI assistant to drive the Unity Editor directly rather than just suggesting code snippets. The README describes around 50 MCP tool entrypoints covering scene and GameObject creation, C# script editing, asset management, test execution, profiling, and building. Practical scenarios mentioned in the docs include:

• Scaffolding scenes from a natural language prompt (the README shows: "Create a cube at the origin and add a Rigidbody"), useful for rapid prototyping.
• Editing scripts and validating them with Roslyn before they land in Unity, reducing the back-and-forth of compile errors.
• Running tests and inspecting the editor from a chat interface, which fits a TDD-style workflow inside a single chat window.
• Routing commands across multiple Unity instances for users running parallel projects or build slaves (documented as Multi-Instance Routing).
• Applying tool groups for VFX, animation, UI, and testing areas so the LLM only sees relevant tools.

It also supports a remote-hosted server with auth for running the editor bridge off a workstation.

How engineers can use it

Setup is documented in three steps:

1. Install the Unity package via Unity → Package Manager → "Add package from git URL" using https://github.com/CoplayDev/unity-mcp.git?path=/MCPForUnity#main (the README notes pinning to #v10.0.0 for that release line, or installing via OpenUPM with openupm add com.coplaydev.unity-mcp).
2. Configure the clients by opening Unity and selecting Window → MCP for Unity → Configure All Detected Clients, which registers the editor with any MCP clients it finds on the machine.
3. Prompt from a connected MCP client and verify the editor responds.

Documented requirements are Unity 2021.3 LTS through 6.x and Python 3.10+ managed through uv. Projects on older Unity versions are not supported. Contributors are asked to branch off beta rather than main, per CONTRIBUTING.md. For full tool catalogs, migration notes, and multi-instance setup that go beyond the README quickstart, the project points to its wiki at coplaydev.github.io/unity-mcp.