Repolode
CoplayDev

CoplayDev/unity-mcp

A bridge that lets AI assistants like Claude, Cursor or VS Code control the Unity Editor through the Model Context Protocol.

Stars
15,189
Forks
1,704
Open issues
106
Last commit
12 hours ago

What is it?

MCP for Unity connects AI assistants to the Unity Editor through the Model Context Protocol. Once installed as a Unity package and configured in an MCP client, the assistant can create scenes and GameObjects, edit C# scripts, manage assets, run tests, profile, and build projects using natural language prompts.

The project exposes a set of MCP tool entrypoints covering scene, asset, script and editor operations. It works with any MCP-compatible client, including Claude Desktop and Claude Code, Cursor, VS Code, Windsurf, Cline and Gemini CLI. It targets Unity 2021.3 LTS through 6.x and requires Python 3.10 or newer, which is run through uv. The package can be installed from a git URL or OpenUPM, and a configuration window can set up detected clients automatically.

The project is MIT licensed and maintained by CoplayDev, with multi-instance routing, tool groups and optional Roslyn script validation for more advanced setups.

Key features

  • 01Create and modify scenes and GameObjects from natural language prompts.
  • 02Read, write and refactor C# scripts inside the Unity Editor.
  • 03Manage project assets and run Unity tests from the assistant.
  • 04Works with any MCP client, including Claude, Cursor and VS Code.
  • 05Supports Unity 2021.3 LTS through 6.x.
  • 06Supports routing across multiple running Unity instances.

Use cases

  • Prototyping a scene quickly by describing objects and components in plain language.
  • Automating repetitive asset or scene edits across a large project.
  • Letting an AI assistant write and check gameplay scripts while you test in the editor.
  • Running Unity tests and profiling from an AI-driven workflow.

Quick start

  1. 1

    Install the package via Unity Package Manager (Add from git URL)

    https://github.com/CoplayDev/unity-mcp.git?path=/MCPForUnity#main
  2. 2

    Configure MCP clients from the Unity editor menu

    Window → MCP for Unity → Configure All Detected Clients
  3. 3

    Prompt your AI assistant to create something in the scene

Things to know

  • Requires Unity 2021.3 LTS or later and Python 3.10+ with uv.
  • Changes in the Unity Editor are driven by AI assistants, so review generated edits carefully.

Topics

FAQ

What is unity-mcp?
A bridge that lets AI assistants like Claude, Cursor or VS Code control the Unity Editor through the Model Context Protocol. MCP for Unity connects AI assistants to the Unity Editor through the Model Context Protocol. Once installed as a Unity package and configured in an MCP client, the assistant can create scenes and GameObjects, edit C# scripts, manage assets, run tests, profile, and build projects using natural language prompts.
Is unity-mcp free to use? What license does it have?
unity-mcp is open source under the MIT license. Check the license terms before using it in commercial products.
Is unity-mcp still maintained?
Actively maintained. The last commit to unity-mcp was on Oct 19, 2026.
What can I use unity-mcp for?
Prototyping a scene quickly by describing objects and components in plain language. Automating repetitive asset or scene edits across a large project. Letting an AI assistant write and check gameplay scripts while you test in the editor. Running Unity tests and profiling from an AI-driven workflow.
How do I install unity-mcp?
Install the package via Unity Package Manager (Add from git URL): https://github.com/CoplayDev/unity-mcp.git?path=/MCPForUnity#main Configure MCP clients from the Unity editor menu: Window → MCP for Unity → Configure All Detected Clients Prompt your AI assistant to create something in the scene
Where can I get unity-mcp?
The source code and installation instructions are on GitHub: https://github.com/CoplayDev/unity-mcp

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects