AI Game Developer (Unity MCP)
A Unity plugin with an MCP server, skills and CLI: the agent drives scenes, assets, scripts and tests in the Unity Editor
- 4,546 ★ on the repository
- Last commit
- Apache-2.0
Broad access
Why this level
- Executes arbitrary C# code in the editor
- Can delete project assets, scripts and packages
Install
- Works in
Manual install
npm install -g unity-mcp-cli
unity-mcp-cli install-plugin ./MyUnityProject
unity-mcp-cli login
unity-mcp-cli open ./MyUnityProjectReplace ./MyUnityProject with your project path.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Executes arbitrary C# code in the editor
- Can delete project assets, scripts and packages
This is third-party code. Review the repository files before installing.
What it does
The plugin installs into a Unity project and starts an MCP server the agent uses to work with the editor. More than seventy tools cover assets, prefabs and packages, scene objects and components, camera and window screenshots, console logs, play mode control, reading and editing scripts, running C# via Roslyn, running tests and the profiler. From the plugin window you can generate skills tailored to the Unity version, OS and installed packages. Custom tools take one line of code, and the plugin also runs inside built games.
Who it is for. For Unity game developers who want an agent to work directly in the editor.
Good fit when
- You need to build a scene or place objects from a description
- You need to write a game mechanic and run tests right away
- You need to find a bug from console logs and fix a script
Not a fit when
- The project is not Unity
- You cannot let the agent execute arbitrary C# in the editor
Example request
Create 3 cubes in a circle with radius 2 and make them rotateMCP
- Transport
- stdio, http
- Authentication
- not required
Limitations
Requires an open Unity Editor with the plugin installed. The unity-mcp-cli flow includes signing in to ai-game.dev in the browser. Asset deletion and code execution tools change the project, so work under version control.
How to disable. Remove the com.ivanmurzak.unity.mcp package via Package Manager and delete the ai-game-developer server from your MCP client config.
FAQ
How do skills differ from MCP setup?
The plugin window can generate skills (recommended) or configure an MCP client; both give the agent access to editor tools.
Can I skip the CLI?
Yes, download the .unitypackage installer or run openupm add com.ivanmurzak.unity.mcp.
How is Unity MCP different from other Unity agent integrations?
The plugin installs into a Unity project and starts an MCP server, plus skills and a CLI. Over seventy tools cover assets, prefabs, packages, scene objects and components, screenshots, scripts and tests in the editor.
How do I install AI Game Developer (Unity MCP)?
Run npm install -g unity-mcp-cli and unity-mcp-cli install-plugin ./your-project-path. The plugin installs into a Unity project. It fits Claude Code, Claude Desktop, Codex, Cursor, Copilot, Gemini CLI, VS Code and OpenCode.
Discussion
No questions or reviews yet.
Related
MCP serversControl the Unity Editor from an agent over MCP: scenes and objects, C# scripts, assets, tests, profiling and builds
An Unreal Engine plugin with MCP: the agent drives the editor via RQL, edits Blueprints and C++, captures viewport shots; cloud or own server
MCP server that connects an AI agent to the Unity Editor: scene, objects, packages and project resources
An MCP server for Unity: control the scene, objects, components, builds and profiling from an agent
A Unity package: the Editor serves MCP over HTTP itself, and agents and the command line drive scenes, assets and builds
An MCP server inside the Unity Editor: the agent builds scenes and scripts, enters play mode and checks the game with screenshots
README in short
The README covers installing the plugin via unitypackage, OpenUPM or unity-mcp-cli, choosing an agent and configuring it from the AI Game Developer window. For manual setup it gives server binary paths per platform and commands for Claude Code, Gemini CLI and Copilot CLI. It then lists built-in tools by section, workflow examples and how to create custom tools.