What Is Funplay MCP for Unity: 76 Tools for Claude Code and Cursor

Funplay MCP for Unity is an MIT-licensed Model Context Protocol server that runs inside the Unity Editor and exposes 79 tools to AI coding assistants such as Claude Code, Cursor, VS Code Copilot, Windsurf, and Codex. It gives those assistants direct, structured access to a live Unity project—handling scene creation, script generation, input simulation, runtime validation, performance profiling, and editor automation through a local HTTP endpoint.

What the Project Is and Why It Matters

Funplay MCP for Unity is a Unity Editor MCP server. After installation, it listens on http://127.0.0.1:8765/ and responds to standard MCP protocol requests from any connected AI client. The server ships with 79 built-in tools organized into two profiles: core, which contains 19 tools for everyday operations, and full, which exposes all 79 tools for complete editor control.

Three facts define the project's scope:

1. 79 built-in tools covering six categories: scene creation, script generation, runtime validation, input simulation, performance analysis, and editor automation. The core profile's 19 tools handle the operations developers reach for most often—querying game objects, reading component properties, creating basic scene elements, and running validation checks. The full profile adds advanced operations like virtual input simulation, profiling hooks, and batch automation.

2. Broad client compatibility. The server works with Claude Code, Cursor, VS Code Copilot, Windsurf, and Codex. Any MCP-compatible client can connect once the server is running—no custom adapters required.

3. MIT-licensed open source. The full source lives at github.com/FunplayAI/funplay-unity-mcp. You can inspect, fork, and modify it without licensing constraints.

For teams already using AI-assisted coding, the practical benefit is direct editor access. Instead of copying error logs or describing scene hierarchies in plain text, the model queries the editor state itself. It reads transform hierarchies, inspects component values, triggers builds, and validates runtime behavior—without you switching away from the terminal or IDE.

Setting Up and Connecting Your AI Client

Installation

The package installs through Unity's Package Manager from a git URL. It requires Unity 2022.3 or later.

Steps in the Unity Editor:

1. Open Window → Package Manager.
2. Click the + button, then select Add package from git URL....
3. Paste the following URL:

https://github.com/FunplayAI/funplay-unity-mcp.git

4. Click Add and wait for the package to resolve.

After installation, start the server via Funplay → MCP Server in the Unity menu bar. The server starts listening on the default address http://127.0.0.1:8765/.

Configuring Claude Code

Add the server to your Claude Code configuration file (claude_desktop_config.json):

{
  "mcpServers": {
    "funplay-unity": {
      "url": "http://127.0.0.1:8765/mcp"
    }
  }
}

Configuring Cursor

In Cursor, open Settings → MCP and add a new server entry pointing to `http