MCP提交

Unitymcpintegration

该项目提供了Unity编辑器与Model Context Protocol (MCP)的无缝集成，使AI助手能够实时理解和操作Unity项目，包括访问场景层次结构、执行代码、监控日志等功能。

开发者工具游戏与游戏化 #Unity集成 #AI协作 #实时交互 #编辑器扩展本地 .C#

2.5分

2025-04-28 14:34:28

概述

安装

工具列表

内容详情

替代品

What is Unity MCP Integration?

Unity MCP Integration is a bridge between your Unity development environment and AI assistants. It allows AI to understand your Unity project structure, access scene information, execute code, and manipulate files - all in real-time through the Model Context Protocol.

How does it work?

The system consists of a Unity plugin and a Node.js server that communicate via WebSocket. AI assistants connect to the server which then relays commands to your Unity Editor, enabling two-way interaction with your project.

When should I use it?

Ideal for: Automating repetitive Unity tasks, debugging assistance, project analysis, rapid prototyping, and getting AI-powered suggestions while developing in Unity.

Key Features

Full Editor AccessAI can inspect and modify your Unity project just like a developer would through the Editor interface

Project File ManagementRead, edit, and create project files including scripts, scenes, and assets

Safe Code ExecutionRun C# code snippets directly in the Editor context with safety checks

Real-time Project MonitoringAccess current scene hierarchy, game objects, and Editor logs in real-time

Pros and Cons

Advantages

Saves development time by automating repetitive tasks

Enables AI-assisted debugging and problem solving

Provides complete project visibility to AI helpers

Secure sandboxed execution environment

Limitations

Requires Unity 2021.3+ and Node.js 18+

WebSocket connection means both systems must be running

Some advanced Editor features may require custom implementation

Getting Started

Install the Unity PackageAdd the package to your Unity project through Package Manager or by importing the unityPackage

Set Up MCP ServerInstall and run the Node.js server that handles communication between AI and Unity

Connect Your AI AssistantConfigure your AI tool (like Claude) to connect to the MCP server

Start InteractingBegin asking your AI assistant to help with Unity tasks

Example Use Cases

Scene AnalysisHave AI analyze your current scene structure and suggest optimizations

Script ModificationRequest AI to modify existing scripts with specific changes

Asset ManagementOrganize project assets with AI assistance

Frequently Asked Questions

Is this safe to use with my production projects?Yes, all file operations are sandboxed to your project directory and code execution includes safety checks. Always use version control for important projects.

What Unity versions are supported?The package requires Unity 2021.3 or later. Some features may work better in newer versions.

Can multiple AI assistants connect at once?The server supports multiple connections but simultaneous edits may cause conflicts. Recommended to use one active assistant at a time.

How do I know what the AI is doing in my project?The MCP Debug window shows all activity, and file edits include diff previews before application.

Additional Resources

MCP Protocol DocumentationOfficial Model Context Protocol documentation

GitHub RepositorySource code and issue tracking

Demo VideoVideo demonstration of features

Smithery InstallationOne-click installation for Claude Desktop

精选MCP服务推荐

Unity MCP 集成

📦 概述

Unity MCP 集成允许 AI 助手通过 MCP 协议与 Unity 编辑器交互。该集成提供了一套完整的工具集，使 AI 助手能够执行代码操作、管理文件系统以及监控 Unity 项目状态。

✨ 主要功能

🛠️ Unity 编辑器工具

get_editor_state: 获取关于 Unity 项目和编辑器状态的详细信息
get_current_scene_info: 获取当前场景的详细信息
get_game_objects_info: 获取特定游戏对象的信息
execute_editor_command: 在 Unity 编辑器中直接执行 C# 代码
get_logs: 检索并过滤 Unity 控制台日志
verify_connection: 检查与 Unity 编辑器是否存在有效连接

📁 文件系统工具

read_file: 读取项目中的文件内容
read_multiple_files: 一次性读取多个文件
write_file: 创建或覆盖指定文件
edit_file: 对现有文件进行针对性编辑，带有预览功能
list_directory: 获取目录中的文件和文件夹列表
directory_tree: 获取目录的层次结构视图
search_files: 查找符合特定模式的文件
get_file_info: 获取特定文件或目录的元数据
find_assets_by_type: 根据类型查找所有资源（例如 Material、Prefab）
list_scripts: 获取项目中所有 C# 脚本的列表

文件路径可以是绝对路径，也可以是相对于 Unity 项目 Assets 文件夹的相对路径。例如，"Scenes/MyScene.unity" 实际指向 <project>/Assets/Scenes/MyScene.unity。

🏷️ 环境要求

Unity 编辑器: 支持 2018.4 及以上版本
Node.js: 版本 16.0 或更高

🔍 安装指南

使用 npm 安装

npm install @quaza/unitymcpintegration

🛠️ 开发环境安装

Unity 插件

在 Unity 中打开项目
修改 UnityMCPConnection/Editor 目录下的 C# 脚本

MCP 服务器端

切换到 mcpServer 目录
安装依赖: npm install
编译服务器: npm run build
运行服务器: node build/index.js

📋 快速上手

初始化 MCP 服务

在项目根目录下运行以下命令：

npx @quaza/unitymcpintegration init

这将创建必要的配置文件和目录结构。

启动 MCP 服务器

运行以下命令启动 MCP 服务：

npm run start

🛠️ 使用说明

📈 调试与监控

在 Unity 中打开 MCP 调试窗口以监控连接状态和测试功能：

进入 Window > MCP Debug
在调试窗口中：
- 检查连接状态
- 测试代码执行
- 查看日志
- 监控事件

🎯 可用工具

Unity 编辑器工具

get_editor_state: 获取关于 Unity 项目和编辑器状态的详细信息
get_current_scene_info: 获取当前场景的详细信息
get_game_objects_info: 获取特定游戏对象的信息
execute_editor_command: 在 Unity 编辑器中直接执行 C# 代码
get_logs: 检索并过滤 Unity 控制台日志
verify_connection: 检查与 Unity 编辑器是否存在有效连接

文件系统工具

read_file: 读取项目中的文件内容
read_multiple_files: 一次性读取多个文件
write_file: 创建或覆盖指定文件
edit_file: 对现有文件进行针对性编辑，带有预览功能
list_directory: 获取目录中的文件和文件夹列表
directory_tree: 获取目录的层次结构视图
search_files: 查找符合特定模式的文件
get_file_info: 获取特定文件或目录的元数据
find_assets_by_type: 根据类型查找所有资源（例如 Material、Prefab）
list_scripts: 获取项目中所有 C# 脚本的列表

📝 示例代码

在 Unity 中使用

using MCPIntegration;

public class MCPExample : MonoBehaviour {
    void Start() {
        // 初始化 MCP 客户端
        MCPClient.Initialize("localhost", 3000);
        
        // 获取编辑器状态
        EditorState state = MCPClient.GetEditorState();
        Debug.Log($"Editor Version: {state.Version}");
    }
}

在 Node.js 中使用

const mcpServer = require('@quaza/unitymcpintegration');

// 初始化服务器
mcpServer.initialize('localhost', 3000);

// 监听文件变化事件
mcpServer.on('fileChanged', (event) => {
    console.log(`File changed: ${event.path}`);
});