Skip to content

Commit f3bfc9c

Browse files
author
NishantJLU
committed
Modernize README with v1.0.0 release info, Quick Start, and Community links
1 parent b1f8993 commit f3bfc9c

1 file changed

Lines changed: 60 additions & 31 deletions

File tree

README.md

Lines changed: 60 additions & 31 deletions
Original file line numberDiff line numberDiff line change
@@ -1,43 +1,72 @@
1-
# Unity-MCP Pro — Agentic Unity Bridge
2-
**Connects LLMs (Claude, GPT) to Unity in real-time via MCP Protocol**
1+
# 🌌 Unity-MCP Pro: The Agentic Unity Bridge
32

3+
[![GitHub release (latest by date)](https://img.shields.io/github/v/release/NishantJLU/Unity-MCP?color=00f2ff&style=for-the-badge)](https://github.com/NishantJLU/Unity-MCP/releases)
4+
[![Status: Production Ready](https://img.shields.io/badge/Status-Production--Ready-00f2ff?style=for-the-badge)](https://github.com/NishantJLU/Unity-MCP)
45
[![CodeQL](https://github.com/NishantJLU/Unity-MCP/actions/workflows/codeql.yml/badge.svg)](https://github.com/NishantJLU/Unity-MCP/actions)
56
[![Node.js CI](https://github.com/NishantJLU/Unity-MCP/actions/workflows/node.js.yml/badge.svg)](https://github.com/NishantJLU/Unity-MCP/actions)
6-
![License](https://img.shields.io/badge/License-MIT-blue.svg)
7+
[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg?style=for-the-badge)](https://opensource.org/licenses/MIT)
78

8-
## 🏗️ How It Works
9-
The bridge facilitates bidirectional communication between Large Language Models and the Unity Engine.
9+
**Unity-MCP Pro** is a high-performance bridge that transforms the Unity Editor into an agentic environment. It enables LLMs (like Claude 3.5 Sonnet and Gemini 1.5 Pro) to interact with Unity in real-time for AI-driven level design, scene auditing, and live construction.
1010

11-
```text
12-
LLM (Claude/GPT) ──▶ MCP Server (Node.js) ──▶ WebSocket/HTTP ──▶ Unity (C# Bridge)
13-
▲ │
14-
└───────────────────────────────────────────────────────────┘
15-
World State & Feedback
11+
---
12+
13+
## 🚀 Quick Start (3-Minute Setup)
14+
15+
### 1. Download & Install
16+
Download the **[Latest Release (v1.0.0)](https://github.com/NishantJLU/Unity-MCP/releases/latest)** and extract it. Double-click **`setup.bat`** to install dependencies automatically.
17+
18+
### 2. Add to Unity
19+
Drag the `UnityScripts` folder into your Unity project's `Assets` folder. Open **Window > AI > Unity MCP Pro** and click **ESTABLISH LINK**.
20+
21+
### 3. Connect to AI
22+
Add this snippet to your `claude_desktop_config.json`:
23+
```json
24+
{
25+
"mcpServers": {
26+
"unity-mcp-pro": {
27+
"command": "node",
28+
"args": ["C:/PATH/TO/Unity-MCP/unity-mcp-server/index.js"]
29+
}
30+
}
31+
}
1632
```
1733

18-
## 🛠️ Prerequisites
19-
- **Node.js**: 18.x or higher
20-
- **Unity Engine**: 2021.3 LTS or higher
21-
- **LLM API Key**: OpenAI or Anthropic key for the MCP client
34+
---
35+
36+
## ❓ Why Unity-MCP Pro?
37+
While basic scripts can move objects, **Unity-MCP Pro** provides a deep, production-grade integration:
38+
* **👁️ AI Vision:** AI can "see" through your Editor camera to reason about layout and aesthetics.
39+
* **⚡ Reflection-Based:** No need to write new tools. AI can access *any* public variable on *any* component via C# Reflection.
40+
* **🛡️ Safety First:** All commands are dispatched via the Main Thread to prevent Editor crashes.
41+
* **📊 Full Context:** AI receives a high-fidelity JSON map of your scene hierarchy, tags, and layers.
42+
43+
---
44+
45+
## 💎 Pro Features
46+
47+
### AI Vision (Visual Feedback)
48+
Empower your AI to capture live snapshots of the Editor, allowing it to reason about material properties, lighting, and UI placement visually.
49+
50+
### Scene Graph Analytics
51+
The AI receives a structured map of the entire scene hierarchy, enabling complex architectural reasoning and automated auditing.
52+
53+
### Reflection-Based Control
54+
Modify deep parameters like Light intensity, Rigidbody mass, or custom script variables without needing pre-defined tools.
55+
56+
---
2257

23-
## 🚀 Setup Guide
58+
## 🏗️ Technical Architecture
2459

25-
### 1. Node.js MCP Server
26-
1. Navigate to the `server/` directory.
27-
2. Run `npm install`.
28-
3. Build the server: `npm run build`.
29-
4. Configure your MCP client (e.g., Claude Desktop) to use the built `index.js`.
60+
- **The Brain (Node.js):** An MCP server that exposes Unity functionality as structured tools.
61+
- **The Hand (C#):** A custom Editor window with a high-performance HTTP listener.
62+
- **Communication:** Low-latency JSON bridge for real-time AI command execution.
3063

31-
### 2. Unity Bridge
32-
1. Open your Unity project.
33-
2. Install the package via UPM: `https://github.com/NishantJLU/Unity-MCP.git?path=/unity`.
34-
3. Open the bridge window: **Window > AI > Game Master Bridge**.
35-
4. Click **Start Bridge**.
64+
---
3665

37-
## 🎮 Use Cases
38-
- **AI-Driven NPC Behavior**: Generate dynamic dialogue and actions based on world state.
39-
- **Procedural Level Generation**: Use natural language to describe and spawn complex environments.
40-
- **Game Testing Automation**: Instruct an AI agent to "find bugs" or "playtest level 1" via the bridge.
66+
## 💬 Community & Feedback
67+
* **Discussion:** Have an idea for a new tool? [Join the Discussion](https://github.com/NishantJLU/Unity-MCP/discussions).
68+
* **Bugs:** Report issues [here](https://github.com/NishantJLU/Unity-MCP/issues).
69+
* **Support:** If this helps your workflow, please consider giving us a ⭐!
4170

42-
## 📜 License
43-
This project is licensed under the MIT License.
71+
---
72+
**Developed by [NishantJLU](https://github.com/NishantJLU)**

0 commit comments

Comments
 (0)