Tuanjie AI Doubles Unity Development Efficiency

This document provides detailed instructions on how to use Codely CLI and its powerful Unity Tools toolkit to directly manipulate the Unity Editor through natural language commands, achieving automation and intelligence in the game development workflow, thereby doubling development efficiency.

1. Why Does Unity Development Need AI?

In Unity game development, developers often encounter the following problems:

• Time-Consuming Repetitive Operations: Basic operations such as creating game objects, adding components, and adjusting properties require repeated clicks in the editor. A lot of time goes into repetitive tasks.

• Difficult Code Writing: Even for simple functions, you need to write complete C# code. Beginners don't know where to start, and experienced developers still need to frequently consult documentation.

• Tedious Debugging Process: When runtime errors occur, you need to check the console, locate the code, modify the script, and recompile. The entire process needs to be repeated many times.

• Scene Setup Overhead: Tasks such as batch creating objects, configuring materials, and setting physics properties require a lot of manual operations, which is both time-consuming and error-prone.

• Inconvenient Project Management: There are many scene files and resources scattered everywhere, lacking convenient viewing and management methods.

• High Team Collaboration Learning Costs: New members need to spend a long time becoming familiar with the project structure, resource organization methods, and code standards.

Codely CLI can help solve these problems. It supports describing what you want to do in natural language, and then automatically operates the Unity Editor to complete tasks, making Unity development simpler and more efficient.

2. Tuanjie AI Unity Tools Core Capabilities

Tuanjie AI provides 13 professional tools for Unity development, covering the complete development process from editor control to asset management:

Core Tools

Tool Category	Tool Name	Core Capability	Typical Scenarios
Editor Control	unity_editor	Playback control, compilation management, window management, tags and layers	Debugging, compilation, project configuration
Scene Management	unity_scene	Create/load/save scenes, query hierarchy	Scene switching, level design
Object Management	unity_gameobject	Create/modify/delete objects, component operations	Scene setup, prefab creation
Log Diagnosis	unity_console	Read console, filter errors	Bug debugging, log analysis
Script Development	unity_script	Create/edit/validate scripts	Code generation, quick modifications
Asset Management	unity_asset	Create materials/textures, search, batch operations	Asset organization, material configuration
Rendering Pipeline	unity_shader	Shader management, SRP detection	Rendering optimization, pipeline adaptation
Dependency Management	unity_package	Install/remove packages	Dependency management, feature extension
Scene Baking	unity_bake	NavMesh/light baking	AI navigation, GI optimization
UI Development	unity_ui_toolkit	UXML/USS management	UI system building
Automation	unity_menu	Execute menu commands	Batch processing, tool integration
Visualization	unity_screenshot	Multi-camera screenshots	Documentation, testing records
Extensibility	execute_custom_tool	Custom tool execution	Project-specific functionality

Core Advantages

✅ Natural Language Interaction: No need to memorize complex APIs, operate Unity by speaking

✅ Batch Automation: Complete operations that originally required dozens of clicks with a single command

✅ Intelligent Error Diagnosis: Automatically analyze console errors and provide fix suggestions

✅ Seamless Integration: Direct communication with Unity Editor with real-time feedback

✅ High Extensibility: Supports custom tools, adaptable to any project needs

3. Practical Guide: Unity Project from Scratch

Next, we will demonstrate through a complete practical case how to use Tuanjie AI to create a simple 3D game scene from scratch, experiencing the AI-assisted development process.

Practical Goals

Create a 3D scene containing the following elements:

• Walkable ground platform
• Controllable player character
• Patrolling enemy AI
• Collectible items
• Complete navigation mesh and lighting

1. Environment Preparation

1.1 Install Codely CLI

Execute the following command in the terminal to install Codely CLI:

# macOS/Linux
curl -fsSL https://codesearch-plugins.tos-cn-shanghai.volces.com/codely-cli/install.sh | bash

# Windows (PowerShell)
irm https://codesearch-plugins.tos-cn-shanghai.volces.com/codely-cli/install.ps1 | iex

After installation, verify the installation:

codely --version

1.2 Login to Tuanjie AI

Codely CLI supports login via Unity ID account (Tuanjie AI OAuth authentication).

When using for the first time, start interactive mode in the project directory:

# Start Tuanjie AI interactive mode
codely

Enter in interactive mode:

/auth

Codely CLI will automatically open the browser and navigate to the authentication page. Follow the prompts to complete login. After successful authentication, your credentials will be cached locally, and you won't need to login again for subsequent use.

2. Project Initialization

2.1 Create Unity Project

Open Unity Hub and create a new 3D (URP) project named TuanjieAIDemo.

2.2 Install Codely Bridge Plugin

In the Unity project directory, start Codely CLI interactive mode:

# Enter Unity project directory
cd /path/to/your/unity/project

# Start Tuanjie AI interactive mode
codely

In interactive mode, enter the following commands in sequence:

# After install, click unity, wait for unity to download and install the plugin
/upm install

/upm refresh

/upm status

After installation, /upm status will display the plugin's installation status and version information. Confirm successful plugin installation:

• Status shows cn.tuanjie.codely.bridge is installed
• Open Unity Editor and check if the Tools-Codely Bridge menu appears in the menu bar
• You can see the Codely Bridge package in Package Manager

2.3 Generate Project Summary

Enter in Tuanjie AI interactive mode:

Analyze the current Unity project and generate a project summary document CODELY.md

Tuanjie AI will automatically scan the project and generate a document containing the following information:

• Project type (3D/2D)
• Rendering pipeline (URP/HDRP/Built-in)
• Installed packages
• Scene list
• Scripting define symbols

3. Scene Setup: Create Game World in One Sentence

The traditional method requires dozens of clicks in the editor. Now in interactive mode, you only need one command:

Help me build a 3D game scene:
1. Create a 10x10 Plane as ground, named Ground
2. Add MeshCollider component to the ground
3. Create a Capsule as player, position at (0, 1, 0), named Player
4. Add Rigidbody and CapsuleCollider to Player
5. Create 3 Cubes as enemies, randomly distributed in the scene, named Enemy1, Enemy2, Enemy3
6. Add Enemy tag to all enemies
7. Create a Directional Light
8. Create a Main Camera, position at (0, 5, -10), facing the center of the scene
9. Save the scene as GameLevel

Execution Process: Tuanjie AI will automatically:

• Call unity_scene.create to create the scene
• Call unity_gameobject.create to create objects one by one
• Call unity_gameobject.add_component to add components
• Call unity_editor.ensure_tag to ensure tags exist
• Call unity_scene.save to save the scene

The entire process is completed automatically without manual operation!

4. Materials and Beautification: Intelligent Asset Management

4.1 Create Materials

Create three materials:
1. PlayerMat.mat - Blue material, metallic 0.5, smoothness 0.8
2. EnemyMat.mat - Red material, metallic 0.3, smoothness 0.6
3. GroundMat.mat - Gray material, using Standard Shader

4.2 Apply Materials

Assign PlayerMat to Player's MeshRenderer,
Assign EnemyMat to all objects with Enemy tag,
Assign GroundMat to Ground

Tuanjie AI will automatically:

• Call unity_asset.create to create materials
• Call unity_gameobject.set_component_property to set material references

5. Script Development: AI-Assisted Programming

5.1 Create Player Controller

Create a PlayerController script that implements the following functionality:
1. Use WASD to control movement, movement speed 5
2. Use mouse to control view rotation
3. Press space to jump, jump force 5
4. Include complete comments and error handling

5.2 Create Enemy AI

Create EnemyAI script to implement patrol functionality:
1. Patrol back and forth between two points
2. Patrol speed 2
3. Wait for 1 second after reaching target point
4. Use Gizmos to visualize patrol path

5.3 Auto-Compile and Validate

Clear console, compile all scripts, wait for compilation to complete, and if there are errors please tell me in detail

Tuanjie AI will:

• Call unity_console.clear to clear logs
• Call unity_editor.start_compilation_pipeline to compile
• Call unity_editor.wait_for_compile to wait for completion
• Call unity_console.get to read errors
• Automatically analyze and provide fix suggestions

6. Component Configuration: Batch Intelligent Operations

6.1 Add Scripts to Objects

Add PlayerController component to Player,
Add EnemyAI component to all objects with Enemy tag

6.2 Configure Component Properties

Set Player's PlayerController.moveSpeed to 8,
Set all Enemies' EnemyAI.patrolSpeed to 3

7. Lighting Configuration

Configure scene lighting:
Set Directional Light intensity to 1.5, color to light yellow

8. Camera Configuration

The camera is a key component for rendering the scene and needs to be configured correctly to see the game scene.

8.1 Configure Main Camera

Configure Main Camera:
1. Set position to (0, 5, -10)
2. Set rotation to (15, 0, 0), making it slightly look downward
3. Set Field of View to 60
4. Set background color to light blue

8.2 Create Follow Camera (Optional)

If you need the camera to follow the player, you can create a camera controller:

Create a CameraController script:
1. Camera follows target object (Player)
2. Maintain relative position (0, 5, -10)
3. Smooth following using Vector3.Lerp
4. Support mouse wheel to adjust distance

Then add the script to Main Camera:

Add CameraController component to Main Camera,
Set target to Player object

9. Debugging and Testing

9.1 Test Run

Start playing the game, pause after 15 seconds, read all errors and warnings from console

9.2 Error Analysis

If there are errors, Tuanjie AI can automatically analyze:

Analyze errors in the console, locate specific code lines, and provide fix suggestions

10. Project Organization

10.1 Asset Organization

Search for all material files and move them to Assets/Materials folder

10.2 Create Prefabs

Save Player as prefab Assets/Prefabs/Player.prefab,
Also save all Enemy objects as independent prefabs

4. Summary

With Codely CLI and Unity Tools, Unity developers can:

✅ Say Goodbye to Repetitive Labor: Natural language commands replace hundreds of click operations

✅ Accelerate Script Development: AI-assisted generation of high-quality, commented code

✅ Intelligent Error Diagnosis: Automatically analyze problems and provide fix suggestions

✅ Batch Automation: Complete batch asset management and scene configuration with a single command

✅ Rapid Prototype Development: Quickly validate game ideas and mechanics

✅ 10-30x Efficiency Improvement: Reduce development time from hours to minutes

Whether for independent developers rapidly validating game ideas or large teams collaborating on complex projects, Tuanjie AI can become a powerful assistant to boost Unity development efficiency.