Back to Projects
PythonComfyUIDiffusion ModelVLM

AI Game Asset Generator

1. Project Overview

An automated system that generates various assets needed for game development using AI. By combining Diffusion Models and LLMs, high-quality game graphic assets can be efficiently created.

2. Key Features

  • Automatic Image Generation: High-quality game art generation based on Stable Diffusion
  • Prompt Optimization: Automatic prompt engineering using LLM
  • Batch Processing: Workflow for bulk asset generation
  • Style Consistency: Maintaining unified art style throughout the game

3. Tech Stack

TechnologyPurpose
PythonBackend development and scripting
ComfyUIImage generation pipeline
Stable DiffusionImage generation model
LLM (GPT-4)Prompt optimization

4. Development Process

4.1. Phase 1: Requirements Analysis

Analyzed the types of assets needed in game development and identified areas that could be automated.

4.2. Phase 2: Pipeline Design

Built a reusable image generation pipeline using ComfyUI.

4.3. Phase 3: LLM Integration

Developed an LLM-based prompt optimization system to improve prompt quality.

5. Results and Achievements

  • 70% reduction in asset production time
  • Improved game quality with consistent art style
  • Created an environment where developers can focus more on creative work

6. Image Assets (PNG)

Use the following path convention to keep markdown image references clean and consistent.

  • Folder: /public/images/projects/ai-game-assets/
  • Markdown path: /images/projects/ai-game-assets/<file-name>.png

6.1. Example Images