✦
21 GitHub stars with active development
✦
Generates 3D models from images and videos using Hunyuan 3D 2.1
✦
Custom ComfyUI nodes for game developers
What is ComfyUI AI GameDev? Features and use cases explained.
About ComfyUI AI GameDev
ComfyUI AI GameDev Nodes is a custom node pack for ComfyUI that brings AI-powered game asset generation into your existing workflow. It includes three main nodes: Hunyuan3D 2.1 for converting images and videos into textured 3D models, a Video Frame Extractor for smart multi-view sampling, and a Prompt Optimizer that uses local vision-language models to enhance prompts for various AI generators. Designed for game developers, these nodes output game-ready GLB/OBJ files and integrate seamlessly with Unity, Unreal, and Godot. Requirements include 10-21GB VRAM, Blender, and optional Ollama for prompt enhancement.
Use Cases
- ▸Generate a 3D prop from a reference photo for use in a game level
- ▸Convert a turntable video into a multi-view 3D model for asset creation
- ▸Enhance a simple text prompt to generate consistent art and audio assets across a project
Features
- ✓Converts photos or concept art into textured 3D models ready for Unity, Unreal, or Godot
- ✓Extracts evenly-spaced frames from videos for 3D reconstruction and multi-view workflows
- ✓Transforms simple prompts into detailed, model-specific prompts using vision-language models
- ✓Supports single or multi-view input (up to 10 angles) for higher quality 3D generation
- ✓Generates optimized GLB files with embedded PBR textures, reducing file size by up to 77%
FAQ
Does this work offline?▼
Yes, all nodes run locally on your GPU. Hunyuan3D requires 10GB VRAM (shape only) or 21GB VRAM (with textures) and Blender installed.
Can I use these nodes in my existing ComfyUI workflow?▼
Absolutely. Each node is a custom ComfyUI node that plugs into any workflow. The Video Frame Extractor outputs a standard image batch, and the Prompt Optimizer generates text inputs for other nodes.
What are the system requirements for Hunyuan3D?▼
Minimum 10GB VRAM for untextured models, 21GB VRAM for textured output. It runs in a separate conda environment to avoid conflicts with ComfyUI's Python dependencies.
Do I need Ollama for all nodes?▼
No. Only the Prompt Optimizer node requires Ollama with a vision model (e.g., qwen2.5vl:7b). The other nodes work independently.
How to install ComfyUI AI GameDev — setup guide and prerequisites.
Installation
Install ComfyUI AI GameDev to get started:
$git clone https://github.com/mattwilliamson/comfyui-ai-gamedev.git
Or visit the project page for detailed setup instructions.
Similar Tools
tilemapgen
Generate isometric dungeon tiles using stable diffusion. Streamlines tilemap creation for game levels with AI-generated art.
Prompt
SpriteFusion Pixel Snapper
Snaps pixels to a perfect grid. Fixes messy and inconsistent pixel art from AI generation.
Leonardo AI
AI image generation for game assets. Generate concept art, textures, sprites, and UI elements with fine-tuned models.
Luma AI
Generate videos and images with AI agents. Automates end-to-end creative workflows for game asset production and promotional content.
Skill
GPTGame
Generate simple 2D games from natural language descriptions. Instant playable prototypes with AI-generated code and assets.
audiocraft
Deep learning library for audio processing and generation. Features EnCodec audio compression and MusicGen for controllable music generation with text and melody conditioning.