█████╗ ██╗███████╗██╗  ██╗██╗  ██╗
  ██╔══██╗██║██╔════╝╚██╗██╔╝╚██╗██╔╝
  ███████║██║█████╗   ╚███╔╝  ╚███╔╝
  ██╔══██║██║██╔══╝   ██╔██╗  ██╔██╗
  ██║  ██║██║███████╗██╔╝ ██╗██╔╝ ██╗
  ╚═╝  ╚═╝╚═╝╚══════╝╚═╝  ╚═╝╚═╝  ╚═╝

AIEXX GenAI Image to 3D

🚀 Enterprise-Grade AI-Powered 3D Model Generation System

Transform Text & Images into Production-Ready 3D Assets in Seconds

🚀 Quick Start • 📖 Documentation • 🎨 Workflows • ⚙️ Tech Stack • 📊 Performance • 🤝 Contributing

🌟 Overview

AIEXX GenAI Image to 3D is a professional, GPU-accelerated AI system for generating high-quality 3D models from text prompts or 2D images. Built on cutting-edge deep learning technologies, it delivers production-ready 3D assets in seconds.

🎉 NEW: Standalone Image-to-3D System

Transform any image into a 3D model locally - no compilation required!

We've integrated TripoSR directly into the project with a custom implementation that works without pytorch3d. Simply drag and drop an image onto the launcher!

# Convert any image to 3D in seconds:
AIEXX_IMAGE_TO_3D.bat your_photo.png
# Output formats: OBJ, GLB, STL, PLY
# Processing time: 30-70 seconds
# Works 100% offline after first download

✅ Features:

• 🚀 No compilation needed (pure Python)
• 🗼️ Drag-and-drop interface
• 💻 Full GPU acceleration (RTX 5060)
• 🔐 100% local processing (privacy)
• ⚡ Professional quality output

📖 See COMO_USAR_IMAGE_TO_3D.md for complete guide

✨ Key Features

🎯 TEXT-TO-3D GENERATION
   ├─ Natural language prompts to 3D models
   ├─ Stable Diffusion 1.5 integration
   ├─ Advanced negative prompting
   └─ 20-60 seconds per model

🗼️ IMAGE-TO-3D CONVERSION
   ├─ Single image to full 3D mesh
   ├─ Multi-view reconstruction support
   ├─ Automated texture generation
   └─ 15-30 seconds processing time

⚡ GPU ACCELERATION
   ├─ NVIDIA RTX 5060 optimized
   ├─ CUDA 12.8 + cuDNN integration
   ├─ BF16 mixed precision training
   ├─ Flash Attention v2 support
   └─ PyTorch 2.7.0 performance mode

🎎 PROFESSIONAL OUTPUT
   ├─ GLB/GLTF format export
   ├─ PBR material support
   ├─ Configurable mesh resolution
   └─ Blender/Unity/Unreal ready

🚀 Quick Start

Prerequisites

Installation

Option 1: One-Click Automated Installation (⭐ EASIEST - Recommended)

# 1. Clone the repository
git clone https://github.com/AIExxplorer/AIEXX_GENAI_IMAGE_TO_3D.git
cd AIEXX_GENAI_IMAGE_TO_3D

# 2. Run the automated installer
0-INSTALL_ALL_AUTOMATED.bat

# That's it! The installer will:
# ✅ Check system requirements
# ✅ Install all dependencies automatically
# ✅ Handle restart (run again after restart)
# ✅ Download all models
# ✅ Launch the system when done

Features:

• ✨ Single command installation
• 🔄 Auto-resume after restart
• 🛡️ Error handling and recovery
• 📊 Progress tracking with checkpoints
• 💡 Clear instructions at each step

See INSTALACAO_AUTOMATIZADA.md for complete documentation.

Option 2: Step-by-Step Installation (Advanced Users)

# 1. Clone the repository
git clone https://github.com/AIExxplorer/AIEXX_GENAI_IMAGE_TO_3D.git
cd AIEXX_GENAI_IMAGE_TO_3D

# 2. Install Visual Studio Build Tools (⚠️ RESTART after this!)
1-INSTALL_VS_BUILDTOOLS_WINGET.bat

# 3. Setup Python environment and dependencies
2-FIX_VENV_AND_INSTALL_TORCH_SCATTER.bat

# 4. Install ComfyUI 3D Pack
3-RUN_INSTALL_3DPACK.bat

# 5. Download AI models
4-DOWNLOAD_3D_MODELS.bat

# 6. (Optional) Download Stable Diffusion models
5-SETUP_SD_MODELS.bat

# 7. Launch the system
6-START_AIEXX.bat

Option 3: Manual Setup

See INSTALLATION.md for detailed manual installation instructions.

First Run

1. Start the System

   6-START_AIEXX.bat
   

2. Access Web Interface

   • Open browser: http://localhost:8188

3. Load a Workflow

   • Click "Load" → Select from workflows/3d/
   • Recommended first workflow: 03_triposr_single_image_to_mesh.json

4. Generate Your First 3D Model

   • Upload an image or enter a text prompt
   • Click "Queue Prompt"
   • Wait ~20-60 seconds
   • Download your .glb file from ComfyUI/output/

📋 Note: See ORDEM_DE_INSTALACAO.md for complete installation order and troubleshooting.

📖 Documentation

🎨 Workflows

Available Workflows

🟢 Basic - TripoSR

03_triposr_single_image_to_mesh.json

Pipeline:

Image > TripoSR > GLB

Specs:

• Time: ~15-30s
• VRAM: 3-4GB
• Best for: Quick prototypes

🔵 Intermediate - Prompt to 3D

basic/AIEXX_prompt_to_3d_triposr.json

Pipeline:

Prompt > SD 1.5 > TripoSR > GLB

Specs:

• Time: ~25-40s
• VRAM: 4-5GB
• Best for: Text-based creation

🛳 Advanced - InstantMesh

advanced/AIEXX_prompt_to_3d_instantmesh.json

Pipeline:

Prompt > SD > Zero123++ > InstantMesh > GLB

Specs:

• Time: ~45-90s
• VRAM: 5-6GB
• Best for: High-quality output

⚙️ Technology Stack

Core AI/ML Technologies

📊 Performance Benchmarks

Processing Times (RTX 5060 8GB)

💫 Acknowledgments

This project builds upon these amazing open-source projects:

• ComfyUI - Node-based Stable Diffusion GUI
• Stable Diffusion - Text-to-image foundation
• TripoSR - Fast 3D reconstruction from single images
• PyTorch - Deep learning framework
• PyTorch3D - 3D deep learning library

Special thanks to:

• Stability AI for Stable Diffusion
• NVIDIA for CUDA toolkit and GPU support
• Hugging Face for model hosting
• The entire open-source AI community

🔧 Support & Community

• Issues: GitHub Issues
• Docs: GitHub Wiki
• Email: aiexxplorer@artificialuniverse.tech

💫 License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.

🚀 Ready to Transform Ideas into 3D Reality?

6-START_AIEXX.bat

⬆ Back to Top

Built with ❤️ by AIEXX

Optimized for NVIDIA RTX 5060 | Powered by AI