🧊 AI-Powered Text-to-3D Model Generation

A GenAI-powered web application that converts natural language text prompts into 3D models using a multi-stage AI pipeline.
The system first generates an image from text and then reconstructs a 3D mesh from that image, all through an interactive web interface.

---

🎥 Demo Video

Text-to-3D.mp4

---

📊 Model Overview

• Pipeline:
  • Text → Image (SDXL)
  • Image → 3D Mesh (TripoSR)
• Input: Natural language text prompt
• Output: 3D model files (.obj, .glb, etc.)

---

🔍 Core Components

• Text-to-Image Generation
  SDXL model accessed via Cloudflare Workers AI.

• Image-to-3D Reconstruction
  TripoSR for single-image 3D mesh reconstruction.

• User Interface
  Streamlit-based interactive web application.

---

⚙️ Enhancements

• Automatic preprocessing of generated images
• Temporary file handling for intermediate pipeline outputs
• Support for multiple 3D export formats
• Modular pipeline design for easy extension
• Streamlit-controlled execution flow
• GPU-accelerated inference support (recommended)

---

🛠 Tech Stack

Layer	Tools Used
UI / App	Streamlit
AI Models	SDXL, TripoSR
ML / Utils	PyTorch, NumPy, Pillow
External API	Cloudflare Workers AI
Deployment	Local (Cloud optional)

---

🚀 How to Run Locally

1️⃣ Clone the Repository

git clone https://github.com/Its-Itachi/Text-to-3D.git
cd Text-to-3D

2️⃣ Create a Virtual Environment

python -m venv venv

3️⃣ Activate the Virtual Environment

Windows (PowerShell):

venv\Scripts\Activate.ps1

or

venv\Scripts\activate

macOS / Linux:

source venv/bin/activate

4️⃣ Install Dependencies

pip install -r requirements.txt

---

🔑 Environment Configuration (SDXL)

Set Cloudflare credentials using environment variables:

CF_ACCOUNT_ID=your_account_id
CF_API_TOKEN=your_api_token

Or via Streamlit secrets:

.streamlit/secrets.toml

---

▶ Run the Application

streamlit run app.py

Open in browser:

http://localhost:8501

---

📦 Output Structure

output/
 └── run_id/
     ├── input.png
     └── mesh.obj

Temporary files are stored in:

temp/

---

👤 Author

Jayesh Dethe

GitHub: https://github.com/Its-Itachi

---

⭐ Support

If you find this project useful:

• ⭐ Star the repository
• 📢 Share it with peers
• 🔔 Follow for more AI + 3D projects

---

📝 Notes

• GPU is strongly recommended for faster 3D reconstruction
• Output quality depends heavily on the generated image
• Complex prompts may produce noisy or incomplete meshes

---

Happy building in 3D! 🚀