Product Concept Generator

A Streamlit-based web application that transforms product sketches into refined concept images, 3D models, and animated videos using AI. Perfect for product designers, concept artists, and creative professionals.

Features

• 🖼️ AI-powered concept image generation from sketches
• 🧊 3D model generation from concept images
• 🎬 Video animation generation with customizable parameters
• 📊 Real-time generation statistics and cost tracking
• 💾 Download capabilities for all generated assets

Prerequisites

• Python 3.8 or higher
• Streamlit
• API Keys for:
  • Google AI (Gemini 2.0 Flash)
  • Fal.ai
  • (Store these in your .env file)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/design-tool-simple.git
   cd design-tool-simple

2. Create and activate the Conda environment:

   conda env create -f environment.yml
   conda activate design-tool-simple

3. Set up your environment variables:

   • Copy .env.example to .env
   • Add your API keys to the .env file

Usage

1. Start the Streamlit app:

   streamlit run main.py

2. Navigate through the three main tabs:

   • Generate Concept: Upload sketches and generate refined concept images
   • Generate 3D: Convert concepts into 3D models
   • Generate Video: Create animations from your concepts

Generation Process

1. Concept Generation:

   • Upload a product sketch
   • Get AI-generated prompt using Gemini 2.0 Flash
   • Generate concept image using FAL.ai's fast-sdxl model
   • Download the generated image

2. 3D Model Creation:

   • Use your generated concept
   • Create an interactive 3D model using FAL.ai's hyper3d/rodin model
   • Download in GLB format with PBR materials

3. Video Animation:

   • Customize action, style, camera motion, and ambiance
   • Choose aspect ratio and duration (5-8 seconds)
   • Generate video using FAL.ai's veo2/image-to-video model
   • Download the animated video

AI Models Used

FAL.ai Models

1. fast-sdxl

   • Purpose: Concept image generation
   • Features:
     • Image strength: 0.95
     • Guidance scale: 7.5
     • Scheduler: DPM++ 2M
     • 30 inference steps

2. hyper3d/rodin

   • Purpose: 3D model generation
   • Features:
     • PBR material support
     • GLB file format
     • Texture generation
     • Interactive viewer support

3. veo2/image-to-video

   • Purpose: Video animation
   • Features:
     • Customizable duration (5-8s)
     • Multiple aspect ratios
     • Action and style prompting
     • Camera motion control

Google AI

• Gemini 2.0 Flash
  • Purpose: Prompt generation and optimization
  • Used for converting user inputs into effective generation prompts

Configuration

Create a .env file with your API keys:

GOOGLE_API_KEY=your_google_api_key_here
FAL_KEY=your_fal_ai_key_here

Project Structure

product-concept-generator/
├── main.py           # Streamlit application
├── .env              # API keys (not tracked)
├── .env.example      # Example environment file
├── environment.yml   # Conda environment specification
├── .gitignore       # Git ignore rules
└── README.md        # Project documentation

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License.