A premium, full-stack web application designed to help creators generate high-converting YouTube thumbnails instantly. Studio AI combines advanced prompt engineering, text-extraction heuristics, and generative AI to craft cinematic backgrounds with punchy, readable text overlays.
- Intelligent Title Extraction: Automatically condenses long video titles into a short hook and highlights the most click-worthy keyword.
- Generative Art Backgrounds: Integrates with Hugging Face's Stable Diffusion XL to generate vibrant, text-free background art based on the user's selected vibe.
- Premium Bento-Box UI: A sleek, dark-mode interface built with Tailwind CSS, featuring glassmorphism, dynamic aurora gradients, and real-time canvas rendering.
- Zero-Loss High-Res Export: Utilizes
html2canvasto composite the AI-generated background and CSS-styled text overlays into a single, high-resolution PNG ready for YouTube. - Fully Responsive: Seamlessly transitions from mobile view to wide desktop displays.
Frontend Architecture:
- Framework: React.js (via Vite)
- Styling: Tailwind CSS
- Icons: Lucide React
- Image Processing: html2canvas
Backend Architecture:
- Framework: FastAPI (Python)
- Server: Uvicorn
- AI Integration: Hugging Face Inference API (
requests) - Environment Management:
python-dotenv
Follow these steps to run Studio AI on your local machine.
- Node.js installed
- Python 3.8+ installed
- A free API token from Hugging Face
Open your terminal and navigate to the backend directory:
cd backend
# Create a virtual environment
python -m venv venv
# Activate the virtual environment
# Windows:
.\venv\Scripts\activate
# Mac/Linux:
source venv/bin/activate
# Install required dependencies
pip install fastapi uvicorn pydantic python-dotenv requests