A modern web application for AI-powered image generation using Flux models on Apple Silicon, with intelligent prompt enhancement via OpenRouter.

• Local Image Generation: Uses mflux for Flux 1 Schnell model on Apple Silicon (MLX)
• AI Prompt Enhancement: Automatically enhances prompts using 300+ AI models from OpenRouter
• Quality Presets: Basic (2 steps), Standard (4 steps), High (8 steps)
• Multiple Resolutions: Square, Landscape, Portrait, Cinematic, Classic, Photo
• Gallery: View, download, and manage generated images
• Real-time Progress: Visual feedback during generation
• Dark Theme: Modern, elegant dark UI

• FastAPI - High-performance async API
• mflux - MLX-based Flux model for Apple Silicon
• OpenRouter API - Access to 300+ AI models for prompt enhancement

• Vanilla JS - No framework dependencies
• CSS Variables - Themeable design system
• Responsive - Works on desktop and tablet

• macOS with Apple Silicon (M1/M2/M3)
• Python 3.10+
• OpenRouter API key (optional, for prompt enhancement)

git clone https://github.com/BTankut/fluxstudio.git
cd fluxstudio

cd backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

export OPENROUTER_API_KEY="your-api-key-here"
export MFLUX_MODEL="schnell"  # or "dev" for higher quality
export MFLUX_QUANTIZE="4"     # 4, 5, 6, or 8 bit quantization

cd backend
source .venv/bin/activate
python -m uvicorn main:app --host 0.0.0.0 --port 8000

cd frontend
python -m http.server 3000

Navigate to http://localhost:3000

fluxstudio/
├── backend/
│   ├── main.py           # FastAPI application
│   ├── mflux_client.py   # mflux integration
│   ├── openrouter.py     # OpenRouter API client
│   └── requirements.txt  # Python dependencies
├── frontend/
│   ├── index.html        # Main HTML
│   ├── app.js            # Application logic
│   └── styles.css        # Styling
├── outputs/              # Generated images
└── README.md

MIT License

Built by BTankut