NectarGraphix 0.1.3

Image generation studio

NectarGraphix

NectarGraphix is a powerful, local image generation studio powered by Stable Diffusion. Run cutting-edge AI models directly on your hardware—no cloud dependency, no usage limits. Generate stunning visuals from text prompts with full control over models, settings, and outputs.

✨ Features

• Local Stable Diffusion: Run models like SDXL, SD 1.5, or custom fine-tunes offline.
• Intuitive UI: Drag-and-drop interface for prompts, images, and settings.
• Model Management: Download, switch, and organize models seamlessly.
• Advanced Controls: Negative prompts, samplers, CFG scale, steps, and resolution tweaking.
• Batch Generation: Create multiple images at once.
• Hardware Optimization: GPU/CPU detection with automatic optimizations (CUDA, ROCm, DirectML).
• Extensions Support: Compatible with popular LoRAs, ControlNet, and embeddings.

💻 Requirements

• OS: Windows 10+, macOS 10.15+, Linux (Ubuntu 20.04+ recommended)
• Python: 3.10 or 3.11
• GPU (recommended):
  • NVIDIA: 4GB+ VRAM (GTX 1060 or newer)
  • AMD: ROCm-compatible (RX 5000+)
  • Apple Silicon: M1/M2/M3/M4
• RAM: 8GB+ (16GB recommended for larger models)
• Disk: 10GB+ free space for models and outputs

🚀 Quick Start

1. Clone & Install

git clone https://github.com/yourusername/NectarGraphix.git cd NectarGraphix pip install -r requirements.txt

2. Download Models

Launch once to auto-download a starter model (SD 1.5), or place models in models/Stable-diffusion/:

models/ ├── Stable-diffusion/ │ └── v1-5-pruned-emaonly.safetensors ├── Lora/ └── VAE/

Popular sources: Civitai, Hugging Face

3. Run the App

python app.py

Or use the one-click launcher: launch.bat (Windows) / launch.sh (Linux/macOS).

The UI opens at http://127.0.0.1:7860.

⚙️ Configuration

Edit config.yaml for custom defaults:

webui: port: 7860 share: false # Enable public URL (ngrok) models: default: "v1-5-pruned-emaonly.safetensors" hardware: gpu: auto # cuda, rocm, directml, cpu

📸 Example Usage

1. Enter prompt: A cyberpunk cityscape at sunset, neon lights, highly detailed, 8k
2. Negative prompt: blurry, lowres, text, watermark
3. Set: Steps=30, CFG=7, Sampler=Euler a, Size=512x512
4. Hit Generate → Save or upscale results.

🛠️ Troubleshooting

Issue	Solution
Out of VRAM	Reduce resolution or use --medvram flag
No GPU detected	Install CUDA 12.1+ or check python detect_hardware.py
Slow generation	Enable xformers: pip install xformers
Model not loading	Verify .safetensors checksum on Civitai

Logs: Check logs/app.log for errors.

🤝 Contributing

1. Fork the repo
2. Create feature branch: git checkout -b feature/amazing-ui
3. Commit changes: git commit -m 'Add dark mode toggle'
4. Push: git push origin feature/amazing-ui
5. Open a Pull Request

See CONTRIBUTING.md for details.

📄 License

MIT License. See LICENSE for details.

🙌 Support the Project

• ⭐ Star on GitHub
• Share your generations on socials with #NectarGraphix
• Buy me a coffee: ko-fi.com/yourusername

---

Built with ❤️ for local AI creators