• Desktop App → - Download and run locally (recommended)
• Landing Page → - Learn more about Z-Image

• Web App → - Deploy your own web version
• Original Z-Image Docs - Model information

Native desktop application for macOS, Windows, and Linux.

• ⚡ Blazing fast - Generate 1024x1024 images in <1 second
• 🔒 Completely private - All processing local, no cloud
• 🍎 Mac optimized - Native Apple Silicon support with Metal
• 🎨 Professional quality - 6B parameter Z-Image-Turbo model
• 📦 Self-contained - Tauri desktop app with embedded Python backend

cd desktop-app
npm install
npm run tauri:dev

Full Desktop App Documentation →

Browser-based interface with cloud deployment options.

• 🎨 Neo-brutalist design
• ⚡ Optimized for speed
• 📱 Responsive (mobile, tablet, desktop)
• 🚀 Easy deployment (Netlify, Vercel, etc.)

Backend:

cd web-app/backend
pip install -r requirements.txt
python main.py

Frontend:

cd web-app/frontend
npm install
npm run dev

Full Web App Documentation →

Deployment Guide →

High-converting marketing page for the desktop app.

Bold neo-brutalist design that emphasizes speed, privacy, and zero cost.

cd landing-page
npm install
npm run dev

1. Push to GitHub
2. Connect to Netlify
3. Configure: Base directory: landing-page, Build: npm run build, Publish: landing-page/dist

Landing Page Documentation →

• ✅ Maximum speed and privacy
• ✅ No server setup/hosting
• ✅ Native app experience
• ✅ Offline usage after model download

• ✅ Share with others (deploy to cloud)
• ✅ No local installation required
• ✅ Access from any device
• ✅ Custom branding/styling

• ✅ Integrate into your own code
• ✅ Batch processing
• ✅ Custom workflows
• ✅ Maximum control

Z-Image is a powerful and highly efficient image generation model with 6B parameters.

• 🚀 Z-Image-Turbo – A distilled version that matches or exceeds leading competitors with only 8 NFEs (Number of Function Evaluations). It offers ⚡️sub-second inference latency⚡️ on enterprise-grade H800 GPUs and fits comfortably within 16G VRAM consumer devices.

• 🧱 Z-Image-Base – The non-distilled foundation model (coming soon)

• ✍️ Z-Image-Edit – Fine-tuned for image editing tasks (coming soon)

Model	Hugging Face	ModelScope
Z-Image-Turbo

import torch
from diffusers import ZImagePipeline

pipe = ZImagePipeline.from_pretrained(
    "Tongyi-MAI/Z-Image-Turbo",
    torch_dtype=torch.bfloat16,
)
pipe.to("cuda")  # or "mps" for Mac, "cpu" for CPU

image = pipe(
    prompt="Your prompt here",
    height=1024,
    width=1024,
    num_inference_steps=9,
    guidance_scale=0.0,
).images[0]

image.save("output.png")

• Z-Image:
• Decoupled-DMD:
• DMDR:

• Cache-DiT - Inference acceleration with DBCache
• stable-diffusion.cpp - Pure C++ inference engine (4GB VRAM!)

• Tauri - Lightweight native app framework (Rust)
• React - UI framework
• Python - ML inference backend
• PyTorch - ML framework
• FastAPI - Backend API

• React - Frontend framework
• Vite - Build tool
• FastAPI - Backend API
• Diffusers - Hugging Face pipelines

• React - UI framework
• Vite - Build tool
• Netlify - Deployment

• macOS 10.13+ (Windows/Linux support for desktop app coming)
• 16GB+ RAM (unified memory on Apple Silicon)
• Apple Silicon (M1/M2/M3) or NVIDIA GPU
• 20GB free disk space
• Python 3.10+

• 8GB RAM (slower, ~5-10s per image)
• CPU only (no GPU required but much slower)

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

This project follows the Z-Image repository license.

• Z-Image-Turbo by Tongyi-MAI
• Diffusers by Hugging Face
• Tauri for the desktop framework
• FastAPI by Sebastián Ramírez
• React by Meta
• All contributors to this project

• Desktop App Issues
• Web App Deployment Help
• GitHub Issues
• Original Z-Image Site