Screen Diffusion

🌌 Screen Diffusion

A real-time AI transformation tool for artists, dreamers, and creators.

🕊️ What Is Screen Diffusion?

Screen Diffusion is a live image-to-image AI renderer that transforms your computer screen into living art — in real time.
Whatever you display — a game, a 3D scene, a photo, a design , a webcam— can be instantly reimagined through SD-Turbo model to reveal new styles, moods, or worlds.

It’s built for artists, developers, streamers, and visionaries who long to see creation through new lenses — to behold the ordinary transformed into the extraordinary.

✨ Features

• 🎞️ Real-Time Transformation — Capture any window or screen region and watch it evolve live through AI.

• 🧠 Local AI Models — Uses your GPU to run Stable Diffusion variants in real time.

• 🎛️ Adjustable Prompts & Settings — Change prompts, styles, and diffusion steps dynamically.

• ⚙️ Optimized for RTX GPUs — Designed for speed and efficiency on Windows 11 with CUDA acceleration.

• 💻 1 Click setup — Designed to make your setup quick and easy.
  

🎁 Free to Use — Support/Tip If You Wish

This project is completely free for everyone to use and explore.
If Screen Diffusion inspires you, you can “pay what you want” or Tip — your support helps sustain development and future features:

• Higher resolution img2img.
• TensorRT support.
• Other model support including (SDXL-Turbo, SDXS)
• Controlnet and LoRA support
• More optimization for higher frame-rate.
• Custom GUI themes and auto-prompt presets.

source code: https://github.com/rudyaa-sd/ScreenDiffusion

🙏 Every contribution, message, and prayer helps this project grow.

🛠️ Prerequisites

• Windows 11 / Windows 10

• NVIDIA GPU with 8GB+ VRAM (RTX 3080 or higher recommended)

• CUDA Toolkit 12.1 installed (Restart required after install)

• Python 3.11.8 (download) https://www.python.org/downloads/release/python-3118/
  ( Click on "Ad Python 3.11.8 to PATH" when installing)

• Download SD-Turbo Model stored locally - (download link) https://huggingface.co/stabilityai/sd-turbo/tree/main
  (Point the app to the folder that contains unet / vae/ textencoder...etc. This folder has to have a file called "model_index.json"

• 

📦 Installation

1. Download the ZIP file and extract it anywhere on your computer.
2. Open "Screen Diffusion.exe"
3. Click on "Enable GPU" button and wait for everything to download.
4. Once download has finished the app is restarted automatically. 
5. Point to your model folder path.
   
6. Click Start.
7. Choose your capture region, adjust prompts, and enjoy the transformation.

💡 Tip: Keep your model (e.g., SD-Turbo) next to the .exe for easy access.

🔧Troubleshooting

• If for some reason Torch isn't being recognized by the app after clicking EnableGPU, cd into the  "_internal"  folder directory and download Torch with code (copy and paste it, it's one line) : pip install --target . --upgrade --no-deps torch==2.1.0+cu118 torchvision==0.16.0+cu118 --index-url https://download.pytorch.org/whl/cu118
• If you get an error about the model, try using the fp16 in the model folder that you downloaded, but rename them by removing the fp16 extension for all three unet/vae/text_encoder.
• Comment down in the comment section on the issue you're having, and I'll try to help you find a solution.

⚠️ Antivirus Notice

Some antivirus software may falsely flag the EXE — this is a common issue with PyInstaller-based apps.
ScreenDiffusion is safe, open source, and completely local.

Please feel free to build your own EXE file from the source code if you want:
https://github.com/rudyaa-sd/ScreenDiffusion

🌄 Closing Thought

May Screen Diffusion inspire you to see creation not just as pixels, but as a living canvas.