Setting up your system for StreamDiffusion TD For Beginners

This video goes through the steps to prepare your computer to run the StreamDiffusionTD.tox created by @dotsimulate , a generative AI tool capable of real-time output of text2img and img2img content.



1) Can your computer run StreamDiffusion TD?

- Must be Windows PC with a CUDA-compatible NVIDIA GPU

- Check your GPU (

- Is that GPU CUDA compatible? Check this list: (


2) How to install the necessary components to make this work?

When installing Python, CUDA, and Git, if you see an option to “Add to system PATH” make sure you check it yes.

a) Python 3.10.9 (

b) CUDA Toolkit (

c) Visual Studio 2022 Community (need “Desktop environment with C++”) (

d) NDI SDK (

e) Git desktop client (

f) TouchDesigner 2023 version ( …

g) Have updated NVIDIA driver? (


3) Are Python, CUDA, and Git in your PATH environment variable?

How to add Python to PATH

Read the section “Manually Add git to the PATH on Windows”




Once you’ve done that, it should work! Watch dotsimulate’s tutorial (the developer of the StreamDiffusion TD tool) on how to get started (   • Real-time diffusion in TouchDesigner ...  )

Github repo of StreamDiffusion researchers (heavy CS content)




00:00 Overview

01:50 Check GPU

02:18 Check CUDA compatibility

03:23 Download Python

05:06 Check add to system PATH

05:32 Download CUDA

06:13 Download Visual Studio

07:05 Download NDI SDK

08:02 Download Git

09:10 How to add to system PATH environment variable manually

12:12 TD 2023 version required

12:34 Update NVIDIA driver

13:24 You're ready to go


Experience level