What you need before you start

Stable Diffusion is an image generation model that runs on your own machine rather than through a web service. Before installation, you need a graphics card (GPU) with enough memory, a compatible operating system, and about 10 to 15 gigabytes of free disk space. The process differs depending on whether you use Windows, Mac, or Linux, and whether you have an Nvidia, AMD, or Apple Silicon GPU.

You do not need to be a programmer, but you will need to use the command line — the text-based interface where you type commands rather than clicking buttons. If you have never opened a terminal or command prompt before, the process will feel unfamiliar but is manageable by following exact steps in order.

The easiest route for most people is to use a graphical installer that handles the technical setup for you. The most common options are Automatic1111 (Windows and Linux), Invoke AI (all platforms), or ComfyUI (all platforms). Each one downloads Stable Diffusion and its dependencies, then gives you a visual interface to generate images.

Key Takeaways

  • Stable Diffusion requires a graphics card with at least 4 GB of memory on Nvidia cards, or 8 GB on AMD or Apple Silicon.
  • The easiest installation path uses a graphical installer like Automatic1111 or Invoke AI rather than manual command-line setup.
  • Installation takes 30 minutes to an hour and downloads 5 to 10 GB of model files the first time you run it.
  • You will need Python installed on your computer first, which is a free programming language that most installers can set up for you automatically.
  • After installation, you generate images by typing text descriptions into the interface on your local computer — no internet upload required.

Check your graphics card and system requirements

Stable Diffusion runs much faster on a dedicated graphics card than on a CPU alone. On Nvidia cards, you need at least 4 GB of VRAM (video memory) for basic use, though 6 GB or more is more comfortable. On AMD cards, you need at least 8 GB. On Apple Silicon Macs (M1, M2, M3 chips), the unified memory works differently and 8 GB of system RAM is the practical minimum.

To check what GPU you have on Windows, right-click the desktop and look for "Nvidia Control Panel" or "AMD Radeon Settings" in the menu. On Mac, click the Apple menu, select "About This Mac," and look under "Chip" for M1, M2, or M3. On Linux, open a terminal and type lspci | grep -i vga to see your graphics hardware.

If you have an older or integrated graphics card with less than 4 GB of memory, Stable Diffusion will still run but will be very slow — generating a single image might take several minutes instead of seconds. You can still proceed with installation, but expect longer wait times.

Install Python first

Most Stable Diffusion installers need Python, a programming language, to run. You likely do not have it installed yet. Download Python 3.10 or 3.11 from python.org — choose the version labeled for your operating system (Windows, macOS, or Linux).

On Windows, run the installer and check the box that says "Add Python to PATH" before clicking Install. This step lets your computer find Python when other programs need it. On Mac, the installer will walk you through the process. On Linux, use your package manager: on Ubuntu or Debian, type sudo apt install python3.10 in a terminal.

After installation, open a command prompt (Windows), terminal (Mac or Linux), and type python --version to confirm it installed. You should see something like "Python 3.10.12" or similar. If you see an error, Python did not install correctly — try the installer again.

Download and run Automatic1111 (Windows and Linux)

Automatic1111 is the most widely used installer for Windows and Linux. Go to the GitHub repository at github.com/AUTOMATIC1111/stable-diffusion-webui and click the green "Code" button, then "Download ZIP." Extract the folder to a location you will remember, like your Documents or Desktop folder.

Open a command prompt or terminal, navigate to the extracted folder, and run the startup script. On Windows, double-click webui-user.bat. On Linux, open a terminal in that folder and type bash webui.sh. The script will download Python dependencies and the Stable Diffusion model files — this takes 10 to 20 minutes depending on your internet speed and is only needed the first time.

When the download finishes, you will see a message with a web address like http://127.0.0.1:7860. Copy that address into your web browser. You should see a text box where you can type image descriptions. Type something like "a cat wearing sunglasses" and click Generate. The first image takes longer because the model is loading into memory, but subsequent images are faster.

Install on Mac with Invoke AI

Mac users should use Invoke AI instead of Automatic1111, because Invoke AI is optimized for Apple Silicon. Download the installer from invoke-ai.github.io/installation and choose the Mac version. The installer is a graphical application — download it, open it, and follow the on-screen prompts.

Invoke AI will ask where to install the application and whether to download the Stable Diffusion model files. Choose yes to download the model files, which takes 10 to 15 minutes. After installation, open the Invoke AI application from your Applications folder. It will open in your web browser automatically.

The interface is similar to Automatic1111 — type a description in the text box and click Generate. On Apple Silicon Macs, image generation is reasonably fast because the unified memory architecture works well with the model. If you have an Intel Mac, Stable Diffusion will run but more slowly.

Use ComfyUI for more control

ComfyUI is a node-based interface that gives you more control over how images are generated. It is more complex than Automatic1111 or Invoke AI but lets you chain together different steps — like upscaling, inpainting, or using multiple models in sequence. Download it from github.com/comfyanonymous/ComfyUI the same way you would Automatic1111.

Extract the folder, open a terminal or command prompt in that location, and run the startup script. On Windows, double-click run_nvidia_gpu.bat (for Nvidia cards) or run_amd_gpu.bat (for AMD). On Mac or Linux, type python main.py. The interface opens in your browser and shows a visual workflow instead of a simple text box.

ComfyUI has a steeper learning curve because you need to understand what each node does, but it is powerful for advanced use cases. If you are new to Stable Diffusion, start with Automatic1111 or Invoke AI and move to ComfyUI later if you want more flexibility.

Troubleshooting common installation problems

If the installer says "Python not found," you either did not install Python or did not add it to PATH. Reinstall Python and make sure to check "Add Python to PATH" on Windows. On Mac or Linux, open a terminal and type which python3 — if it returns a path, Python is installed correctly.

If the download stops or fails partway through, your internet connection may have dropped. Delete the partially downloaded model files and run the installer again. The installer will resume the download. Model files are usually stored in a folder called models inside the installation directory.

If image generation is very slow (more than a minute per image), your GPU may not be detected. On Windows with Nvidia cards, make sure you have the latest Nvidia drivers installed from nvidia.com/Download/driverDetails.aspx. On AMD cards, install the latest AMD drivers. On Mac, update to the latest macOS version.

If you see an error about "out of memory," your graphics card does not have enough VRAM. Try generating smaller images (512x512 instead of 768x768) or using fewer steps (20 instead of 50). Both settings are adjustable in the interface.

Frequently Asked Questions

Do I need an internet connection to generate images after installation?

No. After the initial download of the model files, Stable Diffusion runs entirely on your computer. You do not upload images or text to a server. The web interface that opens in your browser is running locally on your machine, not on the internet.

Can I use Stable Diffusion on a laptop?

Yes, if your laptop has a dedicated graphics card with at least 4 GB of VRAM. Laptops with integrated graphics (Intel UHD, AMD Radeon) will work but generate images much more slowly. Apple Silicon MacBook Airs and Pros work well because the unified memory is efficient.

What is the difference between Stable Diffusion 1.5 and SDXL?

SDXL is a newer, larger model that produces higher-quality images but requires more GPU memory — at least 8 GB for comfortable use. Stable Diffusion 1.5 is smaller and faster on limited hardware. Most installers let you choose which model to download when you first run them.

Can I use multiple models at the same time?

You can install multiple models, but you can only use one at a time without running out of memory. The installer downloads one model by default, but you can add others later by placing model files in the models folder. Switching between them in the interface is instant.

Is it legal to use Stable Diffusion?

Stable Diffusion itself is open-source and legal to download and run. The images you generate are yours to use. However, the model was trained on images from the internet, and some of those images may have copyright or licensing restrictions. Using generated images commercially or for publication may have legal implications depending on your jurisdiction and use case.