What is the best Stable Diffusion model for Mac?

Stable Diffusion 1.5 checkpoint models work well on Mac. DreamShaper from Civitai is a popular choice. Download the .safetensors file and place it in the stable-diffusion-webui/models/stable-diffusion folder.

How do I fix cross attention layer errors on Mac?

Run the WebUI with the --no-half flag: ./webui.sh --no-half. If issues persist, add --loglevel DEBUG. For memory errors, prefix the command with PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.7.

How long does the first Stable Diffusion startup take on Mac?

The first run takes 5–15 minutes on Mac as it downloads and sets up the Python environment and required packages. Subsequent launches are much faster since everything is already installed.

How to Install Stable Diffusion on Mac (M1 & M2)

Introduction

What this guide covers: Installing Stable Diffusion WebUI (AUTOMATIC1111) on an Apple Silicon Mac (M1 or M2) — from installing Homebrew and dependencies to generating your first AI image.

In this guide, you will learn how to install Stable Diffusion on a Mac with Apple Silicon (M1 or M2). Stable Diffusion is a powerful tool for generating AI art, and with the right setup, it runs smoothly on your Mac using Apple's Metal GPU acceleration.

Follow these steps in order and you will have a working Stable Diffusion installation ready to generate images by the end of this tutorial.

System Requirements

Before you begin, make sure you have the following:

1A Mac with an M1 or M2 chip (Apple Silicon).

2Terminal access — built into macOS, found in Applications → Utilities → Terminal.

3A stable internet connection to download Homebrew, packages, and model files.

💡 Tip: Intel Macs can also run Stable Diffusion using the AUTOMATIC1111 WebUI, but performance is significantly better on Apple Silicon thanks to Metal GPU acceleration.

Step 1: Install Homebrew

Homebrew is a package manager for macOS that makes it easy to install the tools Stable Diffusion needs. If you do not have it yet, install it now.

Copy the Homebrew install command

Copy the command below — you will paste it into Terminal in the next step.

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

Open Terminal and paste the command

Open Terminal on your Mac (Applications → Utilities → Terminal), paste the command, and press Enter.

Enter your password and wait

You will be prompted for your Mac password. Enter it and let the installation process complete — it takes a few minutes.

macOS Terminal showing Homebrew installation running successfully — Homebrew installing in the macOS Terminal

Once Homebrew is installed, you will see instructions asking you to adjust your PATH. Copy the last three lines from your terminal output and paste them back into Terminal. They will look something like this:

==> Next steps:
- Run these commands in your terminal to add Homebrew to your PATH:
    echo >> /Users/YOUR_NAME/.zprofile
    echo 'eval "$(/opt/homebrew/bin/brew shellenv)"' >> /Users/YOUR_NAME/.zprofile
    eval "$(/opt/homebrew/bin/brew shellenv)"

macOS Terminal showing Homebrew PATH setup instructions after installation — Copy the three PATH commands from your terminal and paste them back in

Step 2: Install Required Packages

With Homebrew installed, run the following command to install all the packages that Stable Diffusion needs to run on your Mac — including Python 3.10, Git, cmake, protobuf, Rust, and wget.

brew install cmake protobuf rust python@3.10 git wget

macOS Terminal showing brew install command running and installing all required packages — Homebrew installing all required packages for Stable Diffusion

💡 Tip: This command installs everything in one go. It may take several minutes depending on your internet speed.

Step 3: Clone the Stable Diffusion Repository

Create a folder for Stable Diffusion

Create a folder where you want to keep the installation — for example, a folder called AI_Art in your home directory.

Navigate to that folder in Terminal

Use the cd command to navigate into the folder you just created before running the clone command.

Clone the AUTOMATIC1111 repository

Run the command below to download the Stable Diffusion WebUI into your chosen folder.

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui

macOS Terminal showing git clone command downloading the AUTOMATIC1111 stable-diffusion-webui repository — Git cloning the Stable Diffusion WebUI repository

Step 4: Download and Add Model Files

Stable Diffusion needs a checkpoint model file to generate images. Download one from a trusted source and place it in the correct folder.

Download a Stable Diffusion 1.5 checkpoint

Download a .safetensors or .ckpt model file from Civitai or Hugging Face. DreamShaper is a popular beginner-friendly choice.

DreamShaper on Civitai Stable Diffusion 1.5 on Hugging Face

Move the model file to the correct folder

After downloading, move the .safetensors or .ckpt file into this folder inside your cloned repository:

stable-diffusion-webui > models > stable-diffusion

Step 5: Start the WebUI for Stable Diffusion

Now that everything is set up, you can launch the Stable Diffusion web interface. Navigate to the cloned folder first, then start the WebUI.

Navigate to the Stable Diffusion folder

In Terminal, change directory into your cloned repository folder:

cd stable-diffusion-webui

Start the WebUI

Run the launch script:

./webui.sh

⚠️ Note: The first launch takes 5–15 minutes. It sets up the Python virtual environment and downloads required packages automatically. Subsequent launches are much faster.

macOS Terminal showing Stable Diffusion WebUI starting up and downloading required Python packages on first launch — First launch may take several minutes — let it complete without interrupting

Step 6: Generate Your First Image

Once the WebUI opens in your browser, you are ready to generate your first AI image:

Select the checkpoint model you downloaded in the model dropdown at the top.
Type a prompt in the text box — for example: "a dog wearing a hat".
Click Generate and wait for your image to appear.

Stable Diffusion WebUI open in a Mac browser showing the text-to-image interface with prompt field and generated image — The Stable Diffusion WebUI running in your Mac browser — ready to generate images

Fixing Common Mac Errors

If you encounter a cross attention layer error or a blank output on M1/M2, restart the WebUI with the --no-half flag:

./webui.sh --no-half

If the error persists, add debug logging:

./webui.sh --no-half --loglevel DEBUG

If you hit memory errors (MPS out of memory), limit the memory watermark:

PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.7 ./webui.sh --no-half --loglevel DEBUG

💡 Tip: These flags are only needed during troubleshooting. Most users on M1/M2 can run ./webui.sh without any flags once the initial setup works correctly.

Updating Stable Diffusion

To keep your installation up to date, pull the latest changes from GitHub. Navigate to your Stable Diffusion folder in Terminal and run:

cd stable-diffusion-webui

git pull

This fetches the latest commits from the AUTOMATIC1111 repository and updates your local installation to the most recent version of the WebUI.

Frequently Asked Questions

Yes. Stable Diffusion runs on M1 and M2 Macs using the AUTOMATIC1111 WebUI, which uses Apple's Metal Performance Shaders (MPS) for GPU acceleration. You need Homebrew, Python 3.10, Git, and a downloaded checkpoint model to get started.

Stable Diffusion 1.5 checkpoint models work well on Mac. DreamShaper from Civitai is a popular beginner-friendly choice with realistic outputs. Download the .safetensors file and place it in the stable-diffusion-webui/models/stable-diffusion folder.

Run the WebUI with the --no-half flag: ./webui.sh --no-half. If issues persist, add --loglevel DEBUG for more information. For memory-related crashes, prefix the command with PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.7 to limit MPS memory usage.

The first run takes 5–15 minutes on Mac as it sets up the Python virtual environment and downloads required packages. Subsequent launches are much faster — typically under a minute — since everything is already installed locally.

Navigate to the stable-diffusion-webui folder in Terminal and run: git pull. This fetches the latest updates from the AUTOMATIC1111 GitHub repository and keeps your local installation current.

Conclusion

Installing Stable Diffusion on a Mac with Apple Silicon is straightforward once you have the right tools. Homebrew handles the dependencies, Git clones the WebUI, and the --no-half flag fixes most M1/M2-specific issues.

Now that you have a working setup, explore different checkpoint models on Civitai, experiment with prompts, and try LoRA models to control your image style. The AUTOMATIC1111 WebUI has dozens of extensions to extend what Stable Diffusion can do.

Want more speed and flexibility? Try installing ComfyUI on Mac — a node-based interface with more workflow control and faster generation for many models.