From download to your first image.

1. Download the Windows package

   From the latest release on GitHub, download NomadStudio-…-windows-x64.zip. It already contains everything Nomad Studio needs, Python included — nothing else to install.

2. Extract it

   Right-click the ZIP › Extract All… and choose a folder of your own, such as Documents. Avoid Program Files and folders synced by OneDrive.

3. Double-click NomadStudio.bat

   The first time, Windows may ask whether to run a file downloaded from the internet: confirm with Run (or More info › Run anyway). A small console window opens minimised and your browser opens Nomad Studio at http://localhost:5123.

Use the official ComfyUI Portable: a folder with its own Python that you run with a double-click.

1. Download the build for your graphics card

   From the ComfyUI releases page, download the portable .7z for your card: NVIDIA (most users), AMD or Intel. It is about 2 GB.

2. Extract it with 7-Zip

   Install 7-Zip if you don't have it, then right-click the file › 7-Zip › Extract here. Put it on a drive with room to spare: image models take 10–25 GB each.

3. Start it

   Open the extracted folder and double-click run_nvidia_gpu.bat (or run_amd_gpu.bat / run_intel_gpu.bat). When the console shows To see the GUI go to: http://127.0.0.1:8188 it is ready. Leave that window open while you work.

Only the flows that write or rewrite prompts, chat, or describe images use it. Everything else works without it.

1. Install LM Studio

   Download the Windows installer from lmstudio.ai and run it.

2. Start its local server

   Open LM Studio, go to the Developer tab and switch the server on. It listens on port 1234. Or, from a terminal:

   lms server start

3. Let Nomad Studio pick the models

   The setup wizard recommends language models measured with Nomad Studio and downloads the one you choose through LM Studio.

1. Download the macOS package

   From the latest release on GitHub, download NomadStudio-…-macos-arm64.tar.gz. Python and everything else is inside.

2. Extract it

   Double-click the file in Finder. You get a folder with NomadStudio.command and app. Move that folder wherever you like, for example your home folder.

3. Allow it the first time

   Double-click NomadStudio.command. macOS says it can't verify the developer, because the app is distributed outside the App Store: click Done, open System Settings › Privacy & Security, scroll down and click Open Anyway next to NomadStudio.command, then confirm. You only do this once.

   Prefer Terminal? This does the same in one line (adjust the path to where you put the folder):

   xattr -dr com.apple.quarantine ~/NomadStudio

4. Start it

   Double-click NomadStudio.command. A Terminal window opens and your browser opens Nomad Studio at http://localhost:5123. Closing that Terminal window stops Nomad Studio.

On a Mac, use ComfyUI Desktop: the official app for Apple Silicon. It runs on the Mac's GPU through Metal.

1. Download ComfyUI Desktop

   Get the macOS (Apple Silicon) version from comfy.org/download, open the .dmg and drag ComfyUI into Applications.

2. Run its first-time setup

   Open ComfyUI and follow its setup: choose MPS as the device and a folder with plenty of free space — image models take 10–25 GB each. When it shows its workspace, it is ready.

3. Leave it running

   Nomad Studio finds it on its own: the setup wizard looks for ComfyUI on this Mac on the usual ports, including the one ComfyUI Desktop uses.

Only the flows that write or rewrite prompts, chat, or describe images use it. Everything else works without it.

1. Install LM Studio

   Download the Mac version from lmstudio.ai (Apple Silicon, macOS 14 or later), open the .dmg and drag it into Applications.

2. Start its local server

   Open LM Studio, go to the Developer tab and switch the server on. It listens on port 1234. Or, from a terminal:

   lms server start

3. Let Nomad Studio pick the models

   The setup wizard recommends language models measured with Nomad Studio and downloads the one you choose through LM Studio.

1. Download the Linux package

   From the latest release on GitHub, download NomadStudio-…-linux-x64.tar.gz. Python and every library are inside.

2. Extract it into its own folder

   The archive has no top folder of its own, so create one:

   mkdir -p ~/NomadStudio
   tar -xzf ~/Downloads/NomadStudio-*-linux-x64.tar.gz -C ~/NomadStudio

3. Start it

   Your browser opens Nomad Studio at http://localhost:5123. Press Ctrl+C in the terminal to stop it.

   cd ~/NomadStudio
   ./NomadStudio.sh

Two official ways. Choose one:

1. ComfyUI Desktop

   For the .deb: sudo apt install ./ComfyUI*.deb. For the AppImage: make it executable and run it. Then follow its first-time setup and pick a folder with plenty of free space.

   chmod +x ComfyUI*.AppImage
   ./ComfyUI*.AppImage

2. Or the manual install

   Clone it, create its Python environment and install PyTorch for your card — the exact PyTorch line for NVIDIA, AMD or Intel is in the official manual-install guide. Then:

   git clone https://github.com/Comfy-Org/ComfyUI.git
   cd ComfyUI
   python3 -m venv venv
   source venv/bin/activate
   # install PyTorch for your GPU here (see the official guide)
   pip install -r requirements.txt
   python main.py

3. Check it answers

   The manual install serves on http://127.0.0.1:8188; Desktop uses its own port. Nomad Studio's setup wizard looks for both on its own.

Only the flows that write or rewrite prompts, chat, or describe images use it. Everything else works without it.

1. Install LM Studio

   Download the AppImage from lmstudio.ai, make it executable and run it — or use LM Studio's official install script:

   curl -fsSL https://lmstudio.ai/install.sh | bash

2. Start its local server

   Open LM Studio, go to the Developer tab and switch the server on. It listens on port 1234. Or, from a terminal:

   lms server start

3. Let Nomad Studio pick the models

   The setup wizard recommends language models measured with Nomad Studio and downloads the one you choose through LM Studio.