To install ComfyUI on Linux, create a Python environment, clone the repository, install PyTorch for your GPU, install requirements.txt, and run python main.py. These are the steps from the official manual install docs.
Steps
conda create -n comfyenv
conda activate comfyenv
git clone https://github.com/Comfy-Org/ComfyUI.git
cd ComfyUI
Install PyTorch. For NVIDIA GPUs the README gives:
pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu130
Then the dependencies, and run it:
pip install -r requirements.txt
python main.py
ComfyUI is now at http://127.0.0.1:8188. Stop it with Ctrl + C in the terminal.
Notes
- The README says Python 3.13 is very well supported, 3.14 works but some custom nodes may have issues, and a newer PyTorch is strongly recommended over 2.7.
- AMD GPUs use a different PyTorch index (ROCm). See ComfyUI on AMD with ROCm.
- If you see "Torch not compiled with CUDA enabled", see the fix for that error.
- There is also a CLI route:
pip install comfy-clithencomfy install.
Once it runs, the next job is nodes: the ComfyUI custom node install guide has the packs most workflows need. If you would rather not set the machine up yourself, the ComfyUI template on Aquanode provisions the box, installs ComfyUI and starts it when you deploy it; once the deployment is Ready, the Overview tab shows the ComfyUI URL.
Sources
- ComfyUI docs: Manual install. docs.comfy.org
- ComfyUI README: Installing. github.com