Self-hosted ComfyUI install - Ubuntu 24.02

Installing ComfyUI on Ubuntu 24.04 is straightforward but if you want to do it the clean and professional way, this guide shows how to deploy it in /opt/ComfyUI, configure it for GPU acceleration, and run it automatically on system boot with systemd.

Whether you're an artist or an engineer, this setup keeps things modular and system-integrated.

Prerequisites

• Ubuntu 24.04 (fresh or updated)
• An NVIDIA GPU (with at least 4 GB VRAM)
• sudo access

1. Install System Dependencies

sudo apt update
sudo apt install git python3 python3-venv python3-dev build-essential libgl1 ssh systemd-timesyncd -y

2. Install NVIDIA GPU Drivers

sudo ubuntu-drivers list --gpgpu
sudo apt update
sudo ubuntu-drivers install --gpgpu
sudo apt install nvidia-cuda-toolkit
sudo reboot

After reboot, verify with:

nvidia-smi

3. Clone ComfyUI to /opt

sudo git clone https://github.com/comfyanonymous/ComfyUI.git /opt/ComfyUI
sudo chown -R $USER:$USER /opt/ComfyUI

4. Set Up Python Virtual Environment

cd /opt/ComfyUI
python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip wheel
pip install -r requirements.txt

5. (Optional) Install GPU-Accelerated PyTorch

pip install torch torchvision

6. Create Model Folders

sudo mkdir -p /opt/ComfyUI/models/checkpoints
sudo mkdir -p /opt/ComfyUI/models/vae
sudo mkdir -p /opt/ComfyUI/models/clip
sudo mkdir -p /opt/ComfyUI/models/upscale_models
sudo mkdir -p /opt/ComfyUI/models/controlnet
sudo chown -R comfyui:comfyui /opt/ComfyUI/models

7. (Optional) Download Default Models

# Stable Diffusion 1.5 Checkpoint
sudo -u comfyui wget -O /opt/ComfyUI/models/checkpoints/v1-5-pruned-emaonly.safetensors \
https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.safetensors

# VAE for SD1.5
sudo -u comfyui wget -O /opt/ComfyUI/models/vae/vae-ft-mse-840000-ema-pruned.safetensors \
https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.safetensors

# ControlNet Canny
sudo -u comfyui wget -O /opt/ComfyUI/models/controlnet/control_canny-fp16.safetensors \
https://huggingface.co/lllyasviel/ControlNet/resolve/main/models/control_canny-fp16.safetensors

8. (Optional) Install Custom Nodes

sudo -u comfyui git clone https://github.com/ltdrdata/ComfyUI-Manager.git /opt/ComfyUI/custom_nodes/ComfyUI-Manager
sudo -u comfyui git clone https://github.com/Fannovel16/comfyui-reactor-node.git /opt/ComfyUI/custom_nodes/comfyui-reactor-node

9. Create a Systemd Service

Create the user:

sudo groupadd comfyui
sudo useradd -m -g comfyui comfyui
sudo usermod -a -G comfyui $USER
sudo chown -R comfyui:comfyui /opt/ComfyUI
sudo chmod -R 775 /opt/ComfyUI

Create the systemd service file:

sudo nano /etc/systemd/system/comfyui.service

Paste this:

[Unit]
Description=ComfyUI Service
After=network-online.target
Wants=network-online.target

[Service]
User=comfyui
WorkingDirectory=/opt/ComfyUI
ExecStart=/opt/ComfyUI/.venv/bin/python3 main.py --listen 0.0.0.0
Restart=always

[Install]
WantedBy=multi-user.target

Enable and start it:

sudo systemctl daemon-reload
sudo systemctl enable comfyui
sudo systemctl start comfyui

10. Access ComfyUI

If everything started correctly, go to:

http://<your-server-ip>:8188

Integrating ComfyUI into n8n Workflows

One of the most powerful aspects of this setup is how it integrates into my broader automation system using n8n a powerful, open-source workflow automation platform.

I use n8n to trigger image generation jobs in ComfyUI based on external events, such as:

• New RSS feed items from blogs or news
• Webhook calls triggered by my apps or devices
• Timed schedules to generate daily AI art themes
• AI agents creating prompts automatically and passing them to ComfyUI

Workflow Example

A typical workflow looks like this:

1. Trigger: n8n receives a webhook, RSS post, or GPT-generated idea.
2. Prompt Creation: A GPT agent (via OpenAI or Ollama) constructs an image prompt.
3. ComfyUI Call: n8n sends an HTTP POST to a custom ComfyUI API node I created.
4. Image is generated: ComfyUI renders the output and saves it to a known location.
5. Webhook return or upload: The result is sent to Discord, Telegram, or an email.

Why Use ComfyUI + n8n?

This hybrid setup gives me complete flexibility:

• I keep my models local, fast, and private.
• I chain AI tools without coding everything manually.
• I can automate hundreds of images from data streams.

If you're an automation geek or artist, this combo is a creative superpower.