Complete ComfyUI + Wan 2.2 Setup Guide for RunPod

This guide will get you up and running with ComfyUI and the powerful Wan 2.2 text-to-video model on RunPod in just a few minutes.

join to runpod first: https://runpod.io?ref=94hstegb

video: https://youtu.be/vnrUJKRogkM

What You're Getting

One-Click Setup Script

Just copy and paste this entire code block into your RunPod terminal. That's it!

# Navigate to workspace
cd /workspace

# Clone ComfyUI
git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUI

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate

# Install ComfyUI requirements
pip install -r requirements.txt

# Install ComfyUI Manager
cd custom_nodes
git clone https://github.com/ltdrdata/ComfyUI-Manager.git
cd ComfyUI-Manager
pip install -r requirements.txt
cd ../..

# Start ComfyUI in background (it will keep running even if you close terminal)
nohup python main.py --listen 0.0.0.0 --port 8188 > app.log 2>&1 &

# Show logs (press Ctrl+C to stop viewing logs, ComfyUI will continue running)
tail -f app.log

# Download Wan 2.2 LoRA files (these enable 4-step fast generation)
cd models/loras
wget -O wan2.2_t2v_lightx2v_4steps_lora_v1.1_high_noise.safetensors "https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/loras/wan2.2_t2v_lightx2v_4steps_lora_v1.1_high_noise.safetensors"
wget -O wan2.2_t2v_lightx2v_4steps_lora_v1.1_low_noise.safetensors "https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/loras/wan2.2_t2v_lightx2v_4steps_lora_v1.1_low_noise.safetensors"
cd ../..

# IMPORTANT: You still need to manually download these large model files:
# 
# Download to models/text_encoders/:
# - umt5_xxl_fp8_e4m3fn_scaled.safetensors
# 
# Download to models/vae/:
# - wan_2.1_vae.safetensors  
# 
# Download to models/diffusion_models/:
# - wan2.2_t2v_low_noise_14B_fp8_scaled.safetensors
# - wan2.2_t2v_high_noise_14B_fp8_scaled.safetensors
#

After Running the Script

1. Access ComfyUI: Open your browser and go to your-pod-url:8188
2. Download remaining models: Get the big model files from Hugging Face (links in the script)
3. Start creating: Type in your text prompt and generate

Troubleshooting

If something goes wrong, check the logs:

cd /workspace/ComfyUI
tail -f app.log

Need to restart ComfyUI? Kill it and start again:

pkill -f "python main.py"
nohup python main.py --listen 0.0.0.0 --port 8188 > app.log 2>&1 &