Install Wan 2.2 in ComfyUI | Full AI Image-to-Video Workflow (Tested on 8GB VRAM)

This tutorial shows how to install and run the new Wan 2.2 image-to-video AI model inside ComfyUI, using a system with just 8GB of VRAM.

You'll learn how to:
• Download and set up the GGUF 5B model
• Install the required LoRA, VAE, and UMT5 encoder
• Organize all files in the correct ComfyUI folders
• Import a Hugging Face workflow to generate motion from a single image
• Fix the sageattention error automatically using Gemini CLI

🎬 Watch as I test the full pipeline using a wolf image and a storm prompt, all generated from a still frame.

No high-end GPU required. This is a real-world setup for creators, devs, and AI enthusiasts using open-source tools.

📎 Full YouTube version here (if needed):
https://youtu.be/7hUO6KhUsvQ