---
title: "Discussion on \"Wan 2.2 + ComfyUI on 8GB VRAM: Fast Setup & Image-to-Video Test\" | Hashnode"
description: "Discussion on \"Wan 2.2 + ComfyUI on 8GB VRAM: Fast Setup & Image-to-Video Test\". 🚀 Wan 2.2 is a newly released image-to-video AI model that can generate cinematic video from a single still image. The best part? You can run it in ComfyUI with just 8GB of VRAM.\nHere’s a quick setup overview based on my full video tutorial.\n\n🔧 Wha..."
url: "https://hashnode.com/posts/wan-22-comfyui-on-8gb-vram-fast-setup-and-image-to-video-test/688ed38b41a25dba80191d6a"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Wan 2.2 + ComfyUI on 8GB VRAM: Fast Setup & Image-to-Video Test](https://aitechtutorials.hashnode.dev/wan-22-comfyui-on-8gb-vram-fast-setup-and-image-to-video-test?utm_source=hashnode&utm_medium=feed)

🚀 Wan 2.2 is a newly released image-to-video AI model that can generate cinematic video from a single still image. The best part? You can run it in ComfyUI with just 8GB of VRAM. Here’s a quick setup overview based on my full video tutorial. 🔧 Wha...

Aug 3, 2025 · 2 min read · aitechtutorials.hashnode.dev  
Read the article

-   [#wan22](https://hashnode.com/tag/wan22)
-   [#comfyui](https://hashnode.com/tag/comfyui)
-   [#gemini-cli](https://hashnode.com/tag/gemini-cli)
-   [#ai](https://hashnode.com/tag/ai)
-   [#ai-tools](https://hashnode.com/tag/ai-tools)

## Written by

[Jean](https://hashnode.com/@aitechtutorials)

@aitechtutorials

  
[View profile](https://hashnode.com/@aitechtutorials)

ai tech tutorials

AI enthusiast and digital creator exploring the latest in ComfyUI, Qwen Lightning, and AI image generation workflows. Sharing step-by-step tutorials, tips, and insights to help developers and creators get high-quality AI results efficiently, even on mid-tier setups.

## Responses 1

Markdown supported · ⌘↵ to post

[douzatan](https://hashnode.com/@dzt21)

[Oct 31](https://hashnode.com/posts/wan-22-comfyui-on-8gb-vram-fast-setup-and-image-to-video-test/688ed38b41a25dba80191d6a/comment/6ae6248244ed1206f2e1a5a5)

The folder map makes the setup easier to follow. For the 8GB example, could you also publish the output resolution, frame count, render time and exact workflow revision used for the wolf test? Those details would help readers distinguish a working installation from a usable production setup. If a shot is later used in a [Leadde](https://leadde.ai/features/ai-video-creator) explainer, I would first review a small sample against the narration, rather than assume a successful local render is ready for the finished lesson.
