<aside> 🗨️

This guide was generated using Gemini from the Comfy Org livestream by Purz and Fill. Thanks to Purz, Fill, and the Comfy community for the content!

Video Link:

https://www.youtube.com/watch?v=-gOhCVU_ogY

</aside>

<aside> <img src="/icons/network_blue.svg" alt="/icons/network_blue.svg" width="40px" />

**Comfy Workflow (**by @Dream Making & @Flipping Sigmas)

Vace22_Image_Ref_T2V.json

</aside>

Back to:

Nathan’s Notes on Open Source AI

Here is a detailed user guide based on the video about using VACE with Wan 2.2 in ComfyUI to generate video, with steps, tips, guidelines, and timestamps as requested.

VACE with Wan 2.2 in ComfyUI: A Detailed User Guide

This guide provides a comprehensive walkthrough of an intermediate to advanced workflow for generating animated videos using the Wan 2.2 VACE model in ComfyUI, as demonstrated in the ComfyUI official stream.

1. Introduction and Workflow Overview

This workflow is designed for users with some familiarity with ComfyUI and focuses on generating video from an input image and video, using a depth map for control. The VACE model provides more accurate control over the final output compared to other methods.

• Workflow Complexity: This is considered an intermediate to advanced workflow due to its many components and settings [08:01].
• Core Concept: The workflow utilizes a depth map generated from an input video and a reference image to create an animated video. This is similar to ControlNet but with the enhanced control of VACE [08:15].
• Inputs and Output: The process takes an input image and an input video to produce a final animated video output [08:33].

2. Step-by-Step Workflow Guide

Here are the key nodes and steps to build and use this ComfyUI workflow:

Step 1: Setting Up Constants [09:53]

• Define Constants: Start by creating integer constants for Width, Height, and Frame Count. For this workflow, the recommended settings are 640x640 pixels and 81 frames, which is the sweet spot for VACE.