RunningHub workflow：

https://www.runninghub.ai/post/2065361012815122434/?inviteCode=rh-v1497

Sign up through my link to receive 1,000 credits, then claim 100 credits with each daily login.

YouTube tutorial：

🌟 Overview

Say goodbye to massive workflows packed with masks, pose detection, alignment nodes, and separate tools.

Powered by Wan 2.2, this workflow brings ByteDance’s unified Bernini framework into one all-in-one AI video editor. Edit videos with text, replace characters, transfer motion, restyle scenes with reference images, insert new content, and generate coherent videos from multiple images—all inside one workflow.

🚀 Core Features

• V2V Video Editing

Change weather, remove objects, replace backgrounds, or modify actions with a single prompt.

• Reference-Guided Video Editing

Use reference images to restyle scenes, replace characters, and transfer motion without masks or pose detection.

• Content Insertion

Place images or video clips onto billboards, screens, and other tracked surfaces.

• Multi-Image Video Generation

Combine up to five reference images into a coherent HD video with smooth motion and camera movement.

• Automatic Prompt Enhancement

Turn simple instructions into detailed prompts using the Bernini Prompt Enhancer and an online or local LLM.

⚙️ Workflow Notes

Built on Wan 2.2 with WanAnimatePlus ModelLoader and acceleration LoRA. Lower-VRAM users can increase BlackSwap to around 30–40. Reference images and videos can be enabled or disabled with Ctrl+B.