3-Image Reference

This is a multi-image reference video generation application based on LTX 2.3 Dev-Dare + VBVR motion logic enhancement + Crisp image quality enhancement + PromptRelay timeline control + Three-image reference guidance + Three-stage rendering polish. After users upload 3 reference images, the system injects them into the video generation process as background, character, and props/interactive elements respectively, and then uses PromptRelay timeline prompt words to control how they appear, move, interact, and wrap up.

The core advantage of this workflow is: The three images have distinct divisions of labor, allowing background, characters, and props to be referenced separately without blending into a mess. Compared to standard single-image video generation, it is better suited for scenarios such as characters holding products, character-prop interactions, background+character+object combinations, brand showcases, storyboard shorts, and multi-reference composite footage. When using it, it is recommended to focus on controlling three items: 3 reference images determine element sources, global prompts (used to write overall visual rules) determine element relationships, and segmented prompts (describing action changes in each time period) determine interaction rhythm.
🎁Claim RH Coins first, then experience the workflow! Click the avatar in the top right corner → Invitation Code → Enter [rh-v1111] to instantly claim 1000RH Coins, and log in daily to claim another 100 coins~ For more ComfyUI workflows, tutorials, and gameplay, welcome to follow our WeChat Official Account (AIKSK), and stay tuned for simultaneous updates on Douyin / Bilibili / Xiaohongshu / YouTube (AI-KSK)✨