This is a single/dual-person digital human video application based on LTX 2.3 Image-to-Video + PromptRelay Segmented Plot Control + Three-Stage Rendering Refinement + Dual Latent Upscaling. Users upload a character's first-frame image, fill in global prompts and segmented prompts, and the system generates stable character motion videos based on the reference image. This is not an audio-driven version, but a silent digital human generation version: it focuses on controlling character actions, expressions, interactions, positioning, and camera rhythm through text prompts, making it more suitable for solo performances, two-person interactions, plot storyboards, AI short drama clips, and silent talking-head materials.
The core advantages of this workflow are: single-image character fixation, PromptRelay plot control, and three-stage rendering for stable visuals. The first stage establishes basic motion, the second stage enhances details, and the third stage provides final refinement. Combined with Latent Upsampler and tiled decoding, it makes single or dual-person visuals more stable and delicate. Compared to regular image-to-video, it is better suited for "controllable digital human performance," especially for generating single/dual-person short videos with fixed cameras, fixed characters, and fixed left/right positioning. When using it, it is recommended to focus on controlling three items: the character's first-frame image determines identity, the Global Prompt determines overall rules, and Local Prompts determine each segment's action and plot.
🎁Claim RH Coins first, then go experience the workflow! Avatar in top right corner → Invitation Code → Enter 【rh-v1111】 to instantly get 1000RH Coins, and log in daily to claim another 100 coins~ For more ComfyUI workflows, tutorials, and gameplay, please follow our WeChat Official Account (AIKSK). Synchronously updated on Douyin / Bilibili / Xiaohongshu / YouTube (AI-KSK)✨


No creations yet

No creations available.