CIVITAI / Workflows

SCAIL-2 Single-Person Reference Editing Long-Video Workflow

Watch the full video first if you want to understand how this SCAIL-2 single-person reference editing workflow works in practice. The video shows how one reference character can replace the selected person in a driving video, while the workflow preserves pose motion, original scene structure, identity consistency, lighting coherence, and long-video continuity. This ComfyUI workflow is designed for SCAIL-2 single-person biological reference editing. Its main purpose is to place one reference character onto the target person in a single-person driving video. Unlike the single-person driving workflow, this version is not only about making a reference character follow motion. It is explicitly configured for character replacement, using skeleton guidance, subject tracking, colored masks, reference identity encoding, and long-video continuation. The workflow is built around wan2.1_14B_SCAIL_2_fp8_scaled.safetensors as the main SCAIL-2 model. It also uses WAN VAE, UMT5 XXL WAN text encoding, CLIP Vision, SAM3, SCAIL2ColoredMask, WanSCAILToVideo, SamplerCustom, VAEDecode, ForLoop continuation, overlap-frame trimming, ColorTransfer, final video combining, and original audio restoration. A multi-LoRA enhancement chain is also preserved to improve motion quality, visual stability, and final rendering consistency. The most important switch in this workflow is replacement_mode=true. This tells the SCAIL route to perform single-person skeleton guidance with character replacement. The reference image provides the replacement character identity, while the driving video provides the target motion and scene structure. The positive prompt focuses on replacing the selected single target person, following one-person pose guidance, keeping the original scene structure, preserving consistent identity, natural motion, coherent lighting, and smooth temporal consistency. The negative prompt is also designed for this task. It suppresses bad video quality, flicker, wrong-area replacement, identity drift, deformed bodies, distorted faces, extra limbs, missing hands, warped hands, broken anatomy, blur, and low-quality output. This is important because single-person replacement often fails when the mask is inaccurate, the reference image is unclear, or the driving video contains heavy occlusion. The workflow uses strict 512×896 alignment. Both the reference image and the driving vide