Attached is a native Comfyui Wan2.1 Vace workflow.V1: Add a video for control, choose size and lenght, add an image for reference, a prompt, and you're good to go. V2: dynamic mask added (segment anything2, points selection system), you can choose between using mask with controlnet, or mask without controlnet. Dont use teacache native it give some wrong result with controlnet <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/08c5908f-0492-46a4-bdc5-b7c5266ed25e/width=525/08c5908f-0492-46a4-bdc5-b7c5266ed25e.jpeg" /> model Vace preview :<a target="_blank" rel="ugc" href="https://huggingface.co/ali-vilab/VACE-Wan2.1-1.3B-Preview">https://huggingface.co/ali-vilab/VACE-Wan2.1-1.3B-Preview</a> <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/d41592d0-7b91-4fde-9183-a79a7251b85f/width=525/d41592d0-7b91-4fde-9183-a79a7251b85f.jpeg" />

Files with SHA256: e725ccfd611c8b9fc6575f2126e959854b101b8c8ea4d9be5418e70f68ad88e4