Nodes/sweet-tea-nodes/Wan VACE → Video (caps inpaint)
ComfyUI Node

Wan VACE → Video (caps inpaint)

WAN VACE cap-stitch inpaint node for hard clip joins. Treats start/end cap frames as preserved known regions and the middle as unknown generated region, with enforced neutral filler to prevent hidden evidence leakage.

By tea-time-labs·Created 5 months ago·Updated 23 days ago· 0
Wan VACE → Video (caps inpaint)
  • positive
  • negative
  • vae
  • control_video
  • reference_image
  • positive
  • negative
  • latent
  • trim_latent
width832
height480
length81
batch_size1
strength1.00
start_control_frames24
end_control_frames24
ref_strength-1.00
ctrl_strength-1.00
ref_noise_std0.00
Categoryconditioning/video_models

Inputs (15)

NameTypeDefaultDescription
positiveCONDITIONING
negativeCONDITIONING
vaeVAE
widthINT83216–16384
heightINT48016–16384
lengthINT811–16384
batch_sizeINT11–4096
strengthFLOAT1.000–1000
start_control_framesINT240–100000
end_control_framesINT240–100000
control_videooptIMAGE
reference_imageoptIMAGE
ref_strengthoptFLOAT-1.00-1–1000
ctrl_strengthoptFLOAT-1.00-1–1000
ref_noise_stdoptFLOAT0.000–2

Outputs (4)

NameTypeDescription
positiveCONDITIONING
negativeCONDITIONING
latentLATENT
trim_latentINT