DeflickerFrames
ComfyUI custom node for removing brightness flicker and chunk boundary artifacts in AI-generated video sequences (WAN, VACE, FramePack, etc.). bash cd ComfyUI/customnodes git clone https://github.com/karcsiha/comfyUi-deflicker.git Restart ComfyUI. The node will appear under the deflicker category. 1. Download this repository as ZIP 2. Extract to ComfyUI/customnodes/ComfyUI-deflicker/ 3. Restart ComfyUI - PyTorch (already included with ComfyUI) - No additional dependencies Find the node under Add Node → deflicker → Deflicker Frames. The node outputs two images: - images — corrected frame sequence - debugheatmap — visualization of corrections (red = brightened, blue = darkened, black = no change). Connect to a Preview node to inspect. | Parameter | Default | Description | |-----------|---------|-------------| | mode | stepremoval | stepremoval: instant correction of sharp latent space shifts (recommended). both: step removal first, then temporal smoothing. temporalsmoothing: Gaussian window-based correction for random flicker only. | | channels | L | L: brightness only — preserves original colors. LAB: brightness + color correction. | | | | Step Removal | | stepstrength | 1.5 | Step detection sensitivity. 0 = off, 1 = full correction, >1 = more sensitive detection (catches smaller steps) without brightness drift. Ignored in temporalsmoothing mode. | | | | Temporal Smoothing | | smoothstrength | 1.0 | Temporal smoothing strength. 0 = off, 1 = full, >1 = overcorrect. Ignored in stepremoval mode. Caution: values above 1.5 may introduce artifacts. | | smoothwindow | 25 | Tempora…
Deflicker · 1 input · 13 parameters · 2 outputs