Make your dancer appear anywhere.
A ComfyUI custom node that composites a masked dancing person video onto any background image. The person's feet are always anchored to the exact position you specify, regardless of scale — enabling realistic placement of a dancer into any scene.
- Foot-anchored placement — set
pos_x/pos_yand the person's feet always land there, even as you adjust scale - Independent scale controls —
scalesets the overall size (as a fraction of background height);height_ratiostretches or squishes the person without changing width - Mask edge feathering — Gaussian blur on the alpha mask for natural blending
- Brightness matching — optionally blend the person's brightness toward the local background for better lighting consistency
Category: 🎬 Video/Composite
| Parameter | Type | Description |
|---|---|---|
person_video |
IMAGE | Dancing person video frames (batch) |
person_mask |
MASK | Alpha mask per frame — connect from a background-removal node (e.g. InspyreNet, rmbg) |
background |
IMAGE | Background image to composite onto |
pos_x |
FLOAT 0–1 | Horizontal foot position (0 = left, 1 = right) |
pos_y |
FLOAT 0–1 | Vertical foot position (0 = top, 1 = bottom) |
scale |
FLOAT | Person height as a fraction of background height (0.9 = 90%) |
height_ratio |
FLOAT | Fine-tune height only: >1 taller, <1 shorter, width unchanged |
edge_feather |
INT | Mask edge blur radius in pixels |
brightness_blend |
FLOAT | Blend person brightness toward background (0 = off, 1 = full match) |
| Output | Type | Description |
|---|---|---|
composited_video |
IMAGE | Video frames composited onto the background |
[Load Video] ──────────────────────────────────────────┐
↓
[InspyreNet / rmbg] ← video → mask ──→ [🕺 Dance Video Composite] → composited_video
↑
[Load Image] ──────────── background ──────────────────┘
Via ComfyUI Manager — search Dance Everywhere.
Or manually:
cd ComfyUI/custom_nodes
git clone https://github.com/GuardSkill/ComfyUI-DanceEverywhere