comfyui-2dpose-editor

ComfyUI 2D Pose Editor Node

A ComfyUI custom node that embeds an interactive 2D rigging figure directly inside the node. Drag joints to pose the figure, then capture it as an IMAGE output.

ノード内で 2D リギングフィギュアをドラッグしてポーズを編集し、IMAGE として出力する ComfyUI カスタムノードです。

2D Pose Editor Version

---

screenshot

---

Features / 機能

• 🦴 Full body rigging — head, neck, chest, abdomen, arms, hands, legs, feet
• 👁️ Eye control — drag the pupils to move the gaze direction
• 🎥 Camera control — pan (drag background) and zoom (mouse wheel) with reset
• 🖼️ Texture atlas — load a single PNG to skin all body parts via UV mapping (L/R separated UVs)
• 👤 Head toggle — switch between front face and back of head
• 👕 Body toggle — switch between front and back torso
• ✊ Hand toggle — switch each hand between open and closed (viewer perspective)
• 👁️‍🗨️ Rig visibility — show/hide skeleton overlay with stadium-shape outlines
• 📸 One-click capture — saves the current pose as a PNG image (always rig-free output)
• 🔄 Reset pose — return to the default pose instantly
• 🖼 Background compositing — connect a background IMAGE to composite the pose over it
• 📂 Image Input Mode — switch the node to image loader mode (P/I toggle), supports drag & drop
• 📐 Output size mode — Standard / Background / Custom width×height
• 🎨 Background color picker — set a solid background color; ✕ button clears it to transparent
• 📂 Background image loader — load a local image file as background; ✕ button clears it
• 🖼️ Aspect ratio overlay — frame overlay shows the exact output region for Background / Custom modes
• ✂️ Aspect-correct capture — captured image is cropped to the active frame (correct W×H output)

---

Installation / インストール

Option A: Clone into custom_nodes

cd ComfyUI/custom_nodes
git clone https://github.com/ketle-man/comfyui-2dpose-editor.git

Option B: Manual

1. Download this repository as a ZIP and extract it
2. Place the 2dpose_custom_cm folder inside ComfyUI/custom_nodes/
3. Restart ComfyUI

---

Usage / 使い方

1. In ComfyUI, right-click the canvas → Add Node → 2D Pose → 2D Pose Editor
2. Drag the white dots (joints) to pose the figure
3. Drag the black pupils to change the eye direction
4. Use the toggle buttons to switch head / body / hand appearance
5. Optionally load a custom texture via 🖼 Tex
6. Click 🦴 to toggle skeleton overlay visibility before capturing
7. Click 📸 (Capture) to save the current frame
8. Click Queue Prompt — the node outputs the pose as an IMAGE
9. Click RP (Reset Pose) to return to the default pose / RC (Reset Camera) to restore the view

Image Input Mode

• Click the I button to switch to image loader mode
• Click 📂 to load a local image file, or drag & drop an image onto the node
• The loaded image is passed directly as the node output
• Click P to return to pose editor mode

Background Color & Image

• 🎨 BG color picker — sets a solid fill color behind the pose (default #e0e0e0)
  • Click ✕ next to the picker to go transparent (useful when background_image is connected)
• 📂 BG button — loads a local image file as the background layer; displays letterboxed to fit the canvas
  • Click ✕ next to the BG button to clear the loaded image

Output Size

• Std — use the canvas render size (600×600)
• BG — match the aspect ratio of the loaded background image (or connected background_image)
• Custom — specify width × height manually
• An overlay darkens the area outside the active output frame so you can see exactly what will be captured

Camera controls

Operation	Action
Drag background	Pan the camera
Mouse wheel	Zoom in / out
🎥 Reset Camera	Retu