ComfyCanvas

ComfyCanvas
★ 94

交互画布图像编辑ComfyUI 插件实时预览
为 ComfyUI 提供交互式画布界面,便于手绘遮罩、图像合成与可视化控制,增强手动编辑与管线联动能力。
💡 在 ComfyUI 中手绘遮罩与直接编辑图像。
🍴 5 Forks💻 JavaScript🔄 2025-07-10
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/51137d50651f
📄 README

ComfyCanvas

Canvas to use with ComfyUI

Installation

  • Clone this repo into ComfyUI/custom_nodes
  • git clone https://github.com/taabata/ComfyCanvas.git
  • Run setup.py after changing directory to ComfyCanvas
  • cd ComfyCanvas
    
    python3 setup.py 
  • In a terminal/cmd tab,start ComfyUI
  • In another terminal/cmd, run main.py
  • python3 main.py 
  • Enjoy!!
  • Note: Windows users may face issues due to paths, let me know in the issues tab.

    Buttons description & hotkeys

    Set Ref.: select a part of the image to be third output in OutputCanvasNode. (https://github.com/taabata/LCM_Inpaint_Outpaint_Comfy.git) “s”

    Reset Ref.: reset the reference part of the image to be the whole image. “r”

    Crop: crop the image. “c”

    Compare: compare the image with what it was before the generation ends. “o”

    Move: move the image (toggled by clicking on canvas), resize the image by scrolling up and down, flip vertically by scrolling to the left, and flip vertically by scrolling to the right. “m”

    Erase: erase parts of the image (toggled by clicking on canvas). “e”

    Mask: mask parts of the image to manually specify the mask of the image (second output in OutputCanvasNode) (toggled by clicking on canvas). “k”

    Draw: draw on the image, color specified from color input on settings bar, and opacity by first slider on settings bar (toggled by clicking on canvas) “d”

    Fit: resize image to 512×512 pixels. “f”

    New: upload new image (discard the current image). “n”

    Text: add text to image. “t”

    GLIGEN: gligen or regional diffusion. “l”

    Blank: turns the image into a blank image based on selected color and opacity. “b”

    Grid: turns grid on. “g”

    COMFY: generation from ComfyUI (outputs from OutputCanvasNode will be selected).

    Undo: undo last image.

    Reset Mask: reset manual mask.

    Reset Parameters: reset changed workflow parameters.

    Lock To Image: locks selector to only select inside image. “right shift”

    Left,Top,Right,Bottom: blocks selector to respective side of image. “arrow keys”

    Reset Image: resets image, clears drawing, erased part, solves some issues (press if you face an error or generation doesnt start). “spacebar”

    Generate: generate image. “enter”

    “h”: hide settings bar.

    Rec. Routine: record processes on an image as a routine to be applied to other images.

    Start Routine: start a saved routine after selecting it from the dropdown list.

    Other GitHub repos to clone (workflows include nodes from these):

    https://github.com/taabata/LCM_Inpaint_Outpaint_Comfy.git

    https://github.com/kijai/ComfyUI-segment-anything-2.git

    https://github.com/WASasquatch/was-node-suite-comfyui.git

    https://github.com/kijai/ComfyUI-Florence2.git

    https://github.com/JettHu/ComfyUI-TCD.git

    https://github.com/cubiq/ComfyUI_IPAdapter_plus.git

    https://github.com/kijai/ComfyUI-KJNodes.git

    https://github.com/dfl/comfyui-tcd-scheduler.git

    https://github.com/Derfuu/Derfuu_ComfyUI_ModdedNodes.git

    https://github.com/ZHO-ZHO-ZHO/ComfyUI-BRIA_AI-RMBG.git

    https://github.com/Fannovel16/comfyui_controlnet_aux.git

    https://github.com/BadCafeCode/masquerade-nodes-comfyui.git

    https://github.com/taabata/Comfy_Syrian_Falcon_Nodes.git (just download the .py file and put in ComfyUI/custom_nodes)