YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

🖼️ ComfyUI Binyuan · Multi Image Compare (N)

A ComfyUI custom node for comparing N images (and multi-frame batches) side by side — slider A/B mode and grid mode, with optional label baking and a composite grid export. Unlike rgthree's Image Comparer (A/B only), this one supports any number of inputs.

适配 ComfyUI v0.26 / 新前端(comfyui-frontend-package 1.45.x)。


中文说明

这是什么

节点 🖼️ binyuan · Multi Image Compare (N),分类 image。 支持 N 张图片 + 多帧 batch 对比:节点内用滑块拖分割线对比任意两张,或网格并排查看所有图;可把自定义文字烧到每张图左上角;最终把所有端口所有帧横向拼成一张大图导出。

怎么用

  1. 放到 ComfyUI/custom_nodes/binyuan_multi_image_compare/,重启 ComfyUI。
  2. 右键 → image🖼️ binyuan · Multi Image Compare (N)
  3. 把若干 IMAGE(如 VAEDecode 输出)依次连到 image_1image_2… 连满一个端口会自动加下一个空槽(最多 16 个)。也可右键节点 → "➕ 添加图片输入"。
  4. label 多行文本里每行写一个端口的标签(第1行→image_1,第2行→image_2…)。
  5. 运行后:节点内可 Slider / Grid 对比;把 grid 输出连到 SaveImage 存成一张大图。

参数

参数 说明
label 多行文本,每行对应一个输入端口,烧到该端口图片左上角
print_label 是否把 label 烘焙到图片上(默认 true)
label_size 字号 8–200(默认 32,支持中文微软雅黑)

输出

  • grid (IMAGE):所有端口所有帧横向拼成的一张大图。

节点内对比操作

  • 顶部工具栏:Slider/Grid 切换;Slider 模式下用 A vs B 下拉选要对比的两张; 翻转; 重置缩放。
  • 图片区:滚轮缩放(0.2x–10x)、拖动平移、双击重置;Slider 模式下拖动=移动分割线。

典型场景

  • 对比多个模型:3 个 VAEDecode 分别连 image_1/2/3,label 填模型名,运行后导出带名字的合成图。
  • 多帧对比:某端口接 batch(视频帧/多步采样),所有帧都进对比序列和合成大图。

常见问题

现象 解决
图片区不在节点内/乱跑 Ctrl+F5 硬刷新浏览器(JS 没更新)
文字没印上 检查 print_label=true 且 label 填了对应行
改参数没生效 后端改动需重启 ComfyUI(不只刷新浏览器)
字体异常 系统缺微软雅黑,会回退默认字体

屏幕截图 2026-07-01 115316

Uploading 屏幕截图 2026-07-01 120127.png…

English

What it is

Node 🖼️ binyuan · Multi Image Compare (N) under category image. Compare N images and multi-frame batches: in-node slider A/B compare or grid view of all images; bake per-port text labels onto images; export a single composite grid image of every frame from every port.

How to use

  1. Drop into ComfyUI/custom_nodes/binyuan_multi_image_compare/, restart ComfyUI.
  2. Right-click → image → add the node.
  3. Connect IMAGE sources (e.g. VAEDecode) to image_1, image_2… A new empty slot appears automatically when one is filled (up to 16). Right-click node → "➕ 添加图片输入" to add manually.
  4. Write one label per line in label (line 1 → image_1, line 2 → image_2…).
  5. Run: compare with Slider/Grid inside the node; connect grid to SaveImage to save the composite.

Params

label (multiline, per-port text), print_label (bool, default true), label_size (8–200, default 32, CJK font supported).

Output

grid (IMAGE): one wide composite of all frames from all ports.

In-node controls

Toolbar: Slider/Grid toggle; A vs B dropdown to pick two for slider mode; flip; reset zoom. Canvas: wheel to zoom (0.2x–10x), drag to pan, double-click to reset; in Slider mode drag = move the split line.

Files

  • nodes.py — backend: inputs/outputs, label baking, grid composition
  • js/multi_image_compare.js — frontend: compare UI, zoom, DOM overlay positioning
  • __init__.py — registration (WEB_DIRECTORY = "./js")
  • 使用说明.txt — original Chinese manual
  • README.md / LICENSE (MIT)

Install

  • ComfyUI Manager: search Binyuan Multi Image Compare.
  • Manual: git clone https://github.com/yjitiu/ComfyUI-Binyuan-Multi-Image-Compare.git binyuan_multi_image_compare
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support