ComfyUI-ZhiHui
智绘灵箱 - 功能丰富的 ComfyUI 自定义节点插件，提供图像处理、画板绘制、文本操作等多种实用功能
Nodes (42)
Add film grain to a finished render without touching the sampler
The rembg cutout node with fourteen models and a proper mask output
The prompt-based cutout that's still a stub — read this before wiring it up
Composite a cutout onto a new background with three blend modes
Nano Banana 2 in your graph via a reseller API — key required, no local GPU
The Gemini/nano-banana wrapper that points at VectorEngine instead of Google
Scan a whole folder of images into one batch, with filenames to match
Pull frames and audio out of any video in a folder, one index at a time
Brightness, contrast, saturation, sharpness
The colorist's node
The Doubao (高清增强版) API node
Draw, paint and scribble over your generation without leaving ComfyUI
The image-passthrough node that (maybe) deletes files — safety switch included
Dry-run, copy, or move files between folders — with a trigger input that does nothing
The texture generator you didn't know you needed
The grid assembler with cell-size control
Stretch, crop, or pad
Image-to-prompt reversal with BLIP, Qwen2-VL and full generation controls
Crop by pixels or percentage — the exact-cut node with no guesswork
Borders, gaps, spacing and auto layout
A folder stepper that remembers where it was
The batch-vs-counter sibling
A picture selector that won't feed you a dead input
The whole folder, normalized into one batch — with pad/crop/stretch rules
LoadImage that hands you the filename — and a mask you'll want to double-check
Up to 20 loose images, one batch — with free size-matching
A list, not a batch — the node that keeps your images separate on purpose
The node that unloads your models before you OOM
Your favorite prompts, saved to a JSON file you can edit by hand
Roll a structured prompt from ten ingredient lists
One text box full of candidates, one random winner per run
The three-line text cleaner that quietly fixes your prompts
One image in, a 4×4 tile batch out — the grid-split workhorse
Load one image from a folder by index — and survive going out of bounds
Save anywhere, with numbering that never collides and metadata you control
Files named 2026年08月28日14时30分45秒 — the timestamp saver for Chinese users
A real multi-layer PSD straight out of ComfyUI — rare, and yes it needs pytoshop
Walk a tree of folders and pull a slice of images out of each
Turn one text into many — the node that makes ComfyUI run your prompt list
Drag segments on a real timeline, and feed your video model per-shot prompts
Unpacking VIDEO_INFO into numbers you can actually use
Audio muxing, CRF control, and a real filepath out
ComfyUI-ZhiHui (智绘灵箱)
一个功能丰富的 ComfyUI 自定义节点插件，提供图像处理、文本操作、批处理等多种实用功能。
A feature-rich ComfyUI custom node plugin providing image processing, text operations, batch processing and various utility functions.
🚀 安装方法 / Installation
方法一：Git 安装 (推荐) / Method 1: Git Installation (Recommended)
在 ComfyUI 的 custom_nodes 目录下运行：
Run in your ComfyUI's custom_nodes directory:
git clone https://github.com/zhuyungen/ComfyUI-ZhiHui.git
然后重启 ComfyUI 或使用 ComfyUI Manager 重新加载节点。 Then restart ComfyUI or reload nodes using ComfyUI Manager.
方法二：手动安装 / Method 2: Manual Installation
- 下载本项目的 ZIP 文件
- 解压到 ComfyUI 的
custom_nodes目录 - 安装依赖：
pip install -r requirements.txt - 重启 ComfyUI
方法三：ComfyUI Manager 安装 / Method 3: ComfyUI Manager
如果本插件已被 ComfyUI Manager 收录，可直接在 Manager 中搜索 "智绘灵箱" 或 "ComfyUI-ZhiHui" 进行安装。
📋 依赖要求 / Requirements
torch
numpy
Pillow
pytoshop
安装依赖：
pip install -r requirements.txt
🎯 主要功能 / Main Features
🖼️ 图像处理 / Image Processing
- **智绘_画板 **: 支持多图层绘画、抠图、图形编辑
- 智绘_九宫格拼图: 批次图片网格排列拼接
- 智绘_图片切换器: 多图像智能切换选择
- 智绘_制作批次: 单图转批次处理
- 智绘_纯净图片加载器: 增强版图片加载器
- 智绘_批量图片加载器: 文件夹批量图片加载
- 智绘_高级保存图片: 增强版图片保存功能
- 智绘_带时间保存图片: 自动时间戳图片保存
- 智绘_保存PSD: 多图层PSD文件生成
📝 文本处理 / Text Processing
- 智绘_文本分割器: 按分隔符分割文本
- 智绘_去空行: 清理文本空行
- 智绘_随机提示词提取: 随机提取文本行作为提示词
🛠️ 实用工具 / Utilities
- 智绘_文件清理器: 自动清理临时文件
- 智绘_预设加载器: 预设参数快速加载
📚 节点列表 / Node List
图像类节点 / Image Nodes
| 节点名称 | 英文名称 | 功能描述 | |---------|---------|---------| | 🎨 智绘_画板 | Drawing Board | 多图层画板，支持画笔/橡皮/文字/图形/抠图，可接收上游图片输入 | | 🪣 智绘_九宫格拼图 | Grid Image Assembler | 批次图片网格排列拼接 | | 🔄 智绘_图片切换器 | Image Multi Switch | 多图像智能切换选择 | | 📦 智绘_制作批次 | Make Image Batch | 单图转批次处理 | | 🖼️ 智绘_纯净图片加载器 | Image Loader Plus | 增强版图片加载器 | | 📁 智绘_批量图片加载器 | Image Batch Loader | 文件夹批量图片加载 | | 💾 智绘_高级保存图片 | Advanced Save Image | 增强版图片保存功能 | | ⏰ 智绘_带时间保存图片 | Save Image With Time | 自动时间戳图片保存 | | 💾 智绘_保存PSD | Save PSD | 多图层PSD文件生成 |
文本类节点 / Text Nodes
| 节点名称 | 英文名称 | 功能描述 | |---------|---------|---------| | ✂️ 智绘_文本分割器 | Text Split By Delimiter | 按分隔符分割文本 | | 🧹 智绘_去空行 | Remove Empty Lines | 清理文本空行 | | 🎲 智绘_随机提示词提取 | Random Prompt Extract | 随机提取文本行作为提示词 |
工具类节点 / Utility Nodes
| 节点名称 | 英文名称 | 功能描述 | |---------|---------|---------| | 🗑️ 智绘_文件清理器 | File Cleaner | 自动清理临时文件 | | ⚙️ 智绘_预设加载器 | Preset Loader | 预设参数快速加载 |
📖 使用说明 / Usage Guide
详细的使用说明请参考 DATONG_GUIDE.md 文档，包含：
- 每个节点的详细功能介绍
- 参数配置说明
- 使用技巧和注意事项
- 典型应用场景示例
- 工作流推荐
For detailed usage instructions, please refer to DATONG_GUIDE.md, which includes:
- Detailed function introduction for each node
- Parameter configuration instructions
- Usage tips and precautions
- Typical application scenarios
- Recommended workflows
🔧 开发信息 / Development Info
- 版本: 1.1.0
- 兼容性: ComfyUI
- Python版本: 3.8+
- 许可证: MIT License
🤝 贡献 / Contributing
欢迎提交 Issue 和 Pull Request 来改进这个项目！
Welcome to submit Issues and Pull Requests to improve this project!
📄 许可证 / License
本项目采用 MIT 许可证 - 详见 LICENSE 文件
This project is licensed under the MIT License - see the LICENSE file for details.
🙏 致谢 / Acknowledgments
感谢 ComfyUI 社区的支持和贡献！
Thanks to the ComfyUI community for their support and contributions!
注意: 如果在使用过程中遇到问题，请检查是否正确安装了所有依赖项。
Note: If you encounter issues during use, please check if all dependencies are properly installed.