ComfyUI-zyd232-Nodes
ComfyUI nodes including LLM Text Generation for any LLM server (multiple image/video/audio input, preset management, hide API-key, clean VRAM, auto-unload), Merge LoRA Stacks, Image Pixel Comparison, and other nodes for quality-of-life improvements.
Nodes (7)
Are These Two Images Identical? A Strict Pixel Check That Returns a Yes/No
Stitch your 5-second clips without leaving the graph
A Full LLM Inside Your ComfyUI Graph — Local or API, No Heavy Setup
The tiny node that tells Ollama to put your VRAM back
Add, Max, or Average Them Into One
Merge Multiple LORA_STACK Lists Into One
Format, Quality, Workflow JSON, and Date-Stamped Folders
ComfyUI zyd232 Nodes
A collection of custom nodes for ComfyUI, including LLM text generation, image pixel comparison, image save/preview, and mask batch blending.
📖 Full documentation is available in the Wiki — each node is documented individually with its features, inputs/outputs, and usage tips.
📦 Installation
cd ComfyUI/custom_nodes
git clone https://github.com/zyd232/ComfyUI-zyd232-Nodes.git
Restart ComfyUI. The nodes will appear under the zyd232 Nodes category.
🧩 Node Index
| Node | Category | Description | Docs |
|------|----------|-------------|------|
| LLM Text Generator | zyd232 Nodes/LLM | Connect to any OpenAI-compatible LLM service for text generation. Supports multiple images/videos/audio, streaming display, config presets, and result locking. | 📄 Docs |
| LLM Unload | zyd232 Nodes/LLM | Send an unload signal to the LLM Server to release the model from VRAM. All parameters come from the config preset saved by the LLM Text Generator node. Supports selecting a specific model to unload. | 📄 Docs |
| Images Pixels Compare | zyd232 Nodes | Compare two images at the pixel level to check if they are identical; outputs a boolean. | 📄 Docs |
| Save Preview Images | zyd232 Nodes | Save images (PNG/JPG) with quality, metadata, custom path, workflow JSON, and preview options. | 📄 Docs |
| Mask Batch Blend | zyd232 Nodes | Blend multiple masks into one using add / max / average operations. | 📄 Docs |
| Merge LoRA Stacks | zyd232 Nodes | Merge multiple lora_stack inputs into a single lora_stack output. The left-side inputs are dynamic (Autogrow) and can be added freely. | 📄 Docs |
| Join Videos | zyd232 Nodes | Join multiple videos into one, in input order. Accepts file paths and VIDEO inputs through two dynamic (Autogrow) groups; outputs VHS_FILENAMES and VIDEO, joining losslessly when the segments already match the target format. | 📄 Docs |
🚀 Quick Start
Type a node name in the ComfyUI node search box to find it. For detailed usage of each node, click the docs link above or browse the Wiki home.
📄 License
This project is released under the LICENSE.