ComfyUI-oshtz-nodes

Custom nodes for ComfyUI created for some of my workflows.

LLM All-in-One Node

Easy GPT/Claude integration in ComfyUI:

• OpenAI & Anthropic models
• Image-to-text capabilities

GPT Image 1 Node

Generate images via OpenAI's vision model:

• Direct OpenAI API integration
• Image editing capabilities with mask support
• Quality and size customization
• Transparent background option

Easy Aspect Ratio Node

Simplify your workflow with preset aspect ratios:

• Large selection of common ratios (1:1, 16:9, 21:9, etc.)
• Returns width and height values

String Splitter Node

Split text into multiple outputs:

• Up to 10 separate outputs
• Customizable separator

LoRA Switcher Node

Efficient LoRA switch made for API use:

• Switch between up to 40 LoRAs in a single node (10, 20, 40)
• Dynamic LoRA switcher for maximum flexibility
• Fine-tune strength

Image Overlay Node (Beta 🚧)

Combine images with precision:

Current Features

• Basic image compositing
• Position control (X/Y)
• Transparency adjustment
• Auto size matching
• Alpha channel support

Coming Soon

• Advanced positioning
• Transform controls
• Multiple blend modes
• Layer management
• Interactive preview

Installation

First navigate to your ComfyUI installation custom nodes directory.

1. Clone this repository:

git clone https://github.com/oshtz/ComfyUI-oshtz-nodes.git

2. Install the required dependencies:

cd ComfyUI-oshtz-nodes
pip install -r requirements.txt

Requirements

• requests
• torch
• pydantic
• Pillow
• boto3 (>= 1.34.101)

License

This project is open-source and available under the MIT License.