ComfyUI-oshtz-nodes

Custom nodes for ComfyUI created for practical prompt, LoRA, image API, and utility workflows.

Nodes

LLM All-in-One

Easy GPT, Claude, and OpenRouter integration in ComfyUI:

• OpenAI, Anthropic, and OpenRouter providers
• Live OpenRouter model list fetched through the node backend
• Image-to-text capabilities for vision-capable models
• API keys via widget inputs or environment variables: OPENAI_API_KEY, ANTHROPIC_API_KEY, OPENROUTER_API_KEY

LoRA Switcher (Dynamic)

Dynamic LoRA selection powered by a single configurable node:

• Drive LoRA choice via API-friendly widget data
• Unlimited rows with custom strengths managed from the UI
• Keeps workflows simpler than juggling fixed 10/20/40 variants

GPT Image 1 (Direct API)

Generate or edit images through OpenAI's image API:

• Text-to-image generation
• Optional image and mask inputs for edits
• Quality, size, moderation, and transparent background options
• Environment variable fallback via OPENAI_API_KEY

Easy Aspect Ratio

Select common image ratios and output ratio text, width, and height.

String Splitter

Split text into multiple outputs:

• Up to 10 separate outputs
• Customizable separator

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
• pydantic
• Pillow
• numpy

ComfyUI provides the runtime tensor and image stack used by the nodes.

License

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