Getting Started

Installation

Via ComfyUI Manager (Recommended)

1. Open ComfyUI
2. Open Manager → Install Custom Nodes
3. Search for UmeAiRT Toolkit
4. Click Install and restart ComfyUI

Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/UmeAiRT/ComfyUI-UmeAiRT-Toolkit.git
pip install -r ComfyUI-UmeAiRT-Toolkit/requirements.txt

Restart ComfyUI after installation.

Prerequisites

Requirement	Details
ComfyUI	Latest version
Python	3.10 – 3.13
PyTorch	With CUDA (NVIDIA), ROCm (AMD), or MPS (Apple Silicon)
VRAM	≥6 GB recommended (4 GB minimum with GGUF Q4 models)

Optional: aria2c for fast downloads

The Bundle Loader/Downloader uses aria2c for multi-connection downloads when available. Without it, downloads fall back to urllib (single-thread).

Your First Workflow

The UmeAiRT Toolkit uses a block architecture where each node handles one concern:

graph LR
    A["⬡ Checkpoint Loader"] --> D["⬡ KSampler"]
    B["⬡ Generation Settings"] --> D
    C["⬡ Positive Prompt"] --> D
    D --> E["⬡ Image Saver"]

1. Add a ⬡ Checkpoint Loader → select your model
2. Add ⬡ Generation Settings → set dimensions, steps, CFG, sampler
3. Add ⬡ Positive Prompt Input → write your prompt
4. Add ⬡ KSampler → connect all three inputs
5. Add ⬡ Image Saver → connect the pipeline output

Bundle System

For an even simpler setup, use the ⬡ Bundle Auto-Loader which auto-downloads models. See Bundle Auto-Loader.