ComfyUI-RandomSwitch

ComfyUI Random Switch

A ComfyUI custom node that randomly selects one input from multiple inputs of the same type.

Features

• Random Selection: Randomly pick one value from 2-32 inputs
• Flexible Input Types: Accepts any type (images, latents, conditioning, etc.) as long as all inputs are the same type
• Seed Control: Use an external seed for reproducible results, or auto-generate a seed on each run
• Dynamic Input Growth: UI automatically expands to show the next input when all current inputs are connected

Installation

1. Clone or download this repository into your ComfyUI custom_nodes directory:

   cd ComfyUI/custom_nodes
   git clone https://github.com/tttamaki/Comfyui-RandomSwitch.git
   

2. Restart ComfyUI

Usage

Basic Setup

1. Add the Random Switch node to your workflow (search for "Random Switch")
2. Connect 2 or more inputs (images, latents, etc.) - all must be the same type
3. The node will randomly select and output one of the connected inputs

With Seed Control

• Leave seed at 0 to get a random selection each run
• Connect a seed value for reproducible results (same seed = same selection)
• The node outputs both the selected value and the seed used

Example Workflows

• Random Image Selection: Connect multiple images and randomly pick one
• Random Latent Selection: Use with latent inputs for probabilistic sampling
• Random Conditioning: Switch between different conditioning inputs

Inputs

• inputs (Autogrow): 2-32 inputs of any type (must all be the same type)
• seed (Int): Random seed for reproducibility (0 = auto-generate)

Outputs

• output (Any): The randomly selected input value
• seed (Int): The seed value that was used

License

MIT License

Author

tttamaki and GitHub Copilot