A custom node system for ComfyUI that enables efficient prompt management, selection, and interpolation. This extension provides a YAML-based datastore for organizing prompts into categories and includes powerful text interpolation capabilities for dynamic prompt composition.

Features

Prompt Categories: Organize prompts into logical categories
Prompt Selection: Easy-to-use node interface for selecting and using stored prompts
Text Interpolation: Powerful variable substitution and text combination patterns
YAML Datastore: Persistent, human-editable prompt storage
Dynamic Composition: Combine multiple prompts with variable substitution
ComfyUI Integration: Seamless integration with ComfyUI workflows

Installation

Clone or download this repository into your ComfyUI custom_nodes folder:

cd custom_nodes
git clone https://github.com/yourusername/comfyui-promptstore.git
cd comfyui-promptstore

Install dependencies:

pip install -r requirements.txt

Nodes

Prompt Template

The Prompt Template is a simple node - it just allows you to enter some text to be used as a template for your prompts.

This node is here just so you don't have to install other custom nodes just to enter some text.

Prompt Selector

The Prompt Selector node allows you to select and modify prompts from your stored categories with optional variable interpolation.

Inputs

category (required): Select from your defined prompt categories
prompt_title (required): Select a specific prompt from the chosen category
prompt (required): The actual prompt text (can include interpolation patterns)
initial_prompt (optional): An initial prompt to pass through interpolation

Output

Prompt: The processed prompt string (after interpolation if applicable)

Example Usage

Text Interpolation

The text interpolation system allows you to create dynamic prompts by combining strings and substituting variables. It supports three main patterns:

Pattern 1: Placeholder Expansion (`{}`)

The {} placeholder in the text will be replaced with the initial_prompt after processing any variable definitions from it.

For example, a simple substitution can look like this:

Since you can chain the Prompt Selectors you can also get something like this:

Pattern 2: Variable Substitution `${varname:value}`

That's already good, but a bit limited. Sometimes you may want to have more complex prompt creation, where something stays fixed, and other things may change. So you may have a prompt that says:

Henry is standing in the foreground with a wide stance.

And you may have different subjects you want to describe as Henry. What you can do is define the subject name as a variable and then use it in a saved prompt. So - the prompt above - could be written as:

${subject:Henry} is standing in the foreground with a wide stance.

And you can have a saved prompt for your subjects that looks like:

${subject} is a middle-aged Greek fisherman with short gray hair and a weathered skin. Wrinkles on his face and
deep hazel eyes.

{}

The output prompt would be:

Henry is a middle-aged Greek fisherman with short gray hair and a weathered skin. Wrinkles on his face and
deep hazel eyes.

Henry is standing in the foreground with a wide stance.

A picture is worth a thousand words:

Pattern 3: Inline Structural Substitution (`@{var}`)

The @{var} pattern performs structural substitution. It will replace the placeholder in the text with the prompt defined with the same variable name. So, if your prompt is:

The subject is standing in front of @{background}.

The subject is happy, wondering how it all works.

And you selected a prompt for a background like this:

@{background}
a large decorative window

The output will be:

The subject is standing in front of a large decorative window.

The subject is happy, wondering how it all works.

Take a look at this example showing how these can be combined:

A bit about the prompt Storage

Prompts are stored in a YAML file - in user/comfyui-promptstore/promptstore.yaml with the following structure:

categories:
  - portrait
  - landscape
  - character

portrait:
  - A professional headshot portrait
  - A romantic portrait with soft lighting
  - A dramatic character portrait

landscape:
  - A scenic mountain landscape at sunset
  - An expansive valley under clear skies

character:
  - A fantasy warrior with detailed armor ${armor:steel}
  - A mysterious figure shrouded in shadow @{description}

YAML Datastore API

The underlying YAML datastore provides:

Category Methods

get_categories() - Get list of all categories
create_category(category) - Create a new category
remove_category(category) - Remove a category and all prompts
category_exists(category) - Check if category exists

Prompt Methods

get_prompts(category) - Get all prompts in a category
add_prompt(category, prompt) - Add a prompt to a category
remove_prompt(category, prompt) - Remove a prompt from a category
prompt_exists(category, prompt) - Check if a prompt exists
get_all_prompt_titles() - Get all prompt titles across categories

Utility Methods

get_all_data() - Get complete datastore contents

Usage Examples

Basic Workflow: Simple Prompt Selection

Add a Prompt Selector node to your ComfyUI workflow
Choose a category from the dropdown
Select a prompt from that category
The selected prompt text flows to your image generation nodes

Advanced Workflow: Variable Substitution

Create prompts with variables to avoid duplication:

style_prompts:
  - A ${style:photorealistic} portrait of a person
  - A ${style:painting} of a landscape
  - A ${style:digital art} character design

Then in your node, select the prompt and it will automatically substitute the style variable.

Advanced Workflow: Structural Substitution

Build complex prompts by combining base templates with specific details:

templates:
  - A ${mood:peaceful} scene @{details}
  
details:
  - with a sunset in the background
  - with mountains visible in the distance
  - with people enjoying the view

Tips & Best Practices

Organize by Purpose: Use categories that reflect your workflow (e.g., "character_descriptions", "environment_details", "style_modifiers")
Use Variables for Common Values: Create variations with ${var:default} patterns to avoid repeating similar prompts
Chain Prompts: Use the @{var} pattern to build complex prompts from multiple stored segments
Keep Prompts Reusable: Write prompts that work well with interpolation rather than hard-coding specific details
Document Complex Patterns: Add comments in your YAML file (using #) to explain what variables are used in each prompt

Requirements

Python 3.6+
PyYAML
ComfyUI

Troubleshooting

Categories not appearing: Ensure your YAML file is properly formatted and categories are defined
Interpolation not working: Check that variable names match exactly (they're case-sensitive)
Prompt not updating: Restart ComfyUI after modifying your YAML prompt file

License

See LICENSE file for details.

README.MD

ComfyUI Prompt Store