This tool is designed to make image generation easier, faster, and more consistent—especially for users who don’t want to manually write long prompts or user's who may be new to prompting for image/video generation.

Instead of typing everything out, you build prompts by selecting from curated options like character traits, clothing, poses, environments, and more using Danbooru tags. The system then combines everything into a clean, usable prompt.

Overview

Workflow Overview

This workflow provides a modular system of nodes that:

Break prompts into logical sections (character, clothing, pose, etc.)
Provide curated dropdowns instead of freeform typing
Combine everything into a final prompt automatically
Keep things flexible while providing user's with a list of words most models react well to.

Features

Structured prompt building (no prompt knowledge required)
Large curated dataset (~3000+ Danbooru approved tags)
Clean separation of prompt sections
Supports SFW and NSFW workflows
Highly customizable via presets.json
Easily expandable without touching core code

Installation

Option 1 — ComfyUI Manager (Recommended)

Open ComfyUI
Open Manager
Search for: Prompt Builder
Install
Restart ComfyUI

Option 2 — Manual Install

Manually download or clone the repository:

git clone https://github.com/zinigo-creations/comfyui-prompt-builder

Move the folder into:

ComfyUI/custom_nodes/comfyui-prompt-builder

Restart ComfyUI.

How It Works

The system is built around modular nodes that each handle a section of the prompt.

Each node pulls its options from presets.json, meaning everything is data-driven and easy to modify.

Nodes

Style Preset

Style Node

Defines the overall visual direction (anime, cinematic, fantasy, etc).
This has a major influence on the final output.

Preset Character

Preset Character Node

Quick-select from a large list of known characters.
Useful for fast generation and consistency.

Composition

Composition Node

Controls how the image is framed.

Includes:

camera type
angle
shot type
quality tags

Character

Character Node

Defines the subject of the image.

Includes:

body type
hair (color, style, extras)
eyes
clothing
accessories

This is the most detailed and impactful section.

Action

Action Node

Defines what the character is doing.

Includes:

pose
gestures
movement

Helps avoid stiff or static results.

Environment

Environment Node

Controls where the scene takes place.

Includes:

location
lighting
time of day
atmosphere

NSFW (Optional)

NSFW Node

Adds additional layered detail for more mature outputs.

This node is optional and can be ignored entirely.

Combine Prompt

Combine Node

Where the other nodes hook up and to merge all selections into a final prompt.

You can also:

add custom text
override generated output

Customization

All dropdown options are defined in:

presets.json

You can:

add new tags
modify existing ones
create entirely new sections

No code changes required.

Contributing

Contributions are welcome.

How to contribute

Fork the repository
Create a new branch
Make your changes
Submit a pull request

Guidelines

Keep naming consistent (e.g. black hoodie, not Hoodie Black)
Prefer commonly used tags (Danbooru-style works best)
Avoid duplicates and redundant entries (before adding use ctrl+f) to make sure it's not in the list already
Test your changes before submitting

Content Guidelines

Please use common sense and good judgment when contributing.

Do not add illegal, or otherwise harmful content
Avoid content that would make the project unsafe or inappropriate for general use
If a tag feels questionable, it probably is — leave it out

Notes

Not all tags behave the same across models
Simpler prompts often produce better results
This tool is designed to guide, not restrict

License

MIT

Final

If you enjoy the workflow or have suggestions, feel free to open an issue, PR or reach out.