Dan 71c010dd60 Merge pull request #1 from ltdrdata/endpoint
use custom endpoint instead of /view, /upload
2024-09-04 22:15:18 -05:00
2024-09-03 14:21:29 -05:00
2024-09-03 06:54:23 -05:00
2024-09-04 09:30:13 -05:00

ComfyUI Prompt Gallery

What is this?

A custom node that adds a UI element to the sidebar that allows for quick and easy navigation of images to aid in building prompts.

image

Features

  • Snazzy sidebar menu that holds images and metadata for prompt building
  • Built in support for all of Navimixu's Wildcard Files
    • If you get all these you've got access to ~750 goodies!
  • Easily add your own images to the Custom category
  • Search and Sort functions
  • Automatic metadata extraction for prompt tags (only works with default ComfyUI nodes ATM)
  • Drag and drop images with embedded workflows to load it on the canvas (similar to how the new queue works)

Installation

  • git clone this repo into your ComfyUI custom nodes folder

    • There are other ways to do this since it's just a js file but if I ever do add nodes better safe than sorry. The repo is tiny enough you can just download it and stick it in there, I won't tell.
  • Copy the promptImages folder into your ComfyUI INPUT folder (yes I know it's not great but I can't use user directories yet)

  • Download any or all (or none if you're gonna be like that) of Navimixu's archives and extract them into the promptImages folder. If you do it right your folder should look like this with all the images in various subfolders under thumbnails

    image

Requirements

  • ComfyUI 0.1.3+
  • There's no real requirements here since it's just a js file, enjoy!

Usage

  • Click on any image in the gallery to copy its prompt tags to the clipboard to paste into your workflow.
  • Add some images to the sidebar yourself via the Custom category.
  • You can also drag and drop images with embedded workflows onto the canvas to load their workflow.
  • With some editing can support other YAML files, if you're cool like that.
  • If you use Pony in particular the Art Style category makes it super easy to see what various tags can do!

Limitations

  • Uses the Input directory for image storage. (boo)
  • Uses local storage for setting and metadata, not terrible but a bummer if you change browsers or computers.
  • Doesn't do well getting metadata from complex workflows or prompts in non-core nodes.

Roadmap

  • Learn new FE
  • Exposing plugin options via settings
  • Easier support for additional "wildcard" packages
  • Prompt stacking? (build a prompt by clicking on various pictures)
  • Better prompt handling and QoL enhancements
  • Some actual "wildcard" style features
  • Whatever bugs you find / features you submit

Why

image

I got excited by the new front end and wanted to do something solely based in that if I could. It was also fun to just work in FE for a bit. If you're looking for a simple example of something that leverages the new sidebar, toasts, and the new png metadata scraper, this could be just what you're after!

I aslo thought it would be cool to let people play with something that's not LoRA based. Many checkpoints now have a lot of built in knowledge at this point, and whenever you don't need a Lora, the better in my book!

Credits

Navimixu

Comfy Org (Duh)

https://github.com/comfyanonymous/ComfyUI

Compatability

Tested on ComfyUI 0.2.0 Should work on any version 0.1.3+

S
Description
No description provided
Readme GPL-3.0
134 KiB
Languages
JavaScript 93.6%
Python 6.4%