Comfy-Org / Comfy-Org/ComfyUI

Pause node to adjust and add to the prompts

Open
#5,965 0 comments 0 reactions 0 assignees View on GitHub
Feature
Dominant language
Python
Stars
133k
Forks
15.7k
Avg merge
1d 6h
Merged PRs (30d)
155

Description

### Feature Idea

I am currently running Griptape with Ollama's and I made a Prompt reverse engineering workflow which works fine mostly. what it does is it takes an image without prompts or generation data, SEES what is in it (So I need Vision LLM's like Lllama 3.2 vision or gemini 1.5) and then tells me what it sees, then I parse it in to another agent that takes that info and turns that in to a prompt, which gives me "Similar" results but off course never perfectly the same, If I wanted that I would just download the pic much easier.
E.g. the original
![12c](https://github.com/user-attachments/assets/e4bb1fc7-e4f2-420b-8abc-bc17e920d106)
vs Mine
![fluxturbo_8_1 0_667](https://github.com/user-attachments/assets/fb101765-6a80-4181-b2c8-b78e624cc647)

BUT sorry for the long intro, LLM's sometimes get it wrong, or have the tendency to block certain things that appear remotely NSFW or change the prompt in such a way that it is SFW, E.g.
the original:
![34AC79C4F7AB09E918773F3B1360D7423644C5DF3E8AA8457E3A1372E98E5ABD](https://github.com/user-attachments/assets/6da64372-74db-4f6b-a1c0-cb29fc73ee46)
vs mine:

![fluxturbo_8_1 0_752](https://github.com/user-attachments/assets/10256037-9722-4a02-80ed-d690c19daca1)

So here is what I need, I want to parse the prompt outside the griptape and LLM sphere in to a node that pauses and asks me to adjust it, or add to it, just a this is the prompt, change it to your needs, then press continue, NOT JUST for the NSFW sometimes it just get's things wrong like instead of making a brutish male it makes an elegant almost female face etc... I need to change things like that before continuing.

TLDR: Make prompt, parse to pause node, adjust, continue

### Existing Solutions

I do not know of any I knew there was some text clip prompt that had a second window which could contain some extra words to add but I do not know which node that was.
Also it did not pause, it just added those words in.

### Other

You do not need to worry about the griptape nodes interfering, The standing spiderwoman has the workflow, it is a basic flux image generator separate from the griptape nodes, so that is not an issue

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.