# I made a ComfyUI workflow that turns a reference image + a simple idea into a production-ready prompt — retained record

> Retained reddit record from reddit:comfyui.

Canonical: https://highsignal.app/data/records/a8c98a982dd6186c

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:comfyui)  published 2026-09-14 17:52 ingested 2026-09-15 02:40

[original source ↗](https://www.reddit.com/r/comfyui/comments/1wganl3/i_made_a_comfyui_workflow_that_turns_a_reference/)

## retained text



```
I kept finding myself doing the same thing over and over: load a reference image, analyze it, figure out the aspect ratio, write a detailed image prompt, then separately write a video prompt/script.

So I built Prompt Studio to do that inside ComfyUI.

Give it a reference image and a simple direction, and it can generate:

* A detailed production-ready still prompt
* Target-model-specific prompting for Generic, FLUX, Krea 2, or SDXL
* A structured MiniMax H3 video script
* Automatic aspect-ratio detection
* Separate Still and Video directions, so they don't have to describe the same thing
* RECREATE mode for faithful reconstruction or DIRECT mode when you want to change the reference

The demo is using **Krea 2** as the image target and MiniMax H3 for the video script.

It uses OpenRouter for the multimodal analysis/generation. Your API key stays in your local workflow and isn't included in the public download.

GitHub: [ComfyUI-Prompt-Studio](https://github.com/phibbyrizz/ComfyUI-Prompt-Studio)

CivitAI: ComfyUI-Prompt-Studio

Curious whether anyone else has been building this kind of “prompt pre-production” directly into ComfyUI.
```



document fetched 2026-09-15 00:42 · canonical https://reddit.com/r/comfyui/comments/1wganl3/i_made_a_comfyui_workflow_that_turns_a_reference
