A small local web app for generating a series of images from a "story"

I have been messing around with a small web app in Python.
It runs in your browser and talks to ComfyUI on your machine over the API.

Use it when you need a stack of images, each with a slightly (or very) different prompt, and you want captions later. If characters or objects must stay the same across the stack, you describe them once and reuse the names.

A book is one project. Each book has its own settings, story, captions, character stills, and output folder.

The story is a list of scenes. Each scene has a name, then a prompt:

HAR1 a woman in an orange flight suit…
p01_wreck CHAR1 climbing out of a crashed spacecraft
p02_sword CHAR1 pulling a sword from the cockpit

Names like CHAR1 pull that description into the prompt. You can attach up to three reference images per character or object.

You choose the Krea 2 model and LoRAs (other models are untested). You can generate two versions of each scene (or more). Run the book several times, pick the best frame per scene, then attach captions from a flat list — you do not rewrite captions every run.

There is a two-pane option: put p01_start and p02_then side by side as one image. Captions still follow the scene names.

Needs Python and a running ComfyUI. If you want to try it, say so.

Also, see this: https://civitai.red/articles/35064/my-atempt-to-create-a-craphic-novel-or-is-is-a-comic-with-the-script-we-built