swping999/scene-card-studio ? reverse-engineered prompt

Build me a Python project that turns a set of personal photos into a structured visual story.

I want the app to take ordinary snapshots, analyze what is actually in them, turn that into scene cards, then compile those into editable prompts and generate a more directed image result. It should keep the story structure separate from the visual style, so I can swap in different looks like a quiet editorial, family chronicle, travel journal, or dream like version without rewriting the whole narrative.

Please include a clean way to save the scene cards, prompt history, generation results, and review or retry notes so each run is easy to inspect later. If it makes sense, add a simple example workflow and a few sample cases like the ones in the repo, and make the output deterministic enough that the same inputs produce the same layout and narrative decisions.

Look up current docs online if you need to, and make the code easy to run and extend.