A Proof of Concept for an end-to-end AI asset workflow — from Midjourney generation through Python batch processing to a playable React frontend.
Live demo on Netlify · Pipeline docs
Midjourney → manual curation → Python / ComfyUI batch processing → PNG assets → React game
- Generate — structured Midjourney prompts per card category (politician, initiative, intervention)
- Process — Python scripts and a ComfyUI workspace handle batch enhancement and export
- Ship — finished assets land in
public/assets/images/, mapped insrc/data/gameData.ts, consumed by the frontend
Details, prompt templates, and compliance notes: docs/ai-pipeline.md
Portions of this game's visual assets were created with generative AI tools and were reviewed, edited, and integrated by human creators.
Artwork was generated with Midjourney (paid commercial subscription), then curated and modified by the developer. See Midjourney Terms of Service.
React · TypeScript · GSAP · Howler · CSS 3D-transform board scaling · Netlify
Turn-based game engine with AI opponent, deck builder, and 600+ simulated balance matches.
git clone https://github.com/CosmicSlothOracle/Milchcards.git
cd Milchcards
npm install
npm run devProduction build deploys via netlify.toml — push to main and Netlify rebuilds automatically.
npm run build
npx serve -s build