🐱 Le Chaton Fat

Frontier feline intelligence. Somewhere in the millions though.

A polished, meme-native idle/clicker game where you feed a cute French tabby treats to inflate its "parameters." As Le Chaton grows it gets visually bigger, unlocks more absurd treats and upgrades, climbs a fake AI benchmark leaderboard, and keeps turning down Anthropic offers.

Built to look like a serious AI model launch page that slowly devolves into nonsense.

Stack

• Next.js 16 (App Router) + React 19 + TypeScript
• Tailwind CSS v4
• Framer Motion for animations
• localStorage for persistence — no login, no backend

Run it

npm install
npm run dev      # http://localhost:3000
npm run build    # production build

What's in the box

• Click-to-feed loop — feed the cat, watch it bounce, spawn crumb particles and a flying treat, fill the stage progress bar, and occasionally hear a quote.
• 10 treats that unlock as parameters grow (Croissant -> OpenAI Researcher).
• 6 compute upgrades with passive params/sec and 1.25x price scaling.
• 7 cat stages — the cat scales up and eventually overflows the screen.
• Absurd number formatting — 30T -> 999Q -> 5M Q -> "Somewhere in the millions though."
• Anthropic offer events every 30-90s. Turn it down for a 2x boost, or accept and become Claude Fable 6 (then undo it).
• Fake benchmark leaderboard where Le Chaton's scores inflate past 100 into SOTA / Peer review failed while the real models get tiny bars.
• 8 achievements with slide-in toasts.
• Share card with a "Copy Text" button for screenshots.
• Mobile-first, with a sticky feed button on small screens.

All game logic lives in app/page.tsx; data/tuning in app/lib/game.ts; the beret-wearing CSS cat in app/components/Cat.tsx.

---

Le Chaton Fat is not a real frontier model. The benchmarks are not real. The offer was, allegedly, somewhere in the millions though.