BrainRot Index

Drop your tweets, shower thoughts, or 3AM notes and AI scores your brainrot across 6 cognitive dimensions with tarot readings, philosopher roasts, and historical parallels. Powered by Google Gemma.

Comment

Inspiration

What if we could score how chaotic someone's thinking is not to fix it, but to celebrate it? We're all posting unhinged shower thoughts at 3 AM, pitching pigeon-based startups to nobody, and writing tweets that would make philosophers weep. But nobody's actually measuring the rot. BrainRot Index was born from a real cognitive scoring framework I've been building — one that maps to actual research in metacognition, analytical rigor, and creative thinking. I thought: if we're going to understand how humans think, we might as well start with the unhinged stuff first.

What it does

Drop your tweets, shower thoughts, 3 AM notes, startup pitches, or unhinged group chat takes and AI scores your brainrot across 6 cognitive dimensions:

  • 🗣️ Yap Density — how much you say without saying anything
  • 🎨 Unhinged Creativity — the sheer chaos energy of your ideation
  • 🌀 Overthink Spiral — depth of your existential rabbit holes
  • 💅 Delulu Confidence — how delusionally confident your take is
  • 👑 Main Character Energy — how hard you're narrating your own movie
  • 📱 Doomscroll Resilience — your commitment to the bit You get a score out of 100, a personality type (like "Existential Speedrunner" or "Delulu Degen"), a tarot card reading, a philosopher who would've vibed with your take, a historical parallel to your level of chaos, and a shareable Wordle-style card. There's a Hall of Rot leaderboard across 5 categories (Most Cooked, Yap Champion, Delulu Award, Certified Overthinker, Main Character) plus a Rot Archive that tracks your thinking patterns over time with sparkline trend charts. ## How we built it
    • Frontend: Next.js 16 + React 19 + TypeScript + Tailwind CSS
    • AI: Google Gemma 3 12B running locally via Ollama on an RTX 4090
    • Design: Comic book aesthetic — halftone patterns, speech bubbles, comic panels, tarot cards, custom SVG doodle character faces, glitch effects, scanlines
    • UX polish: Sound effects (Web Audio API), confetti on high scores (80+), animated score counter, typewriter effect on roasts, step-by-step loading animation, Cmd+Enter shortcut, clickable example prompts, light/dark mode
    • Icons: Phosphor Icons (duotone + fill weights)
    • Share: Wordle-style text card with dimension bars, share to X Each of the 6 brainrot dimensions secretly maps to a real cognitive science dimension (Analytical Rigor, Creative Range, Metacognitive Depth, Intellectual Humility, Synthesis, Persistence) from a scoring framework I developed for a separate project. ## Challenges we ran into
    • Getting local LLMs to consistently output valid JSON with all required fields (tarot, philosopher, historical parallel) solved with robust JSON extraction and retry logic
    • Making the comic book aesthetic work in both dark AND light mode ended up with 30+ CSS custom properties
    • Balancing "useless and chaotic" with "technically impressive" the app needs to make people laugh AND make engineers ask "wait, how does the scoring actually work?" ## Accomplishments that we're proud of
    • Score 88 on our own app — we're officially Chronically Cooked
    • Google Gemma 3 12B running locally on an RTX 4090 with zero API costs, zero latency — the entire AI stack runs on one GPU
    • 6 brainrot dimensions that secretly map to real cognitive science — the joke has actual research underneath it
    • Full comic book aesthetic in pure CSS — halftone dots, speech bubbles, panel borders, tarot cards, manga action lines — zero image assets ## What we learned Running Google's Gemma 3 locally is surprisingly good for structured output tasks. The comic book CSS aesthetic (halftone dots, speech bubbles, panel borders with box-shadows) is achievable with pure CSS no images needed. And the best way to showcase real technology is to wrap it in something absurd. ## What's next for BrainRot Index The serious version of this scoring framework is called Discuria a platform where crowd-sourced problem solving gets scored across 6 research-backed cognitive dimensions. Think GitHub, but for thinking. BrainRot Index is the chaotic gateway to that vision.

Built With

  • canvas-confetti
  • google-gemma-3
  • motion
  • next.js
  • ollama
  • phosphor-icons
  • react
  • tailwind-css
  • typescript
  • web-audio-api
Share this project:

Updates