Obsession Tracker: An AI That Tells You How Obsessed You Really Are

Ashith Fernandes — Fri, 10 Jul 2026 11:00:47 +0000

This is a submission for Weekend Challenge: Passion Edition

What I Built

Obsession Tracker is a small AI agent that reads your day's work log — a commit message, a rant about a bug, a quick note about what you built — and tells you how obsessed you actually were. It scores your "passion level," gives you a witty one-line verdict, tags the type of grind (debugging, deploying, learning, whatever it was), and tracks it over time in a history chart.

The challenge prompt talked about "the love that fuels late-night side projects." I didn't have to invent that — I'm living it. Over the last few months I've debugged a structured-output integration until 2am, fought CORS errors deploying a backend to free hosting, and rebuilt a browser automation agent's logic from scratch after a library update broke it. None of that is glamorous. Most of it is staring at an error message for three hours. But it's exactly what the theme is about — the obsession that keeps you at the keyboard past the point where it's reasonable.

So I built the tool I wish I'd had this whole time: something that reflects that grind back at you instead of letting it disappear into a commit history nobody looks at twice.

My intended goal was to keep it dead simple — one input, one AI call, one score — so the "obsession" itself stays the star, not the app around it.

Demo

Live app: https://dev-challenge-obsession-tracker.vercel.app/

Video : https://streamable.com/9uao3j

You paste or type a quick log entry, hit Analyze, and Google's Gemini API (gemini-flash-lite-latest) returns a Passion Score (0-100), a one-line verdict, and a category tag. Entries are saved locally, and a running chart shows your passion trend over time. I seeded it with real entries from my own recent projects, so the history isn't hypothetical — it's a log of actual chaos.

Code

https://github.com/spideyashith/dev-challenge.git

Built with Next.js + Tailwind CSS on the frontend, Google's Gemini API for passion analysis (structured JSON output via responseMimeType: "application/json"), and browser localStorage for persistence — no backend database needed. Charts via Recharts. Deployed free on Vercel.

DEV Community: Ashith Fernandes

Obsession Tracker: An AI That Tells You How Obsessed You Really Are

What I Built

Demo

Code