Pítia
A dashboard that reads a Steam library and says what to play next, and why.
- Year
- 2025
- Role
- Personal project, built solo
- Stack
- Next.js, React, TypeScript, PostgreSQL, Drizzle, and Recharts
- Status
- Live

Pítia reads a public Steam library and answers what Steam itself won’t: which genres you actually play, measured in hours, what you never opened, what you dropped in the first couple of hours, what you’re close to finishing, and what to play next, with the reason written out.
The live version is a demo with a single library that stands in for the public profile of the YouTuber Jacksepticeye. The games and store data are real; the hours and achievements are made up. Pítia is not affiliated with Valve. The full version reads any profile whose game details are public, through the Steam Web API.
The recommendations don’t use AI. Each game gets a score that blends affinity with what you play (45%), review score (35%) and length (20%). Affinity weighs hours by their square root, so a single 2,000-hour game can’t define your taste on its own.
The technical decision I like most works around a Steam limit: the store answers one game at a time, and only about once every 1.5 seconds. I built the catalog from 25 requests to SteamSpy, about 86,000 games in around 90 seconds. A new library now opens in about 3 seconds, and the missing details fill in in the background.
More screens



