FindThaGame is a discovery tool that helps you find video games using only fuzzy memories and vague details. It uses a custom scoring algorithm to rank results from the IGDB database so the most relevant titles actually float to the top.
Live Demo: https://findthagame.vercel.app/
Repo: https://github.com/kbtale/findthagame
I got tired of having random games stuck in my head but never being able to find the name, so I built this so people could hunt things down based on the random details they actually remember rather than needing to know the exact title.
Most searches are binary, but this uses a Hybrid Multiquery Search that runs strict phrase matching and broad keyword searches through summaries and storylines simultaneously. A custom scoring engine then ranks the results by adding or subtracting points based on multiple filters.
The project is built with React, TypeScript, and Vite, using Tailwind and Shadcn UI for the styling. It runs on Vercel Serverless functions and uses Groq AI for fast real-time summaries and translations. To keep performance high, I implemented server-side token caching for the IGDB API and a warmup middleware to handle serverless cold starts.