My journey building a Steam price tracker

The Problem:
I have a bad habit: I buy Steam games on sale, only to realize the price drops lower a week later, or that the same "sale" happens every single month. I was getting tired of the FOMO.

The Solution:
I built steamAF (a PWA). Instead of just showing "current sale" like every other store, it checks the current price against ~2 years of historical data to give a verdict: Worth it, No rush, or Wait.

The "Aha!" Moment (Technical Build):
When I first started, I was naive. I set up a script that polled the Steam API for thousands of games every 5 minutes. It was inefficient, cost-heavy, and frankly, a recipe for getting rate-limited.

I had to rethink the architecture. I switched to a Smart Dynamic TTL approach using Supabase:
- If a game is on sale, I cache the data until endsAt + 30 mins.
- If it's at full price, the cache stays fresh for 12 hours.
- No cron jobs, no wasted API calls, and I'm saving 90% of my upstream traffic.

Tech Stack:
I'm running this on Next.js 15, React 19, and Supabase (Postgres). Moving away from "dumb" polling to a dynamic caching strategy was the most satisfying part of this build.

What I'm looking for:
I'm currently at the stage where I’m focused on polishing the "Worth it" verdict accuracy. If anyone here is working on price-tracking tools or e-commerce scrapers, I'd love to hear how you handle upstream API limitations!

Also, if you're a Steam gamer and feel the same "sale FOMO," I'd love for you to give it a spin and let me know if the "Wait" verdict makes sense to you.