Shipping the MVP

Day two of the project, and the plan was to get something a friend could actually click on. Not the finished game, just enough of it: create a room, submit a song, see who won.

The bet underneath all of it was that the game could run without a traditional always-on server sitting behind it watching every move. What that means for a player is simple even if getting there wasn't: everyone in a room sees the same thing update at the same time, without me having to run and babysit a server process around the clock for a free party game.

The core interaction, pasting a Spotify link and having it turn into an actual track with art and a title, worked embarrassingly early, which was encouraging right up until the embed saga started. I added a Spotify embed so songs would play inline. It played at full volume with no way to turn it down. I found this out the way you'd expect, ripped it back out the same day, and moved on. Shipped it, hated it, deleted it, all before dinner.

Then came deploying it, which is its own genre of commit message: "vercel", then "vercel fix?", then a small stack of "fixesss" right after it. None of that is dramatic, it's just what day-one deploys actually look like when you're finding out what breaks in production that never broke locally.

By the end of the day there was a working leaderboard, a room you could create, and a game a friend could theoretically join. Search Console's first recorded day for the site is exactly this one: zero clicks, zero impressions. Google hadn't found it yet. That would take a little longer.