Poker Settle-Up: From Idea to Table Overnight

Overview

I have always wanted to build a small companion app for poker nights, especially for casual games where the chips do not map 1:1 to the cash buy-in, such as a $50 buy-in for $250 worth of chips. Settling up is usually the least enjoyable part of the night: everyone is tired, chips are being counted in uneven piles, and somebody inevitably ends up doing the math for the group.

With a poker game coming up at a social gathering with colleagues, I decided to see what I could build using the AI tools and techniques I had picked up. I spent the night before putting together Poker Settle-Up, then brought it to the table the following day.

The app gives the group a shared room for recording buy-ins in real time. At the end of the game, each player enters the value of their remaining chips, and the app calculates the results and produces a practical settlement plan that minimises the number of payments required.

There are a few useful extras too: a private mid-game estimate of where each player stands, a rules panel showing the poker hand hierarchy, and a host mode that allows one person to enter counts on behalf of anyone who would rather not do it themselves.

I didn’t want the final screen to feel like a receipt so every participant receives a result card complete with a Singapore-flavoured title, such as Huat King, Steady Lah or PayNow Ambassador, along with a rotating subtitle. It began as a fleeting thought of wanting to leave something that people can remember the poker night experience by regardless of their wins or losses, and it ended up becoming one of the most popular parts of the app.

First Night at the Table

The first real test went better than I expected. Everyone joined the shared room on their phones, the buy-ins stayed in sync, and the app required almost no explanation. I have since added a QR code that anyone in the room can generate, making it even easier for other players to access the web app and join the game.

The group really liked the hand rankings available throughout the game, which were particularly useful for casual and first-time players, and receiving a clean list of who should pay whom at the end. These were the practical features I had in mind while vibecoding the app, so seeing them work well at an actual poker game made the time spent building it worthwhile!

The feature that earned actual gasps, followed by a “you actually have this?”, was the discrepancy adjustment. I had included it without expecting that we would need it on the very first night. After everyone declared their remaining chips, the final count contained 40 more chips than should have existed. The likeliest explanation was an innocent mistake by one of us while exchanging larger denominations for smaller chips from the bank, but there was no sensible way to trace it by that stage.

The app allowed us to record the discrepancy, and spread the excess value equally across the table, which everyone was happy to accept.

Several friends have already asked to use the app at future games. Some even joked that I should monetise it and launch it on the App Store.What I took away from this fun project was another reminder that we are living in the fourth industrial revolution. AI know-how, a willingness to experiment, and a practical eye for user experience can now turn a passing thought such as “this could be useful” into something that real people can use, and genuinely want to use, the very next day.

Back to AI Journal