Card Czar

Back in the beginning of 2024, I had a small problem: my Magic: the Gathering card collection was growing out of control. I have a bad habbit of buying players’ entire collections so that I can pick out the cards from the 90’s that I actually want. I would prefer to buy these cards individually from a store, if only there were a store that had all of these old cards in inventory. So, as an excuse to grow my collection further, I decided to open that store.

This created a big problem: I now had to organize my collection so that I could find the cards players ordered from me. Most card stores laboriously sort the collections they buy and integrate these sorted cards into their previously sorted inventory, which is beyond stupid. Due to the size of my inventory (i.e., over 9000), the complexity of Magic’s 30 year history with over 28,000 unique cards across 100+ sets, and other numbers, this was simply too much data for the single-threaded meat-computer in my head to process. I just knew that there had to be a better way.

Given my background in Human-AI Interaction, it seemed appropriate to create an artifically intelligent business partner that would handle the thinking for me. Then, using only the Python packages installed on my computer, I created the Card Czar. The Card Czar classifies images of Magic cards that I put through my scanner and stores their locations in a database. When I sell one of these cards online, the Card Czar tells me where to find it.

The origin of the name Card Czar is a little silly. With its help, I don’t need to know what the cards are, where the cards are, or what the cards are worth; the Card Czar knows that for me. It constantly gathers new Magic card metadata from the internet and prices my inventory using an increasingly sophisticated data-driven policy. This means I have basically no idea what the Card Czar is doing and it’s better that way. I just fetch the cards when it sells them and spend more time enjoying my vintage Magic collection.

Shop Card Czar