Slay The Spire Demake

[AI-ASSISTED CODE]

Play the current iteration of the game! (v0.2)

▶ Click to playarrows move · Z is A · X is B
Enter is Start

What This Is

This project started around March 2026 when Slay the Spire 2 came out. I never played the first game, but played the second game on PC because of all the hype. I had seen some videos of GBA de-makes (like this one for Balatro) and thought that Slay the Spire would actually work really well on a Game Boy Advance. This was also right around the time that AI was getting sorta crazy, so I tried a few prompts and asked it to use Butano to create a Slay the Spire clone. It got REALLY far for only a few amount of prompts. You can play that version of the game below!

▶ Click to playarrows move · Z is A · X is B
Enter is Start

But, when I went to play this game, it didn't feel right. It wasn't as satisfying as playing the real game, even though most of the core mechanics are there. I indulge in a few game development youtube videos from time to time (specifically Razbuten and Game Maker’s Toolkit) and from those videos I hypothesized that the core gameplay loop of my game just wasn't satisfying. It wasn't the same as clearing a minion wave in Deadlock or thinking about how to IGL a round in Valorant. So, naturally, as side projects go, I stopped thinking about it.

Until this week! (September 7th) I saw the posting to be a Valorant Game Designer with no game dev experience and thought this would be a good time to make some progress on this to pad up the ole' resume. (this entire site was actually made for this job post!)

Why I did this

The goal for this project was to upgrade this prototype game I made into something that I, the sole audience, wouldn't mind playing for a few hours. I wanted to make the game fun.

The goals were NOT:

• To learn the nuances of coding a GBA game in C++
• To learn how to make sprites
• To make any sort of profit off of this game

The goal was to make something FUN, and figure out what knobs I could tweak to make this game feel more fun.

What I learned

So yea! Here are some of the things I experimented with, and what I learned.

• Visual feedback matters
  • My initial version did not have health bars — it just had numbers on screen to show the health of an enemy mob. So, when clicking on a card and attacking, I didn’t feel any satisfaction or sense of accomplishment. When I added the health bar, suddenly clicking on a card felt so much more “powerful” because I could see the health bar deplete.
• How much I care about the character matters
  • I didn’t really care or have an attachment to the AI-generated Ironclad or Silent sprites. When I swapped out the icons for Pokémon sprites and swapped the names of the cards out with Pokémon moves, the game became magnitudes more engaging. I didn’t really expect this, since I had playtested the previous version without the sprites and that version was not very fun.
  • Fun fact, but I had the idea to insert Pokémon sprites because of a video I watched about the original Super Smash Bros., where I heard that the game was originally pitched with generic stick figure fighters, but somebody had the idea to plop in Nintendo characters before an executive pitch. I felt the same effect here!
• Information clarity matters
  • Due to the small resolution we were working with (240×160 pixels), I found myself memorizing the distinct icons of each of the cards and not even reading the description of each of the cards. This is slightly different from normal Slay the Spire, where the description can be seen on all of the cards at all times, so you don’t really need to memorize every card.
  • If I were to make some changes, I would really force card icons to be different, even if they didn’t pertain to the card description exactly
• Sound matters
  • Another small thing that had an influence on my enjoyment — adding silly beep boops to clicking a menu item and clicking a card made the game so much more tactile