Card & Conquer

• 

  card & conquer poster

Inspiration

I always have a thing for TCG style game since I'm a kid. I've played like Yu-gi-oh, Pokemon, Duel Master, MTG, and beside the fun gameplay, I really love the "collecting" nature of it. Translating a TCG to a Reddit game gives a fun challenge, especially as a hackathon where I only got limited time to execute my idea. How to create a TCG where the player is anyone in any number, where the selling point is still the "collecting" part. Thankfully by the advance of AI, doing this seemingly daunting task for a hackathon can be achieved. AI image generation is advance enough to give me quick assets that is suitable with my vision for this game as a prototyping tool/hackathon game. But the MVP should be given to Kiro because there's no way I could build this much without utilizing vibe-coding. Kiro basically did EVERYTHING from the start to finish. I only provided the assets without touching the slightest bit of coding.

What it does

Card & Conquer is a collectible card battler featuring 200+ mythological gods from 8 different pantheons (Greek, Norse, Egyptian, Japanese, Chinese, Hindu, Irish, and Korean). Players:

• Collect cards through a gacha system with multiple card variants
• Deploy gods in strategic 10v10 battles against other players
• Trigger real-time combat with 8 unique tactical abilities (FirstStrike, Reinforcements, Precision, etc.)
• Compete globally in a persistent faction war between East and West pantheons
• Climb leaderboards and earn spots in the Hall of Fame

All of this happens seamlessly within Reddit posts without external links, just pure native gameplay.

How we built it

First I created a document about all the description and requirement of the game. After I scaffolded a new Devvit project, I opened the project with Kiro where it recognized the context instantly. From there I asked Kiro to create a spec based on the document that I created earlier. What's good is Kiro automatically created a placeholder for the card where the assets were still yet on progress. This allowed me to run two things at the same time, monitoring the coding progress and creating the assets. Even when I had a change of mind about a feature, there's not much of difficulties creating new spec to handle it.

Challenges we ran into

All bugs can be solved by communicating with Kiro, though what I realized is solving bugs seems like consuming more tokens than building from scratch. Perhaps if I were more specific and detailed about the flow and mechanism of a page or UI's function from the start, I could've save more token, but overall it was a great experience because I also do realize that some things will always be changed in the middle of development.

Accomplishments that we're proud of

In just two weeks I could build a full playable TCG game that I always dream of building. Of course it all thanks to AI advancement especially Kiro.

What we learned

With only 2 weeks, every hour counted. Kiro's ability to generate, refactor, and debug code at high speed was crucial. What would take hours manually took minutes with Kiro. This game can happens in just two weeks, imagine what I can build in months. Or at least, that's where the spirit should be.

What's next for Card & Conquer

There are a lot of things that I want to improve upon Card & Conquer. What I think is top priorities probably:

• Game balancing. Definitely need a lot of players to really understand whether a game is balanced or not. And the ultimate aim of every game maker of course a balanced game
• Different Battle/Game mode. It will be cool if players can play multiple kind of games with the card collection that they have.
• More cards. More characters, more pantheons, perhaps more abilities.

Built With

• devvit
• kiro
• react
• redis
• typescript
• vite