Dungeon Crawler Class

• 

  Navigate the dungeon scenarios

• 

  Discover and share your custom class

Inspiration

Inspired by one of my favorite book series, Dungeon Crawler Carl! I wanted to create something for fans and the Reddit community to discover which class they might have in the dungeon.

Classes in the Dungeon Crawler Carl universe are anything but ordinary. Are you a Doppelgänger Drama Teacher or maybe a Tigran Cat Content Creator? Discover what the Dungeon System AI recommends for you!

What it does

Dungeon Crawler Class is a Reddit app that helps you find the class that best matches you and your decision-making. Navigate through a series of dungeon-based scenarios, uncover your recommended class, and share it with the community!

How I built it

In the Dungeon Crawler Carl universe, when crawlers reach the third floor, they’re given the option to change their race and choose a class. The dungeon’s AI system provides a list of recommendations based on each crawler’s skills and decisions.

With this premise in mind, I used Bolt and Devvit to create a decision-based game (similar to Reigns to emulate navigating the dungeon and making tough choices. I wrote a series of scenarios and used Bolt to build the logic that maps each decision to specific personality traits.

After completing all the scenarios, a personality profile is generated to determine the recommended race and class. Race recommendations are handled through code logic and trait mapping. For class recommendations, the personality profile, race, and a prompt are sent to OpenAI to generate personalized class names and descriptions.

Once a crawler selects their class, they can share and post their unique result in the Reddit thread.

Challenges I ran into

When calling the OpenAI API for class generation, I often encounter internal server errors, which cause long wait times or lead users to an error screen. I’ve tried several solutions, such as early authentication warming and automatic request retries, but the issue persists. I'm continuing to explore ways to resolve it.

I also faced challenges in creating a customized first screen on the Reddit feed.

Accomplishments that I'm proud of

Dungeon Crawler Class now runs on my test subreddit! I’m especially proud of the decision-making micro-interactions that let crawlers swipe or tap to make choices. I also successfully used Bolt and Devvit to integrate with the OpenAI API and generate customized classes for users.

What I learned

I learned how to build a Reddit app using Bolt and how to integrate the OpenAI API to generate personalized class recommendations.

What's next for Dungeon Crawler Class

I want to add more variety to the scenarios. Right now there are only 12 situations but I want to explore what it would be like if we randomize the situations and add branching logic.

I also want to add a customized first screen on the Reddit feed!

Built With

• bolt
• devvit
• openai-api