Dungeons and Dragons Companion

Inspiration

DND is a very hard game to dig into, especially for beginners. While playing online with only a single monitor, it was difficult for me to juggle the game window, Discord, the handbook, and my character sheet all at once, in the heat of combat.

To deal with this, I uploaded the handbook and my sheet to Gemini so that I could simply ask it about aspects of rules and my character so I could quickly make decisions without juggling windows. However, due to the size of the rulebooks, it took an incredible amount of time to find what I was asking for, and it rarely understood the rules in context to my character and the situation.

That's why we came up with Dungeons and Dragons Companion! We are striving to ease the player experience by making it easy to ask questions about the rules and get fast and customized answers. We also have a tool to keep track of the game progress for the whole table to see.

What it does

This allows users to upload a DND (Or any TTRPG!) rulebook, character sheet, and also a lore sheet if desired. An AI model then parses these documents and populates a relational database for quick access to the information later. After the uploads, it moves to a game sheet that lets the table see characters' move timing and inventory. It also has a chat feature that allows the user to talk to a chatbot that can update the data we have, e.g., increasing the level of a character or spell. It can also be asked questions about the rules in context with the player's situation. Giving much faster, well informed and cited responses than a standard general-purpose model would.

How we built it

We used Genie to develop a backend and pipelines. We then used React to create a functioning frontend that achieves immersion and clarity. Inside of that, we developed a chatbot to allow users to update data and ask questions regarding their situation. This is done by having a genie model retrieve data from the database and then putting that information into a digestible and simple explanation.

Challenges we ran into

We ran into two major challenges, first being scope creep, which led to our second issue of wasted tokens due to backtracking.

Accomplishments that we're proud of

Despite roadblocks and a major setback, we continued to push through and made a frontend that we are proud of and feel as though it's well designed. We also were able to complete our first hackathon, which we are very proud of.

What we learned

We learned about the intricate details in Databricks and how to effectively use the AI model genie. We also learned about UI design and UX design since this project is very focused on the user enjoying the game and feeling immersed.