Project Wildspace

Inspiration

Playing D&D with a table full of friends is a lot of fun, but for the Dungeon Master, planning could take countless hours, and coming up with fully involved and engaging people, places, or plotlines is difficult. Project Wildspace was created to revolutionize the current market for D&D concept generators, and is one of a kind in terms of its generation. Most of the tools on the market currently suffer one fatal flaw - context. Helping you create elements for a world is one thing, but for a generator to return something that fits is another challenge entirely. By implementing a world input file management system, we were able to inject world context into each generation request, allowing for some seriously funny and exhilarating pieces of lore.

What it does

Using the Qt Creator Graphical User Interface, Project WIldspace enables the user to create important elements of any fantasy world - NPCs, Buildings/Scenes, Towns, Encounters, and Groups/Factions. Stemming from the canonical D&D concept of Wildspace, this software peers into the infinitely many universes that exist and returns memorable components to any fantasy world.

How we built it

We first created a UI with Qt Creator, including responsive layouts and styling after the thematic Wildspace colors. After, we implemented the functionality and event handlers within the UI and bound them to multiple AI generation objects which we created for each use case. For the AI side of our project, we leveraged the Langchain module to ensure the GPT returns were bound to a strict output format which we could effectively parse through.

Challenges we ran into

Being a team of freshman proves difficult, given that we all had little experience with the software development and deployment process on a team. Working with Github and virtual environments yielded some errors at the start. Furthermore, besides speculating Langchain as the proper approach, the three of us had very little knowledge on how to use it. Lastly, we spent a long time attempting to build this project into an executable and directory, but had trouble in doing so.

Accomplishments that we're proud of

We have nailed the world file functionality that we had hoped for, as it is the sole feature that makes our project innovative and fresh in the D&D community to date. We were able to tell the amount of influence each generation took from the world files, and after testing the software with fantasy worlds of our own, it is safe to say that the core feature of this project works like a charm.

What we learned

Throughout the course of this project, we learned a lot of things. Primarily, we had to learn Qt Creator and PySide6 from the ground up, which made for an interesting team building experiment. Furthermore, Langchain was brand new to us, and experimenting with generative AI modules was really quite fun. We also learned the importance of getting up and walking away every now and then too.

What's next for Project Wildspace

After KnightHacks, we hope to officially deploy this project into various D&D communities. It will take some work with properly building the software, as well as keeping the site up to date with changelogs and instillation instructions, but we truly believe this tool is something that will benefit players across various communities in their worldbuilding adventures.