League of Legends Assist!

Our Logo!

Inspiration

So basically, we were playing League of Legends and we did not know too much about the game. Then me and my friends we had a brilliant idea!!!! we should make a OMI app for league of legends players. We knew that this was very well needed in the OMI community as many people who use AI are also League of Legends players!!!

What it does

So basically, while playing a game of League you can ask your OMI questions about the different champions in the game and it will answer your questions through TTS. This is very epic and cool as you can spend maximum time playing League without alt tabbing and searching up stuff about the champions.

How we built it

We built this app using python to implement a Generative Model which depends on RAG (Retrieval Augmented Generation) to specifically get the answer that the user wants about the champions in the game. The way we implement a RAG-based system was through Web scraping the League of Legends WIKI. Once we got all the champion data we used Generative AI to create answers based on our specific champion data.

Challenges we ran into

So the first challenge we ran through was being poor. We did not have the necessary funds to acquire GPT-4 tokens to use in our model. We decided to Gemini and then since we could not properly download "faiss" to use for an actual RAG, we used a pseudo RAG that we made (basically webscraping and splitting the txt files into catagores). We also could not find out how to use Vercel so we used the actual webhook tester API.