Riftforged

Inspiration

I wanted to analyze match data, player habits, and trends. My inspiration was my own league journey and what I have learned over the years. How I could influence AI analysis with coach-recommended tips to make a personalized coach.

What it does

The app shows game overviews like most common sites, with the option to view a yearly rewind that analyzes performance, habits, and suggests personalized tips to help players improve their gameplay. These rewinds can be customized into showcase cards, which can be shared with friends or downloaded.

How we built it

I used the T3 stack, more specifically I have used Next.js, Tailwind, Typescript, tRPC, AWS Bedrock, and SQLite.

Challenges we ran into

The challenge that made me rethink many things were the low rate limits for the standard keys Riot provides. This challenge had made me learn new ways of fetching data with strict rate limits.

Accomplishments that we're proud of

This is my first project that has ever used generative AI as a feature.

What we learned

The many complications using Gen AI, good system prompting, and how to manage and parse big data to make a product that looks good and works.

What's next for Riftforged

I may continue working on Riftforged, or using the idea of an AI coach and what I have learned with another project.

Built With

• amazon-web-services
• aws-bedrock
• nextjs
• prisma
• sqlite
• t3
• tailwind
• trpc
• typescript