Archon Scout

Home
select team
team event data
gemini3 thinking
scout report common strategies
scout report players and setup
scout report win strategies

Inspiration

Inspired by modern AI decisioning platform and thinking model design

What it does

Leveraging Gemini 3 models capabilities like reasoning thinking, massive long context and support multimedia inputs, we build Archon to enable AI analysis based on massive amount of competition, teams, players and events data for League of Legends esports. Archon Coach analyzes your completed matches using Google's Gemini 3 AI model. It processes match event data to provide personalized insights, macro review, what-if analysis, and individual player performance breakdowns. Coach mode also has a match replay function that is very helpful for coach review Archon Scout analyzes your opponent's recent matches to generate comprehensive scouting reports. Know your enemy before the game starts with AI-powered insights into their strategies, tendencies, and weaknesses.

How we built it

Gemini3 thinking Models, Nextjs, Vercel AI-SDK/AI Elements, JetBrains IDEs and Junie (Anthropic API key), Neon DB etc, Vercel Blob Storage, GraphQL

Challenges we ran into

Work out the event data schema and data points that has implications on match progress and result. We ask AI to help us understand and find the patterns and links between data points.

Accomplishments that we're proud of

We build a cool mini match replay view based on event data. Also optimised how we store/cache event data for matches, and create technical instruction for AI to understand how to read the event data to produce meaningful and accurate analysis

What we learned

Gemini 3 thinking model is very powerful if provided with the correct instruction and resources like unstructured data or json feed.

What's next for Archon

After this hackathon, we would like to explore more about Grid datasets for AI deep learning, adding more datasets to enrich the resources AI analysis is based on. Add support for Valorant