Rise to Challenger — AI-powered LoL coaching app

Show the Community!
1

Hi Streamlit community! Just shipped my first AI product
and wanted to share it here :video_game:

Rise to Challenger is a League of Legends coaching app
that tells players what to fix after each game — not just
shows stats, but gives structured AI feedback benchmarked
against Challenger-level players.

:link: Live demo: https://rise-to-challenger-m9ykf3u3jo5lhyzcvljcbu.streamlit.app/
:package: GitHub: GitHub - Bubble0421/rise-to-challenger: AI-powered League of Legends coaching app with match review, counter guide, and agent-based analysis · GitHub

How it works:

  • Pulls real match data from Riot Games API + Timeline API
  • Benchmarks every stat vs Challenger average for your exact champion + role
  • Multi-agent LangGraph pipeline: Comp → Execution → RAG → Reflection → Synthesis
  • Local LLM via Ollama (Gemma 2B) — private, no API costs
  • RAG with ChromaDB over 900+ Master+ match patterns

Three pages:

  1. Meta Analysis — champion tier list from Master+ data
  2. Player Review — post-game AI coaching report with training goals
  3. Counter Guide — 30-second pre-game matchup plan

This is my first AI product, built as a capstone project
Happy to answer any questions about the architecture or implementation!

game_context
game_context2914×1458 499 KB

2

This is actually a really cool project idea. Combining AI with League coaching through Streamlit makes the app feel interactive and beginner-friendly at the same time. I especially like how gaming and AI are being merged to give players personalized feedback instead of generic tips. It would be interesting to see future features like match history analysis, champion-specific recommendations, or live draft suggestions as well. Great work sharing this with the community!