Login Page
Dashboard

Agnarr — League of Legends Match Analysis Assistant

What Is Agnarr?

Agnarr is a desktop assistant for League of Legends players, focused on delivering actionable, matchup-aware coaching and post-game analysis. It helps players understand their performance and improve with targeted feedback based on completed matches.

What It Does

Agnarr provides:

Post-match analysis and coaching:
- After each match, Agnarr analyzes your performance and provides actionable advice and feedback.
- All feedback is based on completed matches and historical stats, not live in-game data.
Matchup analysis:
- Shows a side-by-side champion matchup header with splash art and names.
- Gives early/mid/late game plans, itemization, lane strategy, roam/gank triggers, and combat patterns tailored to your champion vs. opponent.
Stat comparison & analytics:
- Extracts your match stats (K/D/A, CS, damage, vision, gold, item/rune/summoner spell choices) and compares them to champion-vs-opponent averages.
- Highlights where you outperformed or lagged behind, with color-coded stat cards and difference indicators.
Structured feedback:
- Summarizes what went right/wrong and gives prioritized steps for improvement.
- Feedback includes “Match Review”, “Phase Improvements” (early/mid/late), and “Match Plan & Analytics”.
Data integration:
- Uses Riot API and DDRAGON assets for match data (champions, items, spells).
- Maps Riot/DDRAGON IDs to readable assets and names.
Secure desktop app:
- Built with React + Vite (UI) and Tauri (Rust backend) for fast, secure streaming.
- Keeps API keys and sensitive data out of the browser context.
Streaming AI output:
- Rust backend invokes Gemini 2.0 Flash, streams chat chunks via Tauri events.
- React UI accumulates, sanitizes, and renders Markdown in real time.

Technical Details

Frontend: React 19 + Vite 7; Markdown rendering via marked + DOMPurify; responsive split-panel UI.
Backend: Tauri 2 + Rust; Gemini 2.0 Flash via genai client; Tauri event bus for chunk delivery.
Data: Per-match stat extraction and champion matchup averages; DDRAGON assets for champions/items/spells.

Challenges we ran into

We had to scrape 1000s of data of different summoners from plat+ elo and this took some time
Understanding how to get relevant data for our data analysis efficiently/

Accomplishments that we're proud of

We were able to develop a complete app that we imagined to do
Learnt rust basics in a month We also learnt a new framework to make cross-platform applications