Skip to content

uddhav05-cyber/captain-cool-agentic-premier-league

Repository files navigation

🏏 Captain Cool — Multi-Agent IPL Match Strategist

Built on Google Gemini | Google AI Studio Hackathon 2026

Captain Cool Banner

Captain Cool is an agentic AI system that thinks like an IPL captain — making real-time tactical decisions the way Dhoni, Rohit, or Hardik would. Powered entirely by the Google Gemini stack with 4 named agents that genuinely debate before committing to a call.


✨ What Makes It Different

  • 🎯 4 named Gemini agents that genuinely debate before committing to a call — not a single prompt wearing 4 hats
  • 🧊 Captain Persona Selector — pick Dhoni (calm & calculated), Rohit (aggressive), or Hardik (instinctive)
  • 🌐 Live match scraping via Gemini's URL context tool — paste a Cricbuzz URL, no manual input needed
  • ⚔️ Real agent debate shown in UI — Strategist proposes → Devil's Advocate challenges → Strategist defends
  • 📊 Counterfactual analysis — "If you'd bowled X instead, win probability drops ~11%"
  • 🎙️ Voice commentary — Commentator speaks the final call aloud via Web Speech API
  • 🔴 Confidence score — visual badge per decision (green >7, yellow 5–7, red <5)

🤖 Agent Architecture

User Input (Cricbuzz URL or manual match state)
                    ↓
             Orchestrator
  ┌───────────────────────────────────────┐
  │ Agent 1: Stats Analyst                │ ← Google Search tool call (live player stats)
  │ Agent 2: Strategist (persona-driven)  │ ← Proposes tactical decision
  │ Agent 3: Devil's Advocate             │ ← Challenges the call hard
  │ Agent 4: Commentator                  │ ← Harsha Bhogle-style output + counterfactual
  └───────────────────────────────────────┘
                    ↓
       Final Decision + Debate UI + Voice Output

Agent System Prompts

Agent Role System Prompt Summary
Stats Analyst Data & Tool Use Analyzes pitch, dew, bowler economy, batter weaknesses. Uses Google Search to fetch live player form. Outputs structured JSON with win probability.
Strategist (Dhoni) Decision Maker "Ice in your veins. Back your bowlers, never panic, always think 2 overs ahead. Make ONE decisive call."
Strategist (Rohit) Decision Maker "Bold and aggressive. Back your instincts and your big hitters. Make the attacking call others won't."
Strategist (Hardik) Decision Maker "Unconventional and fearless. Make the call that surprises everyone but works."
Devil's Advocate Challenger "Challenge the Strategist's decision hard. Find the flaw. Propose a specific alternative. Be direct."
Commentator Final Output Harsha Bhogle-style explanation with counterfactual and confidence score out of 10.

🔄 Orchestrator Flow

1. Extract match state from Cricbuzz URL (Gemini URL context) or manual form
2. Stats Analyst → structured JSON with win probability
3. Strategist (selected persona) → proposed tactical decision
4. Devil's Advocate → challenge + alternative
5. Strategist → final defended decision
6. Commentator → cricket-language output + counterfactual + confidence score + voice

🎮 One Full Match Scenario

Input:

  • MI vs CSK, 2nd innings, over 16, CSK need 34 off 24, 4 wickets down
  • Rohit Sharma on strike (67*), Jadeja non-striker (12*)
  • Bumrah has 2 overs left, Chahal has 1, flat pitch, heavy dew, Wankhede

Stats Analyst: Win prob MI 58% / CSK 42%. Dew negating seam movement. Jadeja vulnerable to pace outside off.

Strategist (Dhoni mode): Bowl Bumrah now — back him to nail yorkers despite dew.

Devil's Advocate: Dew renders Bumrah's seam useless. Save him for over 19. Bowl Chahal while Jadeja is on strike.

Strategist (final): Conceded the point. Chahal to bowl over 17.

Commentator: "In dew-heavy Wankhede, Bumrah's famous seam is rendered ineffective. Rohit holds him back for the 19th — smart, calculated captaincy. Confidence: 8/10. If they'd bowled Bumrah here, win probability drops to 51%." 🎙️


🛠 Tech Stack

Layer Technology
Frontend TypeScript + Vite + CSS
AI Model Google Gemini 2.5 Flash
SDK @google/genai
Tool Use Gemini built-in Google Search
Voice Output Web Speech API
Build Environment Google AI Studio

🚀 Run Locally

Prerequisites: Node.js

# 1. Clone the repo
git clone https://github.com/uddhav05-cyber/Captain-Cool-tactic-decisiodecision-
cd Captain-Cool-tactic-decisiodecision-

# 2. Install dependencies
npm install

# 3. Add your Gemini API key
cp .env.example .env.local
# Edit .env.local → set GEMINI_API_KEY=your_key_here

# 4. Run
npm run dev

Get your free Gemini API key at aistudio.google.com


📁 Project Structure

/src
  /agents        ← 4 agent definitions + system prompts
  /components    ← AgentDebate, CaptainSelector, MatchInput, FinalDecision
  /lib           ← Gemini client, orchestrator logic
index.html
server.ts
vite.config.ts
.env.example

🔗 Links

  • 🤖 View in AI Studio
  • 📝 Dev.to Blog ← add after publishing
  • 👤 Built by Uddhav Bhople — Microsoft Learn Student Ambassador | Google Student Ambassador | B.Tech AI/ML @ D.Y. Patil University Pune

Cricket is a captain's game. Built on Gemini. 🏆

About

Multi-agent strategy advisor for the Agentic Premier League powered by Gemini 2.5 Flash

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors