APL

🏏 Building CaptainCool AI: The Multi-Agent Gemini Cricket Strategist

Imagine sitting in an IPL dugout during the 18th over. The stadium lights are blinding, 40,000 fans are screaming, the ball is slick with heavy dew, and you need to defend 24 runs against a world-class finisher.

Do you trust your death-over pacer, or throw the ball to a spinner because the batter struggles against spin on this venue?

Cricket is a captain’s game.

To bring tactical decision-making like :contentReference[oaicite:0]{index=0} and :contentReference[oaicite:1]{index=1} into a live digital assistant, I built CaptainCool AI during a 3-hour vibe-coding session using :contentReference[oaicite:2]{index=2} Antigravity and :contentReference[oaicite:3]{index=3}.

---

🏗️ Architecture

CaptainCool AI is a multi-agent system where separate agents collaborate before making a final cricket decision.

Workflow

text
User Match Input
        ↓
Stats Analyst Agent
        ↓
Tool Call (Live Match Analysis)
        ↓
Strategist Agent
        ↓
Devil’s Advocate Agent
        ↓
Head Coach Agent
        ↓
Final Captain Decision