Captains HUD: AI-Powered Cricket Strategy Simulator

Demo Video Link - https://youtu.be/VudXVtMESMY

Overview

Captains HUD (formerly "Captain Cool") is an interactive, AI-driven web application that simulates high-pressure cricket match scenarios. Users can input a specific match situation (e.g., "19th over, 24 runs to win, heavy dew"), and the application uses Google's Gemini AI to generate strategic decisions, simulate a multi-agent debate, and deliver dynamic live commentary.

Problem Statement

The goal was to build an application that could demonstrate multi-agent AI interactions in a fun, relatable context. Cricket strategy, with its diverse captaincy styles and high-pressure scenarios, provided the perfect backdrop. The application needed to:

• Emulate the distinct personalities of legendary Indian cricket captains: MS Dhoni (calm/methodical), Virat Kohli (aggressive/passionate), Rohit Sharma (tactical/flexible), and Gautam Gambhir (intense/match-up driven).
• Use a multi-agent architecture (Strategist, Devil's Advocate, Commentator, and Stats Analyst) to debate the scenario.
• Present the data in a minimalist, broadcast-style "Sports HUD" with high-end visual interactions (pressure meters, tickers, animations).
• Handle AI rate limits gracefully using mock fallbacks.

Features

• Multi-Agent Simulation: The Gemini API powers distinct agents that analyze the pitch, make decisions, challenge those decisions, and provide live commentary.
• Compare Mode: Run a side-by-side simulation to see how different captains would handle the exact same high-pressure scenario.
• Dynamic Visuals: Features a pulsing pressure meter, an animated news ticker, dynamic CSS "glitch" elements, and a confetti explosion upon the final commentary verdict.
• Resilient Fallbacks: If the API rate limit is hit, the application automatically falls back to a high-fidelity mock simulation, ensuring zero downtime.

Technical Architecture

• Frontend: Pure HTML, Vanilla CSS, and JavaScript (ESM). No heavy frameworks, ensuring a lightweight and lightning-fast HUD.
• AI Integration: Uses the @google/genai SDK to communicate with Gemini models.
• Prompt Engineering: The application uses a "Mega-Prompt" schema, consolidating the debate into a single structured JSON response to minimize latency and API calls.

How to Run

1. Open index.html in any modern web browser (or serve locally using Python).
2. Click the settings icon to input your Google Gemini API Key.
3. Select a captain and enter a match scenario.
4. Click "EXECUTE" and watch the debate unfold!