IPL MCP Server

A Model Context Protocol (MCP) server that provides natural language access to IPL (Indian Premier League) cricket match data. Built using data from Cricsheet with an enhanced sample of 18 IPL matches including Virat Kohli games and CSK vs MI classics.

🏏 Features

• Natural Language Queries: Ask questions about IPL data in plain English

• Enhanced Dataset: 18 carefully selected IPL matches including:

  • Virat Kohli batting performances (99 runs in 4 matches)

  • CSK vs MI classic encounters (3 matches)

  • All major IPL teams represented

• Rich Analytics: Player stats, team performance, match analysis

• Claude Desktop Integration: Works seamlessly with Claude Desktop

• Fast SQL Backend: Efficient SQLite database with optimized queries

• Extensible: Can easily be extended to work with the full 1,169+ match dataset

🚀 Quick Start

Prerequisites

• Python 3.11+

• uv package manager

• Claude Desktop (for MCP integration)

Installation

1. Clone and setup:

2. Install dependencies:

3. Setup database and load data:

This will:

• Create SQLite database tables

• Process 18 sample JSON match files (includes V Kohli & CSK vs MI)

• Calculate player and team statistics

• Takes ~10-15 seconds to complete

4. Test the queries (optional):

5. Start the MCP server:

🎯 Example Queries

Basic Match Information

• "Show me all matches in the dataset"

• "How many matches are in the database?"

• "Which team won the most matches?"

• "What was the highest total score?"

• "Show matches played in Mumbai"

Player Performance

• "Who scored the most runs across all matches?"

• "Which bowler took the most wickets?"

• "Show me Virat Kohli's batting stats"

• "Who has the best bowling figures in a single match?"

• "Show all centuries scored"

Advanced Analytics

• "What's the average first innings score?"

• "Which venue has the highest scoring matches?"

• "What's the most successful chase target?"

• "Which team has the best powerplay performance?"

• "Show me partnership records over 100 runs"

Match-Specific Queries

• "Show me the scorecard for match between CSK and MI"

• "How many sixes were hit in the final?"

• "What was the winning margin in the closest match?"

🔧 Claude Desktop Integration

1. Add to Claude Desktop config:

Edit your Claude Desktop MCP configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%/Claude/claude_desktop_config.json

2. Restart Claude Desktop

3. Test the connection: Ask Claude: "Show me IPL team statistics"

📊 Database Schema

The server uses SQLite with the following key tables:

• matches: Match metadata (teams, venue, date, outcome)

• innings: Innings-level data (totals, wickets, overs)

• deliveries: Ball-by-ball data (runs, wickets, extras)

• player_stats: Aggregated batting/bowling statistics

• team_stats: Team performance metrics

• players: Player registry with Cricsheet IDs

• teams: Team information

🛠️ Advanced Usage

Command Line Options

API Integration

The server can be extended to work with other MCP clients beyond Claude Desktop. The query engine supports pattern matching for natural language understanding.

Adding Custom Queries

Extend the QueryEngine class in src/mcp_server/query_engine.py:

📈 Performance

• Database Size: ~3MB for 18 sample matches

• Setup Time: 10-15 seconds for data load

• Query Response: <1 second for most queries

• Memory Usage: ~50MB typical runtime

🚀 Scaling to Full Dataset

The system can easily handle the complete 1,169 match dataset:

• Full Database Size: ~50MB

• Full Setup Time: 2-3 minutes

• Simply use --data-dir data instead of --data-dir data_small

🔍 Sample Query Results

Query: "Which team won the most matches?"

Query: "Show me Virat Kohli batting stats"

🗄️ Data Source

All data comes from Cricsheet, which provides:

• Ball-by-ball data for IPL matches from 2008-2017 seasons (enhanced sample of 18 matches)

• Player registry with unique identifiers

• Match metadata including officials, venues, outcomes

• JSON format with comprehensive match details

• Full dataset available: 1,169+ matches (2008-2024) can be loaded by using --data-dir data

🤝 Contributing

1. Fork the repository

2. Create a feature branch

3. Add your improvements

4. Test with sample queries

5. Submit a pull request

📝 License

This project is licensed under the MIT License. Data provided by Cricsheet under their terms of use.

🚀 Working with Full Dataset

To use the complete 1,169 match dataset instead of the sample:

1. Reset and load full data:

⚠️ This will take 2-3 minutes to complete

2. Benefits of full dataset:

• Complete IPL history (2008-2024)

• More accurate player statistics

• Comprehensive team performance data

• Better trend analysis capabilities

✅ Verify Installation

Test your setup with these commands:

🔗 Links

• Cricsheet - Data source

• Claude Desktop - MCP client

• MCP Protocol - Protocol specification

Built with ❤️ for cricket analytics and AI-powered data exploration