MCPlucker 🦆

A comprehensive Model Context Protocol (MCP) server for DuckDB, providing advanced data loading, querying, and analysis capabilities.

Quick Start

Prerequisites

• Go 1.24 or later

Installation

git clone https://github.com/dacort/mcplucker.git
cd mcplucker
go build -o mcplucker ./main.go

Usage

MCPlucker runs as an MCP server using stdio transport

Adding to Claude Code:

1. Build the project:

go build -o mcplucker ./main.go

2. Add the MCP server to Claude Code:

claude mcp add duckdb ./mcplucker

3. Start using MCPlucker in your Claude Code sessions!

Available Tools

Core Data Management

• create_table_from_file - Create tables from CSV, JSON, or Parquet files
• list_tables - List all tables and views in the database
• describe_table - Get detailed schema information and sample data
• execute_query - Execute SQL queries with configurable limits

Example Usage

Once connected to an MCP client (like Claude Desktop), you can interact naturally:

Loading data:

"Hello, my friend, can you create a duckdb database from test_data/products.json?"

Exploring tables:

"Can you give me a quick summary of the products table?"

⏺ Products Table Summary:
  - 5 total products across 2 categories (Electronics, Furniture)
  - Price range: $25.50 - $999.99 (avg: $320.09)
  - Stock status: 4 in stock, 1 out of stock
  - Categories: Electronics (4 items), Furniture (1 item)

Running queries:

"Show me all electronics products that are in stock"

⏺ Electronics Products In Stock:

  | ID  | Name    | Price   |
  |-----|---------|---------|
  | 1   | Laptop  | $999.99 |
  | 2   | Mouse   | $25.50  |
  | 4   | Monitor | $299.99 |

  3 electronics products are currently in stock.

Technical usage via MCP protocol:

{
  "tool": "create_table_from_file",
  "arguments": {
    "table_name": "products", 
    "file_path": "test_data/products.json"
  }
}

Features

✅ File Format Support: CSV, JSON, Parquet
✅ DuckDB Integration: Native DuckDB performance and SQL compatibility
✅ Type Safety: Structured input/output with comprehensive validation
✅ Error Handling: Clear error messages and validation
✅ MCP Compliant: Full Model Context Protocol compatibility

Development Status

• ✅ Phase 1: Project structure and basic MCP server
• ✅ Phase 2: MVP functionality (file loading, querying)
• 🚧 Phase 3: Cloud connectivity (S3, Azure, Iceberg)
• 📋 Phase 4: Analysis functionality (query optimization, profiling)
• 📋 Phase 5: Database connectivity (PostgreSQL, MySQL, SQLite)

Contributing

See CLAUDE.md for detailed development guidelines and build instructions.

License

MIT License