MCP Database Server

A Model Context Protocol (MCP) server that provides PostgreSQL database access with query execution and schema inspection capabilities.

Overview

This MCP server enables AI assistants (like Claude) to interact with a PostgreSQL database containing data from multiple systems:

• cs (CustomerScoreView): Customer and credit assessment data

• dms (Data Management System): Document management data

• los (Loan Management System): Loan and payment data

• mls (Mahatheun Leasing System): Contract and asset leasing data

Features

• ✅ Query Execution: Execute SQL SELECT, INSERT, UPDATE, DELETE queries

• ✅ Schema Inspection: List schemas, tables, and column details

• ✅ Multi-Schema Support: Organize data by system (cs, dms, los, mls)

• ✅ Docker Setup: PostgreSQL in Docker with automatic initialization

• ✅ CSV Import: Sample data included and ready to import

Prerequisites

• Python 3.10 or higher

• Docker and Docker Compose

• pip (Python package installer)

Installation

1. Clone or navigate to the project directory:

   cd c:\Users\chaya\project\mcp-database

2. Install Python dependencies:

   pip install -r requirements.txt

3. Configure environment variables:

   copy .env.example .env

   Edit .env if you need to change database credentials (optional).

Setup

1. Start PostgreSQL Database

Start the PostgreSQL container with Docker Compose:

Verify the database is running:

2. Verify Database Initialization

The database will automatically initialize with schemas and sample data. Check the schemas:

List tables in a schema:

3. (Optional) Import CSV Data

Sample CSV files are provided in raw_data/. To import them into the database:

Import a specific CSV file:

Running the MCP Server

Test Locally

Run the server directly to test:

The server will start and listen for MCP messages via stdio.

Configure with Claude Desktop

To use this MCP server with Claude Desktop, add it to your Claude configuration file:

Windows: %APPDATA%\Claude\claude_desktop_config.json

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Add this configuration:

Restart Claude Desktop to load the server.

Available MCP Tools

1. list_schemas

Lists all available database schemas.

Example:

2. list_tables

Lists all tables, optionally filtered by schema.

Parameters:

• schema (optional): Schema name to filter (e.g., "cs", "dms", "los", "mls")

Example:

3. describe_table

Get detailed column information for a specific table.

Parameters:

• table_name (required): Name of the table

• schema (optional): Schema name (default: "public")

Example:

4. execute_query

Execute a SQL query on the database.

Parameters:

• query (required): SQL query string

• params (optional): Array of parameters for parameterized queries

Examples:

Database Schema

cs (CustomerScoreView)

• customers: Customer information and credit scores

• credit_assessments: Credit assessment history

dms (Data Management System)

• documents: Document metadata and storage info

• document_versions: Document version history

los (Loan Management System)

• loans: Loan accounts and terms

• payments: Payment history

mls (Mahatheun Leasing System)

• contracts: Lease contract information

• assets: Asset details for leased items

• lease_payments: Lease payment records

Project Structure

Troubleshooting

Docker container won't start

Connection refused error

• Ensure PostgreSQL is running: docker-compose ps

• Check port 5432 is not in use by another process

• Verify .env file has correct credentials

Import errors in Python

Development

To extend the server:

1. Add new tools in src/tools.py

2. Register handlers in src/server.py

3. Update database schema in db/init.sql

4. Add sample data in raw_data/

License

This project is provided as-is for database integration purposes.