SQL MCP Server

🚀 SQLGenius - Intelligent SQL Assistant

SQLGenius is an intelligent SQL assistant that enables you to query BigQuery databases using natural language. It is built on the Model Context Protocol (MCP), Vertex AI's Gemini Pro, and Streamlit.

🚀 Quick Start

Installation

1. Clone the repository and navigate to the project directory:

cd sql_mcp_server

2. Install the dependencies:

pip install -r requirements.txt

3. Copy the .env.example file to .env and fill in the configuration information:

cp .env.example .env

4. Set your environment variables in .env:

PROJECT_ID=your-project-id
DATASET_ID=your-dataset-id
GOOGLE_APPLICATION_CREDENTIALS=path/to/your/service-account.json
VERTEX_AI_LOCATION=us-central1

Usage

1. Start the application:

streamlit run streamlit_app.py

2. When the Streamlit application starts, the MCP server will run automatically.
3. Use the tabs to perform the following actions:
   • Ask questions about the data in plain English.
   • Write and execute SQL queries directly.
   • Explore the database schema.

✨ Features

• Convert natural language to SQL using Gemini Pro.
• Interactive Streamlit interface with multiple tabs.
• Real-time query execution and visualization.
• Database schema exploration.
• Query history tracking.
• Secure query validation.
• BigQuery integration.
• Based on the MCP architecture.

🎥 Demo

Watch SQLGenius in action! Here is a quick demo video showing how to use the application:

In this demo, you can see:

1. Conversion of natural language queries to SQL.
2. Interactive data analysis visualization.
3. Database schema exploration.
4. Query history tracking.

💻 Usage Examples

Basic Usage

💬 Natural Language Queries

Ask questions in plain English and get SQL results:

• "Show the top 5 customers by revenue."
• "Which products had the highest sales in January?"
• "How many orders were placed last month?"

📊 SQL Queries

Write and execute SQL queries directly:

SELECT * FROM orders 
WHERE order_date > '2023-01-01' 
ORDER BY total_amount DESC
LIMIT 10

📋 Database Explorer

• Browse all available tables.
• View table schema information.
• View sample data for any table.

Advanced Usage

To add custom tools to the MCP server:

1. Edit the register_tools() method in sql_mcp_server.py.
2. Add your custom tools using the @self.tool() decorator.
3. Restart the server.

🔧 Technical Details

SQLGenius uses the Model Context Protocol (MCP) to expose tools, which include:

1. Natural Language Processing: Convert English questions to SQL.
2. Data Exploration: Get schema information and sample data.
3. SQL Execution: Run validated queries on the database.

The architecture consists of the following components:

• MCP Server: Handles database connections and provides tools.
• Streamlit Frontend Interface: The interface for users to interact with the system.
• Vertex AI (Gemini Pro): Provides natural language understanding capabilities.
• BigQuery: Executes SQL queries on your data.

📝 MCP Tools

The following MCP tools are available:

1. execute_nl_query: Execute natural language queries.
2. execute_sql_query: Execute raw SQL queries.
3. list_tables: List all available tables.
4. get_table_schema: Get the schema of a specific table.

🔒 Security Features

• Only allow SELECT queries to be executed.
• Query validation to prevent dangerous operations.
• Secure credential management.
• Error handling and input validation.

🤝 Contribution

Contributions are welcome! Please feel free to submit pull requests.