MCP (Model Context Protocol) Server

Project Structure

├── client.py             # Client-side interaction script
├── server.py             # Main MCP server implementation
├── pg_connect.py         # PostgreSQL database connection
├── lm_config.py          # Language model configuration
│
├── .env.example          # Example environment configuration
├── .env.dev              # Development environment configuration
├── requirements.txt      # Project dependencies
└── .gitignore            # Git ignore file

Prerequisites

• Python 3.10+
• PostgreSQL
• API access to AI providers (Anthropic, Google)

Installation

1. Clone the Repository

git https://github.com/VajraM-dev/Postgres-MCP-Server-With-SSE-Transport.git

2. Create Virtual Environment

python -m venv venv
source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install Dependencies

pip install -r requirements.txt

4. Configure Environment

1. Copy .env.example to .env.dev
2. Fill in the required configuration:

cp .env.example .env.dev
nano .env.dev  # or use your preferred text editor

Configuration Parameters

• POSTGRES_USERNAME: PostgreSQL database username
• POSTGRES_PASSWORD: PostgreSQL database password
• POSTGRES_DB_NAME: Database name
• POSTGRES_HOST: Database host
• POSTGRES_PORT: Database port
• MCP_NAME: Server name
• MCP_HOST: Server host
• MCP_PORT: Server port
• TRANSPORT: Communication transport (sse/stdio)
• ANTHROPIC_API_KEY: Anthropic API key
• GOOGLE_API_KEY: Google API key
• USE_PROVIDER: Default AI provider

Running the Server

Development Mode

python server.py

Client Interaction

python client.py

Key Features

• 🔒 Secure configuration management
• 🗃️ PostgreSQL database integration
• 🤖 Multi-provider AI model support
• 📡 Flexible communication transport
• 🛡️ Extensible tool registration

Supported AI Providers

• Anthropic (Claude models)
• Google (Gemini models)

Tools and Endpoints

Available Tools

• list_tables(): Retrieve database tables
• Custom tools can be easily added via decorators

Endpoints

• /sse: Server-Sent Events endpoint
• Customizable routing and tool registration

Extending the Framework

Adding New Tools

@app.tool()
def custom_tool():
    """Custom tool implementation"""
    # Your tool logic here

Configuring AI Providers

Modify lm_config.py to add or configure new AI providers.