MCP Server with PostgreSQL, uV Python, and AutoGen AI Overview This project is a modular backend server built with uV Python, powered by PostgreSQL for data storage and integrated with AutoGen AI for intelligent data processing and autonomous task execution. The core module, MCP (Modular Command Processor), handles structured interactions between the database and AI agents, supporting real-time operations, import validation, and AI-augmented decision logic.

Key Components uV Python: Lightweight, isolated Python environment used for fast, reproducible deployments.

PostgreSQL: Robust relational database for structured data storage and querying.

MCP (Modular Command Processor): Core logic engine to orchestrate data imports, validation, and transformations using schema-driven rules.

AutoGen AI: Integrated to augment operations with AI capabilities such as natural language query processing, autonomous task execution, and intelligent agent support.

Features Schema-based data validation and transformation

AI-assisted natural language interaction with the database

Streamlined setup using pyproject.toml and uv

Clean modular structure for extensibility and maintainability

Getting Started Prerequisites Python 3.10+

uV (installed via pip: pip install uv)

PostgreSQL running locally or remotely

Setup Instructions

Clone the repository

git clone https://github.com/shrishailwali/mcp-server-postgres

Install dependencies using uV

uv pip install -r requirements.txt # or use pyproject.toml directly Environment Configuration Create a .env file with the following keys:

env

POSTGRES_HOST=localhost POSTGRES_PORT=5432 POSTGRES_USER=postgres POSTGRES_DB=cmms POSTGRES_PASSWORD=yourpassword

Run the Server uvicorn main:app --reload

Integration with AutoGen AI AutoGen is used to:

Parse user input (e.g., file uploads, commands)

Generate SQL queries via AI agents

Automate backend workflows with LLM-powered reasoning

AutoGen integration uses your Azure OpenAI deployment named spick-ai to generate responses and perform actions.