What?

What is What?

What? is a sample application demonstrating Agentic AI using an MCP server that connects to a REST API, PostgreSQL database, and document storage. It is intended for developers learning to build AI agents that interact with multiple data sources via the MCP protocol.

How to use What?

Install UV and dependencies from pyproject.toml, activate a virtual environment, and populate sample financial transaction data with python util/populate_data.py. Configure the MCP client (Claude for Windows) by placing a modified claude_desktop_config.json in the Claude app data folder, then restart Claude Desktop.

Key features of What?

• Sample application for Agentic AI with MCP
• Connects to REST API, PostgreSQL, Document Storage
• Populates financial transaction test data
• Configured for Claude Desktop as MCP client

Use cases of What?

• Learning how to build an MCP server that aggregates REST, database, and document sources
• Prototyping an AI agent that answers queries using financial transaction data

FAQ from What?

What dependencies does What? require?

Python, UV, a virtual environment, and packages defined in pyproject.toml. It also requires an Alpha Vantage API key and PostgreSQL credentials.

How do I populate sample data?

Run python util/populate_data.py after installing dependencies.

How is the MCP client configured?

Copy the sample config file from config/sample_claude_desktop_config.json to the Claude app data folder, rename it, and update the API key and database connection details.