MCP Project

Checkpoint #1: FastAPI + PostgreSQL Setup

This repository marks the first checkpoint of my MCP project. The goal was to get a basic FastAPI app running with a PostgreSQL database, containerized with Docker, and ready to expose through the Model Context Protocol (MCP).

What I accomplished

• Environment setup

  • Installed and configured WSL on Windows.
  • Installed Docker and Docker Compose.
  • Created a Python virtual environment to manage dependencies.

• Backend with FastAPI

  • Built a simple API with the following endpoints:
    • POST /create_appointment → create a new appointment
    • GET /list_appointments → list all appointments
    • PUT /update_appointment/{id} → update an existing appointment
    • DELETE /delete_appointment/{id} → delete an appointment
  • Defined the Appointment model in SQLAlchemy.

• Database with PostgreSQL

  • Set up a Postgres container with Docker Compose.
  • Connected FastAPI to the database and verified table creation.
  • Tested inserts and queries through both the API and psql.

• MCP integration

  • Installed and configured mcp-openapi-proxy.
  • Exposed the API endpoints so they can be used as tools through Claude.

• Version control

  • Initialized the Git repository.
  • Configured Git username/email and SSH key.
  • Made the first commit as a checkpoint.

Current status

• API runs locally on http://localhost:8000
• Database connected and working through Docker
• MCP server loads the API schema correctly
• Repo initialized and linked to GitHub

Next steps

• Improve date formatting (e.g. 18/9/2025 - 8:55pm instead of ISO strings).
• Add automated tests with pytest.
• Create a simple frontend to interact with the API.
• Keep expanding the Claude integration through MCP.

✍️ Checkpoint #1 complete.