MCP Servers

Dual-DB FastMCP — minimal read-only DB tools

Short description

This repository provides a small Python service (server.py) that exposes two read-only FastMCP tools for inspecting PostgreSQL and MySQL databases:

• execute_read_query — run read-only SELECT queries (validated).
• get_database_schema — list tables and columns.

Tech stack

• Docker Compose (optional): Postgres, MySQL, Adminer.
• Python 3.10+ (recommended)
• Node.js/npm (optional, for npx n8n)
• Key Python packages (pinned in requirements.txt): PyMySQL, psycopg2-binary, fastmcp, mcp

Quick start (recommended for local development)

1. Start containerized DBs and UIs (optional):

docker-compose up -d
docker ps

2. Start the Python FastMCP server (terminal A):

python -m venv .venv
(Set-ExecutionPolicy -Scope Process -ExecutionPolicy RemoteSigned) ; (& .venv\Scripts\Activate.ps1)
pip install -r requirements.txt
python server.py

3. Start n8n (terminal B) — n8n editor available at http://localhost:5678:

npx n8n

Notes

• Run each command in its own terminal so services stay active.
• server.py defaults connect to local DBs; override using these env vars: PG_HOST, PG_DATABASE, PG_USER, PG_PASSWORD, PG_PORT, MYSQL_HOST, MYSQL_DATABASE, MYSQL_USER, MYSQL_PASSWORD, MYSQL_PORT.
• The execute_read_query tool only allows SELECT queries; non-SELECT queries are rejected.

Environment file

• Copy .env.example to .env and fill in real passwords for local development. .env is ignored by .gitignore and should not be committed.

Export n8n workflows and push to GitHub

If the endpoint path differs in your FastMCP deployment, adapt the URL accordingly.

**n8n (optional)**

- Access the local n8n UI at http://localhost:5678 (if you started the n8n container).

- Exporting a workflow (manual UI method):
  1. Open the workflow in the n8n editor UI.
  2. Click the workflow menu (three dots) and choose **Export** → **Download JSON**.
  3. Save the downloaded JSON into this repository under a folder such as `n8n-workflows/` (create it if needed).