Homelab

Building and Hardening My Spend Tracker

I built a personal spend tracker to answer one question quickly: where is my money actually going each month? The app ingests ASB statements, categorises transactions, tracks budgets, and runs a local AI assistant for summaries and cutback advice.

Over time it stopped being just a finance tool and became a proper systems project. Once I decided to put it on the public internet, the focus shifted from “does it work?” to “can I trust it when I am not watching it?”

How I Self-Hosted This Blog on a Mini PC

I got tired of the free tier anxiety. Vercel and Netlify are solid until they change their pricing or disappear, and I did not want a recurring bill for a static site. I wanted a machine sitting in my house actually serving the files.

The machine

The site runs on a generic mini PC — an N100 processor, 16GB of RAM, and an SSD. Mini PCs like this are worth knowing about if you are doing homelab work on a budget. They pull around 6W at idle, run silently, and handle a static site without any trouble. Overkill for this use case, but that just means it will never struggle.

Getting PlainSight Off the Ground

This is my capstone project for Unitec, built in partnership with Netbridge. The idea is simple: instead of logging into three different vendor portals to check what is happening on the network, everything feeds into one dashboard that uses a local AI model to explain what it all means in plain English.

What it actually is

PlainSight is three things working together. A Wazuh SIEM backend sitting on Ubuntu via VMware ESXi that ingests syslog from the FortiGate 60E, the Ruckus R650 WAP, and the FS-148F switch. A Streamlit dashboard that talks to Wazuh via its REST API and surfaces everything in one view. And a local phi3:mini model running through Ollama that reads the alerts and writes a plain English summary with MITRE ATT&CK tags attached.