Montouir
Nov 27, 2025
Experimental Technologist
What's The Situation?
As a long winter break loomed, I thought of a number of useful tools that I could build - namely in support of my hobby building computers and conducting hardware monitoring. I wanted to create a tool that would help me analyze the cost of a system's power usage with real-time analytics data.
At the same time, I had been hearing the constant online battles of AI - was it a danger to careers? Was it a pseudo-friend? Or was it really what all technology becomes - a tool to be learned and used to achieve a goal? To answer these questions and to create my tool, I turned to Claude Code, having learned a decent bit of Python and Javascript earlier on in my academic career. With my combined experience and intrigue, I set out to build my little power monitoring experiment - Montouir.
Taskmaster
To accomplish Montouir's construction, I set out with a few pieces of info in mind:
1. This was a learning experience to construct a personal tool - not a commercial project meant to replace hiring proper development support.
I wanted to experiment with connective tech while sticking a small core set of features. Proper electricity cost monitoring, system utilization, power trends over time, and logging data to be exported and analyzed as I needed. At the same time, I wanted to export this data to a cloud service, enabling remote system monitoring without a large on-machine power draw, therefore skewing numbers more drastically than needed.
This project required an extensible codebase from which I could add features and utilities as I saw fit. That ended up including custom-length CSV export, a drag-n-drop dashboard for enhanced customization, and a live-updating tray icon to indicate monitoring status.
As I reached the end of initial ideation, I felt that the app should be a standalone application rather than a simple web app - that way I could focus on optimization of application performance/overhead versus everchanging web conditions.
Next Steps & Implementation
I then set about implementing these features using Claude, starting with my monitoring dashboard. I chose to go the React route based on availability of troubleshooting information, with Electron enabling the desktop app deployment and Python for the backend due to its straightforward syntax and data handling simplicity.
I eventually ending up with a workable MVP, though it needed refinement.
What Have We Created?
By the end of the process, I had created a standalone desktop app that solved my cost-calculating power monitoring problem. I'd persevered through feature implementations, long documentation reading sessions, late-night troubleshooting, and build errors/packaging issues galore. I had learned so much about how modern web apps can actually be constructed, how to leverage AI in a healthy and self-contained manner, and problem-solved using a combined set of skills augmented by emerging technologies. I'm planning an open source alpha release of the app in the coming weeks for the OSS community that might find the tool interesting to poke at.
