I created the Money Toolkit because these were tools that I wanted while managing my own personal finances. I was tired of my budgeting app being in one place, planning tools in another, and then still having to break out Excel when I couldn't find some niche thing online.
This project also marks me getting over my own personal intimidations, by (finally) deploying a project that I've built. For this blog feature, I don't know what it will become. I just wanted a place to share things about the project as I move through it. It will be nerdy; sometimes in a programming/techy way, sometimes in a finance way.
Why Finances?
Growing up, we weren't dirt poor but not with a whole lot more, managing my finances was important to me in my early career. Without much guidance other than "don't go broke", I quickly found Dave Ramsey, liked what I heard, and Dave's financial principles have served me well. I know he's not everyone's cup of tea. (That's okay, no one is. And this tool won't be a beacon "no debt" anyway.) But I just wanted more financial tools than what I could find in typical budgeting apps.
This isn't a place for recommendations. The tools I build for this site are really made out of morbid nerdy curiosity.
How you run your finances is your business.
What about the tech?
They techy bits. Here is how I built it.
Tools and Languages
- Next.js
- JavaScript & TypeScript
- Tailwind
- Convex (data backend)
- Clerk (authentication)
Why those?
The site is hosted on Vercel, and uses Convex for a data backend. I've tried several tech stacks, and connected a few databases to Next.js...it always felt like things were unnecessarily difficult. I was hesitant on Convex (sorry, I like my SQL) but it turned out to be the missing piece I really needed.
I wish I had some thought out, evidence-backed technical reason, but...building it with these tools felt good. So, that's what I did.
AI
The hot question right now is: Did I use AI?? Yes.
I wrote the core calculations myself for things like the amortization schedule, but used Codex for a bunch of it. I kept things focused. I wasn't trying to "one-shot" anything, or try out some hot shot agentic whatever. I found success using a planning mode, reviewing the "plan" Codex would produce, and then letting it implement it.
I still found myself needing to be very involved in the process. Breaking down features into tasks, describing acceptance criteria and expected behavior, and reviewing the code instead of blindly merging everything. I was definitely further removed from the code, but it's not all that bad and I find I can still move around the codebase pretty well. There were still a couple times I had to fix things myself, though.
Anyway. My big focus, at this point, was on learning how to get something "done, not perfect" so I could actually ship it. So, AI was very beneficial for this.
Final Thoughts and Monetization
Quickly on monetization. I chose a donation-based approach like Buy Me a Coffee, because as of now the site isn't offering anything that you couldn't find for free with a quick Google search. (Or, vibe code your own Excel spreadsheet) I didn't feel good putting that behind some paywall. I'd like to create a full budgeting app and more features, which will need to be paid features, but that's all for later.
Anyway. If this tool helps, then it did its job. Enjoy!