Last year I sat down and added up what I was paying monthly for developer productivity tools. The number was embarrassing — not individually, but in aggregate. A focus timer here, an API dashboard there, a nutrition tracker because I kept forgetting to eat while shipping code. Around $35/month for tools that each did one thing.
So I built my own.
The Problem With SaaS for Simple Problems
Don't get me wrong — SaaS is great for complex products. I'm not about to build my own Figma. But some tools solve problems so narrow that a subscription feels like paying rent on a closet.
Take API cost tracking. I use multiple LLM providers daily — OpenAI, Anthropic, Google. Each has its own dashboard, its own billing page, its own way of showing token usage. I was either constantly switching tabs or just... not checking. Then getting surprised by the bill.
Or focus apps. I tried three different ones. They all wanted to block entire applications. But I don't need to block Safari — I need to block the Twitter feed inside Safari. The app isn't the problem. The feed is.
What I Built Instead
1. A menu bar token counter
Instead of checking dashboards, I built TokenBar — a tiny macOS menu bar app that shows real-time token usage across providers. One glance, no tab switching. It sits right next to my clock and gives me a running total. When I see the number climbing, I naturally optimize my prompts.
2. A feed blocker (not an app blocker)
This was the big one. I built Monk Mode to block feeds specifically — Twitter's timeline, YouTube's recommended section, Reddit's front page — without touching the rest of the app. I can still search YouTube for a tutorial. I just can't fall into the recommendation rabbit hole.
The difference in my productivity was immediate. App blockers always felt like punishment. Feed blocking feels like removing a trap.
3. Photo-based food logging
This one's personal. I kept skipping meals while deep in code, then eating garbage at 11 PM. Manual calorie tracking was too tedious to stick with. So I built a phone app where I just snap a photo of my food and AI handles the rest. No barcode scanning, no searching databases.
The Economics Actually Work
Here's the math that surprised me:
- Previous SaaS costs: ~$35/month ($420/year)
- Building time: ~2 weeks total across all three
- Ongoing cost: $0 (they run locally)
But beyond saving money, I got something better: tools that work exactly how I think. No feature bloat. No onboarding flows. No "upgrade to pro for this basic feature."
When to Build vs. Buy
I'm not saying everyone should build their own tools. My rule of thumb:
- Build when the problem is simple but personal — your workflow is specific enough that no SaaS will nail it
- Buy when the problem is complex and shared — collaboration tools, cloud infrastructure, design software
- Build when you'd use it every single day — daily tools justify the investment
- Buy when you'd use it occasionally — don't spend a week building something you use twice a month
The Unexpected Benefit
Building your own tools forces you to deeply understand the problem. I learned more about attention and distraction building Monk Mode than I did from any productivity book. I understood my actual API spending patterns only after instrumenting them myself.
Sometimes the best tool is the one you build because the act of building teaches you what you actually need.
All three tools started as personal projects. Two of them (TokenBar and Monk Mode) are now available for other devs who have the same problems. Lifetime purchases, no subscriptions — because that was kind of the whole point.
Top comments (0)