Software for One

Back in 2020, Robin Sloan detailed his experience creating BoopSnoop, a bespoke messaging tool for his family. While the project only took a week, he spent half that time trapped in the frustrations of code-signing.

"In a better world, I would have built this in a day, using some kind of modern, flexible HyperCard for iOS."

The Arrival of the "Personal" Era

Fast forward to 2026, and that sentiment has resurfaced. A user named Thariq noted on X that while "personal software" might have been a premature concept in 2020, it is now a reality. He likened it to a mass-produced frozen dinner home-cooked meal or a handwritten letter.

Thariq and his spouse developed a baby tracker specifically because they didn't want the bloat of modern SaaS:

• No user profiles
• No gamified badges
• No tiered subscription models
• No unnecessary "extra" features

---

My Suite of Bespoke Tools

Over the last half-year, I have leaned heavily into this philosophy, building a variety of hyper-personalized applications:

1. The Sleep Orchestrator: Converts a PDF from our sleep consultant—which was essentially a complex set of conditional logic (nap caps, wake windows, etc.)—into a live, shared schedule for my wife, our nanny, and me. It dynamically adjusts when a nap ends early.
2. The Dynamic Fitness Hub: A central node that connects disparate data sources to provide insights no commercial app offers.
3. The Race-Based Marathon Planner: Instead of using generic templates, this derives a plan from my actual Strava history and past race results.
4. Jazz-Lingo Duolingo: A custom quiz app for practicing chord voicings based on my specific piano lessons.
5. Medical Gap Finder: A tool designed to highlight missing records before visiting a specialist.

Deep Dive: The Fitness Ecosystem

My fitness app is the crown jewel of this approach. It doesn't just track; it calculates. It knows my marathon schedule and adjusts my nutritional needs in real-time.

Logic Flow:

• Input: Morning run distance + intensity $\rightarrow$ Output: Adjusted calorie target.
• Input: Upcoming long run $\rightarrow$ Output: Specific carb/protein reminders and "carb-loading" alerts.
• Input: Daily training volume $\rightarrow$ Output: Precise weights for the 8 ingredients in my morning smoothie.

Data Integration Architecture

The app functions by aggregating data that usually lives in silos:

---

The Philosophy of Sovereignty

The greatest advantage here is ownership. I am no longer subject to:

• $\text{Sudden UI redesigns}$
• $\text{Intrusive advertisement floods}$
• $\text{Corporate pivots to chase a different demographic}$

If my wife loves a specific feature in the sleep app, it stays there indefinitely.

The Technical Blueprint

I've utilized a consistent stack for a few years that balances speed with granular control.

My Development Stack

Note: For those less comfortable with a terminal, tools like Claude Artifacts, Replit, or Lovable can achieve similar results without deep technical overhead.

The Financial Breakdown

The monthly overhead is approximately $160, broken down as follows:

While this exceeds what I'd pay for a bundle of subscriptions, the Claude subscription is a baseline cost I'd incur anyway. Furthermore, the Neon and API costs are driven by larger public projects like nycjazz.guide and claudecodedaily.com. For a hobbyist with one or two apps, the cost could be as low as $25–$30/mo using free tiers.

---

The New Development Paradigm

A few years ago, building these would have been unthinkable due to the maintenance burden. Now, the process is radically different.

The Workflow:

1. Identify a bug or need a feature.
2. Take a screenshot of the issue.
3. Feed it to Claude.
4. Deploy the fix.

The "cost" of change has plummeted. I no longer fear the "one feature request" trap that Sloan mentioned; adding functionality is now trivial.

The Pattern of Personal Software

Most of my apps follow a specific mathematical and logical shape: $\text{Data Sources} \rightarrow \text{Aggregation} \rightarrow \text{LLM Context} \rightarrow \text{Personalized Insight}$

This allows me to build for audiences too small for a VC-backed company to care about, such as NYC jazz enthusiasts or developers using Claude Code.

The API Economy

I recently migrated from Cronometer to FatSecret simply because the latter offered a superior API. As this trend grows, the value of an app will be measured by its API and MCP (Model Context Protocol) quality—how easily it can be plugged into a user's personal ecosystem.

The Irony of Creation

There is a psychological hook to this. Instead of passive consumption (Netflix or X), I spend my nights building. The proximity to the "next unlock" is addictive.

The Paradox: I have spent countless sleepless nights coding applications specifically designed to optimize my health and sleep.

Looking Forward

Currently, this path requires technical literacy—comfort with git, databases, and deployment pipelines. However, we are moving toward a world of "single-prompt" software. As Sam Altman recently hinted, the future is as simple as telling an AI to "plan a trip," and having the software manifest around that need instantly.