Personal Systems Can't Be Downloaded from the App Store

Here we are, in the second half of 2026. If you build software for a living, you've felt the big shift going through this industry. LLMs are changing what it means to be a developer. Some people are struggling to find meaning in work that used to define them, and others are sure AI will take all of our jobs. Meanwhile, some influencers hype AI so much that it's hard to take any of it seriously. There are enough texts about all of that already.

This one is about the other side of it. For the first time, people like me are building software that is truly personal and fully suited to their digital needs.

It started for me early in 2026, when the models got good enough to change how I work. For most of my career I've been a front-end developer who knew a bit about databases, DevOps and design. Now I can take an idea from start to finish on my own. Things that used to take months now take days, and it feels like I could build almost anything, even manage my own servers, databases, and deployments, things I always used to leave to others while I focused on the interface.

For my personal projects, that meant leaving Vercel and managed platforms behind. Now everything runs on a Hetzner VPS in the EU, deployed with Dokploy, and it's all open source. The whole stack is mine now, not just the front end, and I can put whatever I want on it: small experiments, data visualizations, learning tools, project monitoring, and complete, secure web products with their own APIs.

With all of that in my hands, I started thinking, like many others, about building my own tools. I wanted software built around how I actually live and work, and searching online for something that already did what I needed started to feel like more work than just building it myself.

So that's what I did. I built a system for myself. Part of it is public: my writing, links I've saved, a way to contact me, the usual things you'd expect on a personal website. Behind it are four small tools only I use:

• A place to write, where this text starts
• A link saver, with Chrome and Firefox extensions that save straight to my database
• A small Anki-like app for learning German, with spaced repetition, voice input, ElevenLabs reading words and examples aloud, and Mistral AI filling in grammar and examples
• A dashboard that shows what's running on Dokploy, where all my projects live

That's all of it. A bit over 350 commits, one Next.js app, one SQLite file.¹

It's also a PWA, and that part still feels a little unreal. It sits on my iPhone home screen next to apps from the App Store, with its own icon and no browser around it, and it opens just as fast. Nobody has to approve it. I push a commit and it's there.

The German trainer is the part I enjoy most. I couldn't find a tool that worked the way I wanted, so I made one. When I hear a new word or sentence while walking through Vienna, I add it right there on my phone, and AI fills in the grammar for me. Learning a language with something I built for myself feels really good.

Since it's all one app with one database, adding something new doesn't take much. Notes would be a table and a form. Reminders would be a date and a quick check every morning. No new account, no new subscription, no new app to open, just another screen in something I already use every day. That's what makes it a system and not a bunch of separate apps.

So when I need something, I build it instead of buying it, usually in an evening. And the ideas keep coming. Next might be a shared list for my family, or my Hue lights turning on when the No. 12 tram is almost home. How far can you take something like this, and how crazy can it get?

Pretty far, it turns out. The hard part is knowing when to stop. Every time something small bothers me, I'm tempted to build a fix for it, and most of the time I probably shouldn't. Building a feature is quick now. Keeping it working, simple, and maintained for years is not.²

It isn't free, either. Nobody else updates the server, sets up backups, or fixes things when they break late at night. If you own everything, you own the annoying parts too. So before I add anything new, I ask myself if I'll still care about it a year from now. Which leads to a bigger question.

Why not just use something that already exists?

Each of these four tools has a better commercial version. Anki has years of research behind it and plenty of plugins. Ghost or WordPress would give my writing section far more features than it has.

But nobody sells all four together, working as one. Four products means four accounts, four subscriptions, four different interfaces, often clunky, with APIs that are hard to work with or not there at all. And all of them are made for an average user who doesn't exist.

I also got tired of switching. I saved links in Pocket until it shut down, then in a Notion database. Every change meant moving my data and hoping someone else's plans matched mine. Now I build it myself, and it does exactly what I want.

And it isn't just me. If you work online, you probably have the same problem I had: too many subscriptions that don't talk to each other, with your data spread across all of them. You could have one system instead, one that does all of it.

I think this is where software is going. Not for everyone, and not all at once, but away from apps made for a million people and toward systems made for you. Systems that fit your data, your language, your habits.

You won't find them in the App Store, or on any other platform. They have to be built around how you live, learn, teach, and run your business. The tools are here now, and so is the time for fully personal systems.