I Moved to Portugal and Somehow Ended Up Building a Language App

Stanislav M. — Fri, 09 Oct 2026 18:41:06 +0000

I Moved to Portugal and Somehow Ended Up Building a Language Learning App

When I moved to Portugal, I started learning Portuguese. I thought finding a good language learning app would be the easy part, considering how many of them are available today. But I quickly noticed something annoying: a lot of the Portuguese learning materials I found were focused on Brazilian Portuguese.

There's nothing wrong with Brazilian Portuguese, of course. But I live in Portugal, and I wanted to learn the words, expressions and pronunciation I actually hear around me. Sometimes the differences are small, sometimes they're quite noticeable, especially when you're a beginner and already struggling to understand what people are saying.

I'm a QA engineer, and I also develop Android apps and small games in my free time. So naturally, instead of spending more time looking for a suitable learning tool, I decided to build my own. Because apparently I don't know how to solve small problems in a normal way.

That's how Retelingo started.

It was supposed to be a simple vocabulary app

My original idea was pretty straightforward. I wanted something where I could learn useful words, practice remembering them and review them regularly. I wasn't interested in building another huge language learning platform with hundreds of features, complicated learning paths and endless notifications. Just a practical tool I could use every day.

I decided to use active recall and spaced repetition, with FSRS handling the review scheduling. Instead of simply showing the same words over and over again, the app should help you remember them and decide when it's time to review them again. It's not some magical method that makes you fluent in a month, but it gives vocabulary practice a bit more structure.

Another important decision was offline support. I don't think an application needs a permanent internet connection just to show vocabulary cards and remember your progress. Retelingo downloads dictionaries when needed, but the actual learning process works offline. I also didn't want mandatory registration, because creating an account just to memorize a few Portuguese words seems a bit unnecessary to me.

At this point, everything still sounded relatively simple. Unfortunately, the development part had other plans.

The app was easier than the dictionaries

One thing I underestimated was how much work the language content would require. Writing the application logic is one problem, but maintaining dictionaries across different languages is something else entirely. Translations need to make sense, regional differences matter, pronunciation isn't always straightforward, and the same word can have several meanings depending on the context.

And when you start adding more languages, all those problems multiply. It isn't enough to generate a large collection of vocabulary and call it finished. You also need to think about content quality, consistency between dictionaries and what happens when the content changes after people have already started learning it.

For example, imagine someone has been studying a word for several weeks and already has review history associated with it. Now you discover that its translation needs fixing, or the dictionary structure needs updating. Ideally, you should be able to improve the content without destroying the user's existing progress. These are the kinds of problems that aren't particularly exciting to show in screenshots, but they matter quite a lot.

I'm still working on improving the dictionaries and expanding language support. And the more I work on the content, the more I realize that adding another language is probably the easy part. Making sure it's actually useful for someone learning that language is much harder.

Being a QA engineer doesn't make development easier

There's another funny thing about building your own software when your main job is testing other people's products. You already know too many ways something can go wrong, and unfortunately, you can't really ignore that knowledge when working on your own app.

Every relatively simple feature creates a bunch of additional questions. What happens if someone closes the application during an update? What if imported vocabulary contains broken data? What if the phone is offline, the dictionary version changes, or some local data becomes corrupted? And, of course, what happens when someone uses the application in a completely different way than I expected?

This is probably why my small projects tend to become much larger than planned. I start with a feature, then find edge cases, then write tests, then discover that the original implementation needs changing. Somewhere in the middle of all this, I occasionally remember that I was supposed to be making a language learning app.

And yes, I still manage to introduce plenty of bugs myself. Working in QA doesn't seem to provide any special immunity against that.

Where the project is now

Retelingo has grown beyond my original idea of learning European Portuguese. It now supports multiple languages, with more content and improvements still in development. I'm trying to keep the learning process simple while making the underlying system reliable enough to handle new dictionaries, updates and different learning scenarios.

I don't know yet how far the project will go. I'm developing it alongside my other apps and games, and there are still plenty of things I want to improve. What started as a personal problem has become an interesting combination of language learning, software engineering, content management and QA.

If you're curious about the project, you can find it here: https://morozhenko.com/rt/dv/po. I'm the developer, and I'm interested in feedback, especially from people learning European Portuguese or working on similar offline applications. I'm much more interested in hearing what doesn't work than collecting polite comments saying everything looks nice.