Why you should not depend on a language app

The plateau is built into the format

A course is a fixed sequence. That is what makes it a course, and at the very beginning it is an advantage: when you know nothing, almost any reasonable order works, and having the order chosen for you removes the decision that stops most people.

But a fixed sequence has to be written for an average learner who does not exist. Early on that barely matters — everyone needs the alphabet and everyone needs 'water'. By the time you have a few hundred words, the words you need next are determined by what you are actually doing with the language: the meeting you sit through, the textbook chapter you were assigned, the drama you are watching, the in-laws you are meeting. No sequence written in advance knows any of that.

This is why the plateau feels like a personal failure and is not one. You did not stop trying. The content stopped being about you.

Recognition is not production

Tapping the correct tile out of four is recognition. Choosing the right word out of everything you know, conjugating it, and saying it out loud before the other person looks away is production. They are different skills and they improve separately.

Most app exercises train recognition, because recognition is what a phone can grade instantly and reliably. That is a reasonable engineering decision and a bad model of the thing you are trying to learn. It is also why people who have finished an entire tree can still freeze in a two-line conversation — the freezing is not a gap in their knowledge, it is a skill they never practised.

The fix is uncomfortable and cheap: say things out loud, from memory, badly, every day. Write three sentences about your own day rather than translating someone else's. Anything that makes you retrieve rather than recognise.

A streak measures attendance, not ability

Streaks work. They are the most effective habit mechanic anyone has shipped in this category, and that is exactly the problem: the thing being reinforced is opening the app, and 'opened the app' is the metric that will keep going up while the thing you care about stops.

Notice what a streak cannot tell you: whether you could hold a two-minute conversation today that you could not hold last month. If you have a 400-day streak and cannot answer that, the streak has been measuring something else for a long time.

Replace it with a milestone you could fail. Record yourself talking for ninety seconds without stopping, once a month, and keep the recordings. It is a worse feeling than a streak and a much better instrument.

Input has to be both slightly hard and genuinely interesting

The uncontroversial part of second-language research is that you need a lot of input you can mostly understand — a bit above your current level, not far above it. The part people skip is the second condition: you have to want to consume it, in the volume required, for months.

A fixed course can be at the right level or it can be interesting, and it can usually manage both for the first few weeks. It cannot manage both indefinitely, because what interests you is specific to you and what is at your level changes weekly.

Which is why the people who get past the plateau almost all did the same unglamorous thing: they found real material they actually liked — a show, a podcast, a YouTuber, a class, a book slightly too hard — and stayed with it long enough to be bored and then not bored.

What to do instead

Pick a source of real material and commit to it for a month. A class, a tutor, a textbook, a channel, a show with subtitles in the target language. It does not have to be optimal; it has to be yours and it has to keep producing new material.

Produce something out loud every day, even badly. Ninety seconds is enough. The point is retrieval under mild pressure, not performance.

Review your own mistakes rather than a generic word list. The sentence you got wrong yesterday is worth more than the twentieth most common noun, because it is evidence about you specifically.

Space that review out. Spaced repetition is the one piece of learning-science folklore that survives contact with the evidence: revisiting something just before you would forget it is dramatically cheaper than relearning it after you have.

Keep an app if it is doing a job. A script drill, a five-minute warm-up on a bad day, a habit anchor. The mistake is not using one, it is letting one be the whole plan.

And langNote is also an app

So the honest version: everything above applies to us wherever it applies to anyone, and the distinction we draw is narrow. langNote has no curriculum. It does not decide what you learn next, because it does not know anything you have not given it. It reads material you already have — a photo of the worksheet, a recording of the class, a textbook page — and turns that into practice and a spaced-repetition schedule.

That is a real difference from a fixed course, and it is also a real limitation. If you have no material and no course, we are the wrong tool and there is nothing clever to say about it: get a class, a textbook, or yes, an app with a curriculum, and come back when you have something to review.

Do not use langNote if you need to work offline, if you want to own your data as a local file, or if you are at the very start of a writing system — a free drill app will do that better than we will. Do use it if the reason your review has stalled is that turning today's lesson into practice takes longer than the lesson did.

We charge for it, which is a reason to read the paragraph above twice.

Questions people actually ask

Are language apps a waste of time?

No — they are good at the beginning and at habit-building, and bad at everything after that. The mistake is treating one as the whole plan rather than as one tool inside it.

Why do I plateau after a few months?

Because a fixed curriculum has to be written for an average learner. Early on almost any order works; once you have a few hundred words, the words you need next depend on what you are actually doing with the language, and no pre-written sequence knows that.

Is a streak a good measure of progress?

It measures attendance. A useful substitute is a milestone you could fail: record ninety seconds of unscripted speech once a month and compare the recordings.

What should I use instead of an app?

Real material you chose and will stay with, daily production out loud, review of your own mistakes rather than a generic word list, and spacing that review out over time.

Isn't langNote an app too?

Yes. The narrow difference is that it has no curriculum — it only works on material you bring. If you have no material and no course, it is the wrong tool and a course is the right one.