The Current State of Language-Learning Apps

Open any app store and you'll find language learning is booming. Hundreds of millions of people have a streak going somewhere. The apps are polished, gamified, and genuinely fun. And yet a quietly uncomfortable question hangs over the whole category: how many of those people can actually hold a conversation?

This isn't a takedown. Modern language apps got a lot right, and they pulled millions of people into trying — something traditional classrooms never managed at this scale. But it's worth being honest about where the category is, because the gaps point directly at where it's going.

What today's apps do well

Credit where it's due. The current generation of apps solved real problems:

• They removed friction. Five minutes, on your phone, on the bus. Learning a language stopped requiring a class, a textbook, or a tutor.
• They made it a habit. Streaks, reminders, and tidy daily goals are genuinely effective at getting people to show up day after day — and showing up is the hard part of any long-term skill.
• They lowered the fear. Tapping answers in private is far less intimidating than speaking in front of a class. For a lot of people, that's the difference between starting and never starting.

These are not small wins. Consistency and low anxiety are two of the biggest predictors of whether someone sticks with a language at all.

Where they quietly fall short

The trouble is that the metric most apps optimize for — daily engagement — isn't the same as the outcome learners actually want: being able to communicate.

A few specific gaps show up again and again:

Recognition isn't recall, and recall isn't conversation. Choosing the right answer from four options is much easier than producing the word yourself — and producing a word in a drill is much easier than producing it live, in a conversation that won't wait for you. Many apps stop at the easy end of that ladder.

Sentences without situations. Translating "the bear drinks beer" teaches you vocabulary, but language lives in context — who's speaking, why, and what they want. Disconnected sentences are quick to grade and slow to transfer to real use.

Streak guilt over real progress. Gamification is a double-edged sword. The same loops that build habits can quietly shift your goal from learning the language to protecting the streak — and those aren't the same thing.

Little real speaking. The single most important skill — talking with another person in real time — is also the hardest to build into an app, so it's often the first thing left out or reduced to reading sentences aloud.

Why it's built this way

None of this is because the teams are lazy or cynical. It's because the easy-to-build loop and the easy-to-measure metric happen to line up — and the genuinely effective stuff is inconvenient. Spaced retrieval means showing people things they got wrong. Comprehensible input means producing or sourcing a lot of level-appropriate material. Real conversation means handling the messy, open-ended, unpredictable shape of how people actually talk.

For a long time, that last one was simply too hard to do well at scale. That's the part that's now changing.

What comes next

The most interesting shift in the category is that the hardest gap — real, responsive conversation — is finally tractable. A learner can now practice speaking with a patient partner that adapts to their level, responds to what they actually said, and corrects gently instead of buzzing red. That doesn't replace human conversation, but it makes the on-ramp to it dramatically less scary.

The apps that matter over the next few years won't be the ones with the cleverest streak mechanics. They'll be the ones that close the loop between the three things that actually produce fluency: understanding real input, speaking early and often, and remembering what you learn at the right moments.

That's the bet we're making with Coral. If you want the longer version of the why, our companion piece on how people actually learn languages walks through the research the whole category is slowly converging on.