What is SpeakBeats?
Your brain has been trying to tell you something about how it learns languages.
You probably can't remember what you had for lunch on Tuesday. But you can almost certainly recite, word for word, a song you haven't heard in fifteen years. Not because you tried. Because your brain decided.
Every flashcard app you've abandoned, every textbook gathering dust, every streak you let die at day 47 - they were all fighting that. SpeakBeats is built around it.
Why most language apps quietly fail you
The apps you've tried aren't bad. They're just built around the wrong assumption - that learning a language is a discipline problem. Show up daily, drill the cards, check the box. The brain you actually have doesn't work that way. It hoards what it finds emotionally meaningful and forgets the rest, no matter how many gold stars you collect.
That's why the streak dies. Not because you lacked willpower. Because nothing in the material was reaching the part of your brain that actually decides what's worth keeping.
Music is how memory was meant to work
Long before writing, before classrooms, before grammar books - every culture on earth used song to carry its language across generations. Epics, prayers, lullabies, nursery rhymes, the alphabet. We didn't pick melody as a study aid. We picked it because it's the format human brains hold onto best.
The research caught up to the intuition. A few of the studies that shaped how we designed SpeakBeats:
- Sung words stick harder than spoken words. In one well-known experiment (Schön et al., Cognition, 2008), adult learners picked up new vocabulary roughly twice as well when it was sung versus spoken - same words, same exposure, just melody added.
- Music lights up the reward circuit. Salimpoor and colleagues (Nature Neuroscience, 2011) showed that music you enjoy triggers a dopamine release in the same neural pathway that consolidates memories. The pleasure isn't decoration - it's literally what tells your brain this is worth remembering.
- Music and language share a brain. Aniruddh Patel's OPERA hypothesis (2011) showed that music and language use overlapping neural machinery for pitch, timing and structure. Train one carefully, and you scaffold the other.
- Songs review themselves. The "earworm" everybody complains about is just involuntary musical imagery - your brain rehearsing a melody on its own clock. When the lyric is in Japanese, that's spaced repetition you didn't have to schedule.
Put together: a song you actually like is a delivery mechanism that bypasses the bored-and-grinding part of your brain entirely, and lands the language straight into the part that won't let go.
You don't review the songs. The songs review you.
But the lyrics have to be the real language
Here's where most "learn a language through songs" attempts fall apart. Pop songs are full of poetic contractions, slang, and lines no one would ever say to a coworker. Listening to them is fun, but it teaches you a register you can't actually use.
SpeakBeats songs are written backwards from how songs are normally written: we start with the conversational Japanese we want you to internalize, and then build a song around it. Every lyric is something an actual Japanese speaker would actually say. The melody you can't get out of your head is also, line for line, vocabulary and grammar you can drop into a real conversation.
How a SpeakBeats lesson works
Every lesson is built around one song. We pick the vocab and grammar points first (aligned to the JLPT levels, starting at N5), teach them with examples and short drills, and then - only then - write the song. By the time you hear the track, every word in it is a word you've already met.
Tap any lyric while it's playing and you'll see the word, its reading, and its meaning. Tap a grammar pattern and you'll get a quick explanation in plain English. Spaced review keeps surfacing the words you almost-remember, so the lessons compound instead of evaporating.
You're not memorising a list. You're learning a song that happens to be made of Japanese you can use.
Two ways in
We don't think learning a language has to look one specific way. There are two honest paths through SpeakBeats, and which one fits depends on how serious you are right now - not on how serious you'll ever be.
The casual path
Just listen.
Open the player. Stream the full SpeakBeats playlist. Sing along, or don't. Tap a word when one catches your ear. Close the tab when you're done.
This is the slowest path on paper and one of the most underrated in real life. People who just listen, week after week, end up absorbing more conversational Japanese than people who grind apps and quit after a month.
The accelerated path
Take the course.
Same songs, but now they're embedded in a structured JLPT course. Vocab introductions, grammar explanations, dialogue practice, exercises, drills, and spaced review - all built around the song that closes each lesson.
Every minute in the course makes the songs in the player land harder, and every song in the player reinforces what the course taught you. They're the same engine at two different speeds.
It's free
The whole thing - the player, the full curriculum, every lesson, every song - is free right now. No signup wall, no trial countdown. We're building this for ourselves and the friends who keep asking us to teach them, and we'd rather have you actually using it than gated out of it.
New lessons drop on YouTube regularly.
The honest promise
We're not going to tell you you'll be fluent in 30 days. Languages don't work like that and neither do brains. What we will tell you is this: if you spend time with SpeakBeats - really spend it, casually or seriously - you're going to catch yourself humming a Japanese line on the train, and realize halfway through that you know exactly what it means.
That's the moment everything else gets easier from. We just built the shortest bridge to it we know how to build.