Wakanai — casual Japanese for your Japan trip

The problem

Apps teach you to tap. Japan asks you to listen and answer.

A 500-day streak doesn't order food at an izakaya. Real conversation means catching real speech at real speed — and that skill only grows one way: lots of listening you can actually understand, then opening your mouth. That's the whole app.

How it works

Three passes. Zero grind.

Each session is one real podcast episode, three times, each pass with a job:

一 · Listen

Pictures and ears only

Watercolor scenes switch in time with the story, carrying the meaning. No text anywhere — you understand from context, the way kids do.

二 · Replay

Now with subtitles

The same episode with Japanese + English line by line, core words lighting up as they're spoken. Confirm what you understood; catch what you missed.

Replay pass with Japanese and English subtitles

三 · Speak

Say it out loud

Each new word plays in a clear studio voice. You say it back — and on-device speech recognition tells you honestly how it went.

Word test asking the learner to speak Japanese

Join the beta →

Memory that argues back

Your voice is the test.

Reviews are spoken, not tapped. Hear the Japanese, say the meaning in English — the app checks your answer. No self-graded "yeah, I knew that."
Pronunciation is graded sound-by-sound, forgiving what natives forgive, catching what matters.
FSRS scheduling brings each word back right before you'd forget it.
Wrong verdict? Override it. The recognizer is a judge, not a tyrant.

A review card showing Correct! after speaking the meaning

Kanji without the wall

Hard words wait until they're easy.

Kanji are built from smaller pieces, so Wakanai teaches the pieces first: 電球 (light bulb) stays out of your queue until you've met 王 and 水 — then it unlocks by itself. New words arrive five a day, always one step from what you already know.

Private by design

Your learning never leaves your phone.

Speech recognition on-device. Your voice is never uploaded — the Japanese and English models both run locally on your iPhone.
No account. No cloud. No telemetry. The app doesn't know who you are, and doesn't want to.
Handmade by machine, curated by a human. Every episode's illustrations, translations, and voices are produced by the developer's own AI pipeline — you just download the app and press play. (No, you don't need a Mac.)

Wakanai home screen with today's session

Questions

Before you ask

Is this for total beginners?

Yes — that's exactly who it's built for. Day one starts with こんにちは-level words and everything harder waits until the app has proof you're ready. If you already read manga in Japanese, this will feel slow. Curious why listening-first works? Read about the method.

Why iPhone only? Why iOS 26?

The private, on-device speech engine Wakanai is built on ships with iOS 26. No iOS 26, no honest voice grading — so for now it's iPhone (or iPad) on iOS 26 and newer. Android folks: join the list anyway and we'll count you.

Do I need a Mac or any setup?

No. The developer uses a Mac to produce episodes; you just install the app from TestFlight and press play.

Is my voice uploaded anywhere?

Never. Recognition runs entirely on your iPhone. The app has no server, no account, and no analytics.

What happens after I join the list?

You'll get one email with a TestFlight link when your wave opens. Beta is free. Reply "stop" to any email and you're off the list.

Beta

Get on the list.

Wakanai is iPhone-only (iOS 26+). Invites go out in small waves, oldest signups first.

Signups aren't wired up yet — the API endpoint hasn't been deployed. (This notice is only visible while SIGNUP_ENDPOINT is a placeholder.)