Japanese study guide
Furigana, Audio, and Hover-to-Translate Sentences
Furigana you can switch off, audio on every word and sentence, hover-to-translate, and kana typing with no IME: reading support in Cotobara.
Most people don't quit Japanese because the vocabulary is hard. They quit because of friction: an example sentence with three kanji they can't read, a word they've never heard pronounced, a phone with no Japanese keyboard. Each one is small. Together they turn a five-minute study session into a chore.
Cotobara's reading tools exist to delete that friction. Here's what they are.
Furigana: on when you need it, off when you don't
Every Japanese word and sentence in Cotobara can display furigana: the kana reading printed above the kanji, rendered with the browser's native ruby text so it sits properly rather than crowding the line.
It's a single toggle in Settings, and the right answer changes as you improve:
- Learning to read? Leave it on. An unreadable sentence teaches you nothing, and guessing at a reading you've never seen just risks memorizing it wrong.
- Getting comfortable? Turn it off. Furigana is a crutch that quietly becomes a habit: if it's always there, your eye learns to read the kana and skip the kanji entirely, and you never build kanji recognition.
There's no single right setting, which is why it's a toggle and not a policy. Many learners flip it off, struggle for a week, and leave it off.
Natural-sounding audio on everything
Every vocabulary word, every example sentence, and every kana character in the beginner course has audio, in both male and female voice sets. It's generated with high-quality neural voices rather than recorded by voice actors, the class of voice you'd hear from a modern phone assistant, not the robotic text-to-speech of a decade ago.
That matters more than it sounds. Japanese is full of pitch-accent distinctions and long/short vowel contrasts that written romaji simply doesn't capture: 箸 (chopsticks) and 橋 (bridge) are the same syllables with a different melody. If you only ever read a word, you learn a version of it that native speakers won't recognize, and that's much harder to unlearn later than to learn correctly the first time.
Audio plays automatically when you want it to (there's an autoplay toggle) and on demand from a speaker button everywhere a word appears. One deliberate exception: a lesson check never autoplays before you answer it. A question that pronounces its own answer isn't a test. Once you've answered, it plays like everywhere else.
Hover-to-translate example sentences
Every word comes with a real example sentence, because a word in isolation only half-teaches you: you need to see what it attaches to, which particle it takes, and what register it belongs to.
The problem with example sentences at a beginner level is that they contain other words you don't know yet. Traditionally you'd stop, open a dictionary, look up three words, and lose your place.
In Cotobara, point at a word in a sentence and, where we have a gloss for it, you get its reading and its meaning, right there. No new tab, no dictionary app, no losing your place. The sentence is broken into real words rather than characters, so you're hovering 食べます, not 食.
Two touches that keep it from being annoying:
- Particles and punctuation are skipped: you're not getting a tooltip on は.
- The word you're studying is skipped: the sentence would otherwise hand you the answer.
Where this feature actually stands: we'd rather say what it can't do than oversell it. Most words in a sentence have a gloss, but not all of them do: some words come up blank, and some show the wrong meaning. The hard part is that plenty of Japanese words are several unrelated words sharing a spelling (おく is both 置く "to put" and 奥 "the inner part"), and picking the sense that fits that sentence is a problem we haven't fully solved. We're improving the coverage and the accuracy level by level.
So treat it as a helpful shortcut rather than a dictionary you can trust blindly. If a gloss looks wrong, it might be.
Type kana without an IME
On reading questions you type the answer in kana, and you type it in romaji, which converts as you go. ta becomes た as you finish the syllable, kya becomes きゃ.
No Japanese input method, no OS settings, no keyboard switching. This works on a laptop you don't control and on a phone without touching your keyboard settings. Setting up a Japanese IME is a genuine barrier for beginners, and it's a strange one to put between someone and their first study session.
A couple of details that keep it from being frustrating: hiragana and katakana answers are treated as equivalent, so you're never marked wrong for the script you chose. And if you leave stray romaji that never converted, the field nudges you to fix it instead of quietly marking you wrong.
It all shows up on the public word pages too
You don't need an account to see most of this. Every word in Cotobara has a public page (here's 飛行機) with its reading, furigana, meaning, part of speech, pitch accent, example sentences with hover glosses, and audio. Look up a word from a search engine and you get the real thing, not a signup wall.
Start for free, or read the complete feature tour. New to reading Japanese? Start with the hiragana and katakana guide.