Chapter 2
Core features
NipponDaily opens one Japanese word a day with what its sources say: the reading, meaning, pitch accent and example sentences and, where they record it, its morphemes, its layer of the vocabulary, the processes that shaped it and the evidence behind every claim, with a calendar to look back through every word so far.
What a reader gets
A New Word Every Day
One entry opens each day at midnight in Japan (JST), the same word for every reader, all from the JLPT N5–N1 vocabulary, first the words chosen for their origin, then the rest, easiest list first, and last the words that share a spelling with an earlier one, easiest list first again. Each section shows what its source holds for the word and otherwise says the data is missing: the origin needs a Wiktionary Etymology section, the pitch accent an accent list and a second dictionary that agree, the sentences a Tatoeba match. 7959 words are written, May 2005 to February 2027, and 7830 have opened so far.
A Calendar to Look Back Through
/words is a month grid. A day that has arrived shows its word and links to the full entry; an upcoming day shows only its date, and the API refuses to serve it, so a word can't be read early; a day the catalogue has no word for is left blank, except that once the catalogue has run out each open day shows the word of its lap, marked ↻ with a legend entry saying the word was first opened on another day (its link goes to that day). A short note under the heading explains that a word sits on the day it first opened, and each entry page is headed “First opened” with its date. Jump to any year and month from the picker above the grid: a year menu with earlier and later buttons, so it stays one row however many years the catalogue spans, and the twelve months of that year. “Filter the calendar” opens the Explore filters (search, level, layer, process, part of speech, how common, any or all): the month's matching words are outlined in the grid, days whose word doesn't match fade but stay links, each month in the picker shows how many matches it holds, and a month with none points to the nearest months that do. The month and filters live in the URL, so a view can be shared.
Explore by How Words Are Built
/explore searches every word that has opened — by the word, its reading (katakana, hiragana or romaji, so taberu finds 食べる) or its meaning — and narrows by JLPT level, layer (including “not stated”), process, part of speech, how common JMdict says the word is, or the pattern of its pitch accent (flat, head-high, middle-high, tail-high, or none shown). Pick several options in a group and choose whether a word needs any or all of them. Each option shows how many words it would leave, the list shows the newest matches with “Show more” for the rest, and the filters live in the URL so a search can be shared. The calendar uses the same filters, folded behind the same “Filter” button and one row of group buttons. Upcoming words are never searched.
Patterns Across the Vocabulary
/patterns counts the same entries across words: how the vocabulary splits by layer, how each JLPT level's mix differs, which processes are most common and which turn up together, in pairs and in sets of three to four. Every bar links into Explore, and the page says what the counts can't tell you — they describe these entries, not the language. It also lists which sounds voice inside a word (ひ → び, か → が…), with the parts and words that show each change — a small, parser-derived sample, not a rule of the language. Two more sections count how the pitch accents the entries show split into patterns for each layer, and how long each level's words are in morae and kanji, and a last one lists the readings several words share and how often the pitch alone tells them apart.
The Parts, Across Words
/parts gathers every part a “Taken apart” row has shown. Open one — 日, say — to see each word it turns up in, grouped by the reading it takes there (び, ひ, か, にち), with rendaku shown. Words that merely contain the character, with no breakdown naming it, are listed apart and claim nothing.
The Kanji
/kanji lists every kanji a word that has opened is written with. Open one — 日, say — for KANJIDIC2's record of it (on'yomi, kun'yomi, meanings, school grade, stroke count, newspaper rank), the order its strokes are drawn in (KanjiVG's, shown only where KanjiVG and KANJIDIC2 count the same number of strokes) the shapes EDRDG's KRADFILE lists it as containing (a lookup aid, not an etymology), with a few other kanji that contain the same shape, and every word so far that is written with it, grouped by JLPT level, easiest first. Each entry links its own kanji, and the page says a word doesn't use every reading or meaning of its kanji.
In a Sentence
Most entries show up to two sentences that use the word, with their English translations, from Tatoeba's community collection. They are picked by fixed rules from one dated export, quoted unchanged, linked to their Tatoeba pages, and credited. Furigana is Tatoeba's own, drawn over a kanji only where it passes a check against the dictionaries and a contributor or a second source stands behind it; a word is shown in a sentence whose kanji are all read where one exists, and a reading that cannot be settled is left bare, not guessed. Nobody has reviewed them one by one, and the page says so.
More Like This
Under each entry on its own page (/words/<date>), a few other words that have opened and share something with it: a part, a process or a layer. Each card says exactly what is shared and links to where you can see more of it. A word that shares only very common tags is not offered, and upcoming words never are.
Words That Sound the Same
Under an entry, the other words so far with exactly its reading and another spelling (橋, 箸 and 端 are all はし), each beside the entry with the pitch accent its own entry shows, so you can see where the pitch tells them apart. Only the readings are compared.
The Same Day in Other Years
Under each entry and under the word on the home page, the words that opened on the same day of the year in other years, oldest first. They share only the date, and the page says so.
Opposites and See Also
Where JMdict gives the sense an entry's meaning came from an antonym or a "see also", the entry page names the words in a block of its own, as JMdict writes them. A word that has opened links to its day; one that has not is only named, never linked, and a spelling with several open words (上 is うえ, かみ and じょう) is not linked at all. Nothing is added to what JMdict says.
Taken Apart
Where Wiktionary splits a word into parts with glosses, each morpheme is shown with its reading and that gloss, but only if the parts spell the word and join to its reading. When the text gives no split of a word written with kanji (and the okurigana after them), each part is one of its kanji with KANJIDIC2's reading and meaning, and the page says so. Otherwise no breakdown is shown, since any other split would be a guess.
How It Is Said
Where the pinned Kanjium accent list gives a word's exact spelling and reading and UniDic gives the same accent, the entry shows its pitch accent: the reading with a line over the high morae and a ↓ after the one where the pitch falls, named flat (heiban), head-high (atamadaka), middle-high (nakadaka) or tail-high (odaka). A word the two do not agree on shows only the accents they share, and says what each gave; where both give several accents they are all shown.
Which Layer, Which Process
Every entry shows JMdict's part-of-speech tags, a “Common” mark when JMdict counts the word as common, JMdict's own register, field and dialect tags where every sense the meaning matches carries them (colloquial, polite, baseball, Kansai-ben…), its layer when the spelling, KANJIDIC2's readings, JMdict's loan source or the quoted text establish it — native 和語, Sino-Japanese 漢語, loanword 外来語 or hybrid 混種語 — and the processes its Wiktionary text mentions, or JMdict records (compounding, rendaku, clipping, ateji, sound change…), each defined on the page.
The Story, in Wiktionary's Words
Where the word comes from is quoted line by line from Wiktionary's Etymology section for that reading, never paraphrased. Lines that hedge (“probably”, “unknown”) are flagged “Not settled” instead of being turned into a verdict. Where the same page also has sections for other readings of the spelling (抱く だく and いだく), they are quoted in a closed block of their own, labelled with their readings and never mixed into the word's origin.
Evidence for Every Claim
Each entry quotes the Wiktionary lines it shows, taken from one dated dump, with a link to the page and its CC BY-SA 4.0 license. Only the one-line headline is hand-written.
Share a Word
A shared link unfolds into a card: the word, its reading, meaning, level and layer, and the headline, in the colours of the season it falls in. It never shows a word before its day.
Practise Without Keeping Anything
An entry can cover its meaning until you tap it, and a sentence can hide the readings over its kanji, so you can test yourself on the word. Both live in the page alone: nothing is saved, and the next word starts uncovered.
Verified in CI
Every entry is regenerated from JMdict, KANJIDIC2 and the pinned Wiktionary snapshot and compared, and checked independently: reading, level and meaning against the pool, part of speech, register, field and dialect against JMdict, morphemes against KANJIDIC2 or the cited text, priority, loan and ateji claims against JMdict, each pitch accent against the committed accent snapshot and the word's count of morae, each kanji's strokes against KANJIDIC2's stroke count, each example sentence against the committed export snapshot, and every quote against the section for its own reading.
Four Seasons
The site follows the Japanese calendar: spring, summer, autumn and winter. The browser works the season out from the date in Japan and checks again at every midnight. Each one changes the palette, the shapes of cards, buttons and badges, and the falling petals, bubbles and fireflies, leaves or snow.
Pick Your Season
The season button in the header lets you choose any season, or “Follow the calendar”. Your choice is kept in this browser only.
Season Music
The header's music button plays a looping background track for the season on screen. Every track is mastered to the same loudness, and changing season while the music plays crossfades into the new track. It is off every time the page loads, and the volume you choose is remembered in this browser.
Kana Reference
A hiragana/katakana chart with romaji and shape mnemonics at /kana, laid out as 学業守 omamori charms and ema plaques.
Nothing Stored About You
No accounts, no tracking. The site fetches words, the month grid and the counts (the season is worked out in your browser), and sends nothing about you back. Only your color mode, season choice and music volume are remembered, in your own browser.
Dark Mode & Fallbacks
A light and a dark palette for every season. If a fetch fails, a retry card appears instead of an empty page, and once the catalogue runs out the home page starts another lap from the first word.