N2 is where vocabulary stops being a list you can finish and becomes a volume you have to manage.
The grammar and the reading are their own problems, but vocabulary is the one that quietly caps everything else. If you cannot read a word, you cannot parse the sentence, and no amount of grammar study rescues a sentence you cannot decode.
This is a plan for handling that volume with spaced repetition, sized for someone with a job or classes rather than an open calendar.
What N2 vocabulary actually requires
One thing to get straight first: the JLPT does not publish official vocabulary lists. There is no canonical N2 word list issued by the test makers. The lists you find online are compiled from past papers, textbooks, and publishers' materials, which is why they disagree with each other at the edges.
That has two consequences worth accepting early.
First, no list is complete, so treating any single list as the syllabus will leave gaps. You are building coverage, not checking boxes.
Second, because there is no official list, you cannot measure readiness by "words remaining." You measure it by whether you can read the material you are trying to read.
Where to get words from
The strongest source is the material you are already studying, because those are the words that will keep appearing.
- Your textbook or the N-level vocab books you already own. These are graded and roughly ordered by difficulty, which makes them good card fodder.
- Past papers and practice tests. Words that show up in questions are, by definition, words the test cares about.
- Reading you do outside studying. Articles, news, short stories. Note words you had to look up twice.
- Dictionaries and frequency tools. Useful for checking whether a word is worth learning, less useful as a primary source.
The order matters. A word you met in a sentence you were trying to understand comes with context attached, and context is what makes it stick. A word met only inside a list is a word you will forget.
Card format: the decision that matters most
For Japanese, the card format is the whole game. Get this wrong and you will build a deck that tests your ability to recognize kanji you already recognize while never testing the thing you actually need.
The core problem is that Japanese has several different things to know about one word:
- the reading — how it is pronounced
- the meaning — what it refers to
- the usage — which particles, which register, which collocations
A card that shows the kanji and asks for the meaning only tests recognition. That is necessary but not sufficient, because the exam asks you to read sentences, and reading requires the reading.
A practical split:
- Kanji → reading and meaning. The everyday card. Front is the word in kanji, back is the reading plus a short definition.
- Reading (kana) → meaning. Tests whether you know the word when it is not written in kanji, or when it appears in a sentence.
- Meaning → word. The production direction. Harder, and the one that reveals whether you actually own the word or merely recognize it.
- Word in a sentence → meaning and reading. The most useful format, and the most expensive to write. Worth it for words you keep failing.
If you can only do one direction, do kanji to reading plus meaning. If you can do two, add the sentence card for the words you miss. The kana-to-meaning direction is the one people skip and the one that costs them on the reading section.
A twelve-week weekly plan
This assumes roughly one to two hours a day, six days a week, with one lighter day. Adjust the volume, keep the shape.
Weeks 1–2 — build the habit, not the deck. Learn 20 to 25 new words a day. Review daily. The goal is not coverage; it is proving to yourself that the daily review happens. If you cannot hold the review here, more words will not help.
Weeks 3–6 — hold the rate. Stay at 20 to 25 new words a day. By now your daily review is mostly old cards, which is the point: the ratio of new to review should tilt toward review as the weeks go on.
Weeks 7–9 — switch emphasis. Drop new words to about 15 a day and add reading practice that uses the words you have learned. This is where vocabulary converts into comprehension, or fails to.
Weeks 10–11 — straight review, no new words. Clear the deck. Anything you have failed repeatedly, convert into a sentence card or delete it.
Week 12 — past papers. Stop learning and start testing. Note every word you had to look up, and those become your final small deck for the last days.
The shape to notice is that adding slows down while reviewing stays constant. That is what makes the plan survivable.
The trap: recognition versus production
Almost every N2 study failure with flashcards has the same cause. The learner builds a deck where every card shows the kanji and the deficit is invisible, because recognizing 語彙 when you see it is not the same as being able to read it in a sentence under time pressure.
The check is simple. Take twenty words you consider learned. Turn the cards around — show the meaning and the reading, ask for the word. If that goes badly, your deck has been testing recognition, and the exam does not stop there.
A practical compromise: keep the recognition direction as your daily workhorse, and add production cards only for the words you keep missing. Production cards are slow and demoralizing if you use them for everything.
How to tell it is working
Not by deck size. Deck size measures how much you have written down.
Judge it by whether reading got easier. Can you get through a page of your source material and look up fewer words than a month ago? Can you read a past-paper passage without stopping on vocabulary, even if the grammar still gives you trouble? That is the signal.
The second signal is the count of words you know without a card. Some words stop needing review because you now see them constantly in real material. When a card keeps passing easily, retire it.
Where a tool helps and where it does not
Building a deck by hand from a textbook is the part that stops most N2 plans, because it is mechanical and it never ends. If your source is a document — a vocab list, a chapter, a set of past-paper notes as a PDF or Word file — the first pass of card-writing is a good candidate for automation. WakeCards turns that file into question-and-answer cards in the same language as the source, which for a Japanese word list means it produces Japanese cards.
What it will not do is decide what to learn, check your readings for accuracy, or schedule your reviews. Readings in particular: generated cards are a draft, and with Japanese you should verify kanji readings against a dictionary before you commit them to memory. Memorizing a wrong reading is worse than not memorizing it.
The five-step loop works the same in any language. The vocabulary does not care what tool you used to write the card.
A first week you can actually do
- Pick one source. Your textbook chapter or your vocab book, not all of them.
- Make cards for one unit only. Kanji → reading plus meaning.
- Review daily, even for ten minutes.
- At the end of the week, turn the cards around for ten of them and see how that goes.
- If it went badly, add sentence cards for the words you missed.
That is the plan. It is not complicated; it is just daily, which is the hard part.
Frequently asked questions
Is there an official JLPT vocabulary list?
No. The test does not publish official word lists for any level. Published lists are compiled from past papers and teaching materials, and they disagree at the edges, so treat any list as a good estimate rather than a syllabus.
How many new words a day should I add?
Somewhere around 20 to 25 is a rate most people can hold while still clearing reviews. If your reviews are piling up, cut the new words before you cut the reviews — new cards are what creates the backlog.
Should I learn kanji readings in isolation?
Rarely. A reading learned as an isolated fact has nothing to attach to. Learn it attached to a word, and let the words accumulate. The exception is readings you meet constantly and keep forgetting.
Can I use a pre-made N2 deck?
Yes, but check it. Pre-made Japanese decks vary a lot, and errors in readings are the most common problem. If you use one, spot-check readings against a dictionary and be willing to delete cards.
Does WakeCards support Japanese documents?
Cards are produced in the same language as your source, so a Japanese document produces Japanese cards. There is no OCR, so the file needs selectable text. Check important readings against a dictionary, since generated cards are a first pass.