If you use WaniKani, you can probably recognize hundreds, maybe thousands, of kanji. Then you pick up a pen and freeze. This is not a failure of WaniKani, and it is not a failure of yours. It is exactly what WaniKani is built to do, and what it is not.
What WaniKani is great at
WaniKani is one of the best tools for reading. Its spaced-repetition system drills meaning and readings at scale, and its radical-based mnemonics make characters stick visually in a way few tools match. If your goal is to read Japanese, it earns its place, and a WaniKani user at a high level has genuinely impressive recognition. The point it does not address, on purpose, is handwriting. It was never sold as a writing app, so the gap you feel is not a bug.
Why recognition does not transfer to writing
Recognizing a kanji is matching a shape you are shown to a meaning. Writing one is producing the shape with no prompt at all. These are different skills, and they do not transfer as much as you would hope, which is why recognition is not recall, the idea at the centre of recognition versus recall. WaniKani trains recognition constantly, through every review. Your writing recall, meanwhile, gets almost no practice, so it stays weak while your reading races ahead. The gap widens the longer you study, and it is completely normal.
There is also a structural reason. To write a kanji you must retrieve its exact components, their arrangement, and the order of the strokes. To recognize it you only need a rough visual match. Writing asks for far more detail, and that detail is what fades first when it is never practised.
What WaniKani builds, and what writing adds
WaniKani is excellent at its half of the skill; writing covers the other half:
| Skill | WaniKani builds it? | Writing app builds it? |
|---|---|---|
| Recognize kanji and readings | Yes | No |
| Vocabulary through SRS | Yes | No |
| Write a kanji from memory | No | Yes |
| Stroke order | No | Yes |
Reading down the columns shows why the two pair so well: almost no overlap, and together they cover the whole skill.
Add a writing layer, keep WaniKani
The fix is not to drop WaniKani. It is to add the one thing it skips: producing the character by hand. A writing app for WaniKani users should:
- prompt you with the meaning and reading, then ask you to write from memory before any answer appears,
- offer a stroke-order guide on demand, so your hand learns the sequence rather than guessing,
- use spaced repetition aimed at writing, not just reading, so weak characters return sooner.
Kanji Write Practice is built for exactly this pairing. It is writing-first and free in early access. You keep WaniKani for reading, and add writing on top. If you also use flashcards, the same logic applies to Anki users, and if you want a free option comparison, see our free Skritter alternative guide.
A simple WaniKani plus writing routine
Because your recognition is already strong, the writing comes back faster than learning from zero. A workable daily loop:
- Do your WaniKani reviews as usual.
- Take a small handful of characters you already recognize, five is plenty, and write each from memory from a meaning and reading prompt.
- Reveal a stroke-order guide only when stuck, then write the character again without it.
- Let the ones you miss come back sooner, and let the clean ones drift to longer intervals.
A few minutes of writing a day is enough, because one act of recall builds more durable memory than several recognition reviews, the testing effect again. Producing a character by hand also strengthens the recognition WaniKani is building, so the two are not competing for your time. For the bigger picture, see why you recognize kanji but cannot write it.
A worked example
Take a kanji you have burned in WaniKani, say the character for language. Your recognition of it is solid, you read it without thinking. Now cover it and try to write it. Most likely you remember it has the speech radical on the left and something on the right, but the something is fuzzy. That fuzziness is the writing gap in miniature: you stored enough to recognize the character, not enough to produce it. The fix is to rebuild it as parts you can produce, the speech radical, then five over mouth on the right. Write it that way from memory three times over three days and the fuzzy half resolves into something your hand knows. Do that with ten characters a week and the gap closes quickly, precisely because your recognition gave you a running start. This is why WaniKani users tend to add writing faster than absolute beginners: the meaning and reading are already in place, and only the production needs building.
