I have this Anki deck of 1478 notes that I use to practice reading, speaking, and listening. https://ankiweb.net/shared/info/3355859590 With one reading, one listening and one speaking card for each note, the deck turns into 1478*3 = 4434 actual flash cards.

The cards are all in hirigana and katakana and it makes it very hard to read when its more than ten kana. I want to convert all 1478 cards into kanji where appropriate.

For example, I would like this:

せんしゅう とうきょうで じしんが ありました

turned into this:


ideally with furigana too, inserted like this with Anki square brackets:


I think Anki automatically generates the furigana like I did in sentence 3 above, but only if you type each note in the actual Anki deck editor, manually.

If I cannot convert the kana, I should probably disable all my "reading" flash cards because it's pointless to read all kana. Arguably I would read kana faster, but it's very difficult and my energy would be better spent specializing in listening and speaking.

My first thought was to select all the "kana" entries in the Anki SQlite database directly and use some kanaToKanji() function. I could not find this function. I found one online but it didn't work with my test case above.

My second thought was to use the built-in Mac OS X Japanese input which converts kana to kanji and write a keyboard script that iterates through the actual Anki deck editor and types each sentence keystroke by keystroke. That way Anki would automatically insert furigana at the same time.

Is this the best approach?

  • 3
    Whichever automated way you can find to convert kana to kanji, it will be bad enough to consider not converting it at all.
    – Earthliŋ Mod
    Commented Mar 17, 2016 at 19:18
  • 1
    Ya, what @Earthliŋ said. There are simply far too many homophones in Japanese (i.e. words that are spelled identically in kana), which is a large part of why kanji are still used -- to make meanings clear. just せんしゅう alone could be spelled in kanji as 先週・専修・選集・泉州・千秋・千週・撰集・千周・撰修... Which one to use depends on context, and context is an enormously difficult programming problem. Any automated dumb-computer approach is thus doomed to lots of (hopefully amusing) failure. Commented Mar 17, 2016 at 20:54
  • 2
    Practically speaking, when you're used to reading in general it'll probably be easier with kanji, but a sentence like the one above should be easy to read in all kana, too. After all, you need to understand it when it's spoken, and no one speaks in kanji.
    – user1478
    Commented Mar 17, 2016 at 23:15

1 Answer 1


Try out Justsystem ATOK IME (free for 30 days) ... if you can read the installation instructions which is, unfortunately, available only in Japanese :-(

It is considered as the best IME for Japanese.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .