tl;dr at the bottom. The beginning is explaining why I need something different than what I've found.

The languages I normally use on my computer are English and Greek. However, sometimes I want to type in Japanese as well, without a Japanese keyboard.

On Windows this works well, as IME is just another keyboard layout. My primary OS is Linux though (specifically Mint KDE 17.1) and there are problems. It seems that the only way to use a similar method there is something that replaces the entire keyboard layout system, such as ibus or SCIM.

The above systems work well enough if I just want to use English and Japanese, but adding Greek to the mix causes more issues, as stresses stop working normally (they should work with dead keys, such as '+α producing " ά ", but instead if I use ibus or SCIM the result is " 'α ". Also, wanting to use a compose key is another problem, as the implementation depends on the default keyboard layout system.

I've tried ibus, SCIM and fcitx but none of the systems worked well with all 3 languages and the compose key. So I'm wondering if anything exists that would do the job.


Is there a piece of software that can turn Romaji into Japanese (primarily Hiragana and Katakana) in one of the following 3 formats?

  1. A Linux program, such as a simple window one can type into. Something that works under Wine or Mono would do the job as well.
  2. A website that can do it.
  3. An input method for Linux that implements Anthy or another backend and does NOT replace XKB (as ibus, SCIM etc. do) but can be enabled and disabled on-the-fly, while I keep using XKB for the other languages.

4 Answers 4


DuckDuckGo has instant answer feature for that

Merged pull request of this feature

Example of DuckDuckGo conversion of romaji to hiragana


I've written a few sed-based filters for this purpose.

Example converting Romaji to Kana:

romaji2kana <<< arigatou

More binary alternatives i know:

  • uconv -x 'Any-Hiragana' or uconv -x 'Any-Katakana'
  • kakasi (untested by me)
  • mecab (untested by me)

There is a perl module that seems to be capable of this - Lingua::JA::Kana. This alone would of course require some perl programming, but you may be able to find a premade script or program that utilizes this module. If I come across anything I'll be sure to post back.


Since you mentioned a website, you can try jisho.org

curl -s https://jisho.org/search/ro-maji | grep "Searched for"

Will give you:

            Searched for <strong><span class="hilite_1">ろーまじ</span>.</strong>

which can be processed by a regex to find ろーまじ.

Obviously this is vulnerable to any change on the website format.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.