KanjiTomo is a program for identifying Japanese characters from images:

KanjiTomo main screen

Kanji lookup is done by pointing the mouse to any image on screen (either from a file, program or web page). Dictionary lookup is done at the same time. Horizontal and vertical text are both supported.

How it works?

Description of KanjiTomo's algorithm is here


The latest version of KanjiTomo can be downloaded here Unzip the package to any directory and double-click KanjiTomo.jar or launch.bat to run the program. If you are upgrading from previous version of KanjiTomo, you can use Settings -> Import settings to load your old configuration.

Java is required to run the program. It is recommended to install Java JDK instead of JRE, it will increase recognition speed by about 25%. KanjiTomo is a standalone program, Java browser plugin is not required and can be disabled.

KanjiTomo has been tested on Windows 7 and 10 operating systems. Other operating systems might also work but are not supported. Japanese fonts must be installed.

Release history

1.0.0 (2019-12-15)
- improved OCR quality and speed
- KanjiTomo's algorithm is now available as a Java library at GitHub

0.9.13 (2018-11-25)
- fixed issue that prevented launch with Java 11

0.9.12 (2016-01-24)
- improved text orientation detection

0.9.11 (2015-04-12)
- support for Chinese language

0.9.10 (2014-05-18)
- ability to save identified words to a list and export them 
  to file or clipboard 

0.9.9 (2013-06-17)
- zoom mode
- ability to click and drag to manually select
  word location (in file or zoom mode) 

0.9.8 (2013-02-10)
- dictionary for Japanese names
- improved detection of white text over complex background

0.9.7 (2012-11-25)
- clipboard support
- fullscreen mode
- two-page spread mode
- file history

0.9.5 (2012-10-07)
- four character compounds
- selectable text orientation
- selectable text color
- hotkeys
- faster startup speed

0.9.2 (2012-09-23) 
- initial release


When you first start KanjiTomo, no file has been opened and the program is in Automatic OCR mode. This means that you can point the cursor anywhere on the screen and the program tries to identify the image under cursor as Japanese text. This is useful if you are using web browser or other program to display images. If you have static images on your computer, it's best to open them directly with KanjiTomo from the file menu.

If the image has large characters or complex background (for example in title page), it might be necessary to manually mark the text. This can be done by click and drag with left mouse button over image (file mode) or zoom frame.

When the program identifies characters, they are marked by red squares. OCR Results panel now shows a list of matching characters for each character image. If the correct character is not at top of the list, you can select it manually. If too many characters are selected, click the selected row to clear it. You can also use hotkeys Alt+W and Alt+Q to add or remove characters from selection.

Selected characters are show in Search field. You can also input characters here yourself. All words starting with inputted characters are shown; add space to also search for ending words.

Search results panel shows all words in a dictionary that contain the text in search field. Click a word to show a detailed description in Word details. Hotkey Alt+Z can be used to copy the result to clipboard.

The dictionary used is Jim Breen's EDICT. Terms used in descriptions are explained in EDICT documentation.

Hotkeys can be used to control the program, see below for examples. Hotkeys must first be enabled by setting ENABLE_HOTKEYS=1 in config.txt file.

Control panel

File menu

Settings menu

Zoom frame

Zoom frame can be used to mark characters for OCR with click and drag. You can move the frame by dragging from title bar or with middle mouse button; red rectangle show the zoom target location. Zoom can be targeted to specific location by clicking the middle button over image (file mode) or with Alt+R. Middle button also closes the zoom frame.

Names dictionary

In addition to EDICT, KanjiTomo also includes ENAMDICT, a dictionary for Japanese Proper Names by Jim Breen. This dictionary can be accessed from Names button or by the hotkey alt+D.

List of saved words

You can save identified words to a list and later export them to file or clipboard. After a word has been selected, click Add button to add the word to the list. A dialog is shown where you can edit the details of the saved word. Hotkey Alt+1 or double-click in results list also opens the Add word dialog.
By default, words are saved or copied to clipboard in tab-delimited format; you can modify the save format by editing WORD_LIST_SAVE_FORMAT in config.txt file.

If Japanese characters are not displayed correctly after opening the saved file, make sure that your text editor supports UTF-8 encoding. If you are using Windows 7 operating system, try to open the file with Notepad instead of WordPad.

Chinese support

KanjiTomo has experimental support for Chinese language, it can be enabled by setting CHINESE_DICTIONARY=1 in config.txt. CHINESE_TRADITIONAL can be used to select between traditional and simplified characters.

Configuration file

Default settings can be changed by editing config.txt file. See the file comments for more information.


If you have any questions about the program, feel free to send an email to kanjitomo@gmail.com