JEDict supports several methods of kanji search:

Search by meaning or reading

In the Kanji Search field type any text (English or Japanese) containing a meaning, reading or kanji itself and press Enter key. The matching kanji characters will be displayed in a Kanji List Table.

In figure above, the Kanji Search field is located in top right corner and the Kanji List Table is at the bottom.

Search by kanji index

To find kanji by index, enter into the Kanji Search field a code of the corresponding index.
This code consists of a character prefix referring the index type and a decimal number standing for index value as shown in a table below.



Example for kanji 日


kanjidict index



Gakken index



Halpern index



Haig index



Nelson index



O'Neil index



SKIP code



Stroke number



Unicode (hexadecimal)


A triangle located to the left of search field opens a reminder popup menu listing all codes.

Search by kanji components

This is the most effective method of kanji search. 6355 kanji from the KANJIDIC are classified using 440 graphical components. These components (graphemes or pictograms) include 214 radicals and 226 non-radical elements and are grouped by stroke index.

To find composite kanji you need to specify one or more simpler kanji components, and JEDict will display all kanji containing chosen components.

To find appropriate kanji component (radical/grapheme) click on Grapheme stroke selector, and JEDict will display graphemes with given stroke count.

Once grapheme is found, drag it to the Grapheme filter. Alternatively you may send grapheme to Grapheme filter by double clicking on it.

If you choose List auto-update option, one click on graphemeis enough to send it to Grapheme filter. Each time the Grapheme filter is modified, a list of displayed kanji is updated correspondingly. Simultaneously the Grapheme selector is also updated to show only allowed graphemes, i.e. those graphemes that may appear together with selected grapheme combination.

If List auto-update box is unchecked, no filtering occurs during grapheme list editing. You compose the list of component by dragging graphemes or double clicking on graphemes. Single click in this case just expands the grapheme, i.e. displays kanji containing this grapheme in Kanji view, whereas double click send it to Grapheme filter. To initiate search of kanji containing chosen components, simply click on small search button near Grapheme filter.

If there is no kanji with selected graphemes combination, all views stay unchanged. A warning messsage is displayed at the bottom of the page.

Example. How to find kanji 捜 with activated Auto-update option.
In kanji 捜 three components are clearly distinguishable. Let starts with 才, To find this components, first click on button 3 in Grapheme stroke selector.
In view below a group of graphemes consisting of three strokes will be displayed.

To tell JEDict: "I want to find a kanji containing component 才", simply click on 才 pictogram (you may also drag and drop this component to any field of Grapheme filter).

Once chosen grapheme appears inside Grapheme filter, a Kanji view below will be filled with all kanji containing component 才.

The resulting list contains 231 kanji, i.e. the list is still big to extract the wanted kanji.
Let's do another step and define a second filter component 由. First let's click on button 5 to show five-strokes graphemes, then we can pick a 由 component.
There is only one kanji 捜 containing 才 and 由 components and this kanji is displayed as an output.
The searched kanji was found in four clicks!

In most cases two or thee components are well enough to find a kanji.

An important note. There are some graphemes that may look similar and can be easily confused especially in handwritting. For example kanji 田、甲、由 and 申. A fuzzy match check box specify if all similar graphemes should be treated as the same grapheme. In this case you receive a wider list of eight kanji all containing 才 grapheme and another grapheme similar to 田.

As default, the resulting kanji list is ordered by strokes, but you can choose another sorting criterion using sorting popup menu.

There are a Backspace button (<--) to delete last chosen component (or selected component) and a Clear button (x) to clear a whole filter.

Handwritten Kanji Recognition

JEDict implements an original handwritten kanji recognition engine providing a robust recognition of all 6355 kanji characters. An important feature - recognition does not depend on the stroke drawing order. The handwriting view is located inside Kanji Info view. When drawing a kanji, make sure you have no extra lines. Pressing the Find button will display the most relevant kanji. Beside the Find button, the view contains a Clear Last Stroke button and a Clear All button . A stroke count is shown in the top right corner.

Kanji Info View

A detailed information about a kanji (the one selected in Kanji List Table) is displayed in Kanji Info view.

This view contains three elements:

  • An animated SVG image of a kanji at the left.
  • The right area is for kanji handwriting. See the paragraph above.
  • The rest of this view displays a detailed kanji information. The displayed kanji indices may be chosen in Preferences. This view also contains the examples of the kanji usage taken from the most common words of EDICT dictionary.

    The two buttons at the button of SVG field are (1) to choose color/black&white drawing and (2) to play the stroke order animation. A triangle popup is to select the recently displayed kanji.
    The SVG kanji drawing is based on data from KanjiVG project by Ulrich Apel.

    [ First ] [ Previous ] [ Next ] [ Last ]