Unicode Hanja IME Guide

Open the app — Unicode Hanja IME

Have you ever been transcribing an old text and come to a full stop at a character you cannot read?
Korean input converts hanja from its reading, so without the reading you cannot even open the conversion list.
And once you have dug the reading out of a dictionary, the character is often missing from the list anyway.

The built-in converter on Windows mostly covers the basic block.
Characters from the extension blocks — common in classical texts — never show up.

This IME changes two things.
You can enter a character by breaking its shape into radicals, no reading required, and all 93,189 Unicode hanja are offered as candidates.
It also holds 251,751 words, so modern material flows just as easily.

How to Use

The screen, and the basic loop

The screen splits in two. The left is the editor where your text collects; the right holds the input box and the candidate list.

1. Type Hangul into the box on the right — reading, meaning-plus-sound, or radicals, all in the same box.
2. Candidate characters appear ten at a time.
3. Press 0–9 and that character drops into the editor on the left.
4. When you are done, [Copy] puts it on the clipboard and [Save] downloads a .txt file.

There is a reason the two panes are separate.
While you are picking characters, symbols, digits, spaces and line breaks go straight into the left pane.
So you are not just collecting characters — you can type a whole document start to finish.

The Jaechun method — entering what you cannot read

The Jaechun method calls a character by the readings of the radicals it is made of.
If the character is in front of you, you know its shape even when you don’t know its sound, so you spell out the radicals you recognize.

Each radical has a fixed name. It feels odd at first, but half a dozen common ones cover most of what you meet.
The full list lives at Jaechun radical list.

Searching by meaning and sound

For a character you do know, typing its gloss and reading together is faster.

The reading alone works too: type 휴 and every character read that way is listed.

Wildcards — when you only know some of them

A forgotten radical is not a dead end. Leave the gap as a symbol.

? and * never appear in the left pane — they exist only for searching.

Entering whole words

With [Include words] checked above the input box, multi-character words join the candidates.
Type 가감 and you get 加減 · 可堪 · 可鑑 as complete words — far quicker than one character at a time.

Candidates list single characters first, words after, because looking for one character is the more common case.
Uncheck the box to see single characters only.

How candidates are ordered

A single reading can carry dozens of characters, so ordering happens in two steps.

1. Exact matches for what you typed come first, then entries that merely start with it.
2. Within each group, order follows the Unicode blocks: basic, then Extension A, compatibility, then Extension B and beyond.

Common characters live in the earlier blocks, so you usually finish on the first page.

Keyboard only

Nothing here requires reaching for the mouse.

To fix a typo in the right-hand box, Esc is cleaner than Backspace — backspace deletes from the left-hand editor.

Symbols common in old texts get their own shortcuts.

Display and Settings

Extension-block badges

The small mark beside a candidate is the Unicode block that character lives in.
Two characters can look alike and still differ in whether your screen can draw them.

The badge exists to warn you in advance about characters that look fine here and turn into empty boxes somewhere else.
If a red-badged character is going into a document you will send to someone, spare a thought for their setup.

The Unicode blocks covered

Hanja are not gathered in one place in Unicode. Each time more characters were encoded, another block was appended.
The C or F on a badge is the name of one of these blocks.

For the details of each block, see the Character Code Charts.

What the IME actually holds is 93,189 characters.
That is fewer than the table totals, because Extensions G, H and I are not included yet.
Characters from those blocks will never appear as candidates, so keep that in mind if a search comes up empty.
The other blocks also fall slightly short of their full size, since not every code point in a block is assigned.

Counting characters

The toolbar shows total characters and hanja characters separately.
The total counts spaces and symbols; the second counts only hanja.
Useful when you have a length to hit.

View settings

• Font: switch between sans and serif. Serif renders hanja strokes more distinctly.
• Theme: light or dark.
• Text size: + and −. Dense extension characters are much easier to verify enlarged.

Settings persist in the browser and are still there next time.

Good to Know

Your text is kept

The left-hand editor saves itself to the browser automatically.
Close the tab by accident and it is still there when you return.

But it lives inside that browser, so it does not follow you to another device.
For anything you will work on for a while, take a .txt copy with [Save] now and then.

When characters show as boxes

If something you picked appears as □ or blank, the character went in correctly — your system simply has no font that draws it.
This is most common from Extension C onward.
Installing a font with wide extension coverage, such as Source Han Serif, resolves most cases.

Copying to another program still carries the character itself; only its appearance follows whatever fonts that environment has.

Where it runs

A browser is all you need — nothing to install, no account.
After the data loads once, everything runs locally and nothing you type is sent to a server.

On a phone

It runs on a phone just the same; only the layout changes to fit the narrow screen.

The screen splits top and bottom rather than left and right — the editor above, the input box and candidate list below.
To leave more room for candidates, the editor is squeezed to a single line, and the input box sits on one line with its label and the [Include words] checkbox.
Only one line of the editor shows at a time, but nothing is lost — just scroll it up and down.

You pick candidates by tapping them, in place of the number keys.
On a tall screen the candidates stack ten deep with their meanings, as on the desktop.
On a short one they fold into two columns of five, and you page through them with ◀ ▶ beside the list.

Settings live under ☰ at the top right.

Space and Enter on the Hangul keyboard behave exactly as they do on the desktop:
Space inserts what you typed as Hangul rather than converting it, and Enter inserts the selected character.

Goes well with

• To look up what a character you just entered means, try the Korean Studies Web Dictionary.
• To convert a date from the material you are transcribing, there is the East Asian Chronology.
• If an existing document mixes variant forms of the same character, Hanzi Normalization will unify them.
• To dock all three beside your browser, there is the Korean Studies Toolkit Chrome extension — select a word and look it up in place.

Open the app — Unicode Hanja IME