Full width and half width

How it works

1. Paste the text. The result updates as you type.
2. Choose the direction: to half width for a form that rejects full-width digits, to full width for a system that expects them.
3. Switch off the kinds of character you want left alone — converting the letters but not the katakana, for instance.
4. Copy the result.

Frequently asked questions

What is the difference between the two widths?

Japanese typesetting has two forms of the same characters: full width, which takes the square of a Japanese character (ＡＢＣ１２３), and half width, which is the ordinary Latin form (ABC123). Katakana exists in both too. Which one a system wants is arbitrary and rarely explained, which is why this conversion is an everyday chore in Japan.

Does it handle voiced katakana properly?

Yes, and this is where simple conversions go wrong. Half-width katakana writes バ as two characters — ﾊ followed by a voiced mark — while full width writes it as one. Going to full width joins them; going to half width splits them again. ﾊﾟ and the small forms are handled the same way.

Is my text uploaded?

No. It is converted in this page as you type. Names, addresses and customer records are exactly what these conversions are for, and none of it leaves the device.

Why is there a separate option for spaces?

Because the ideographic space (U+3000) is a character in its own right and is a character wide. A form that rejects it will not tell you why. Keeping it on turns it into an ordinary space going one way, and back going the other.

What is left alone?

Hiragana, kanji and anything outside the ranges being converted. Only the characters that genuinely have two widths are touched.