- Ver 1.3.3 (released 2026-07-25)
- Please uninstall any previously installed version before installing.
- Older versions are available on the Releases page, and the version history is in CHANGELOG.md.
naver.github.io/d2-coding-font is a live specimen
with an editable playground, the complete ligature list read out of the font, and three tools for
pinning down a rendering problem before you report it: a size ladder from 8px to 24px, a coverage
check against the real cmap, and a report template that captures your settings and environment.
D2 Coding runs on Windows, macOS and Linux, and works with a wide range of development tools. The simplest install is a double-click: download a TTF from fonts/ttf or the release archive, open it, and press Install (Windows) or Install Font (macOS).
The terminal does it in one command, standard and ligature builds both, for the current user; the scripts are in install/.
Linux and macOS:
curl -fsSL https://github.com/naver/d2-coding-font/raw/master/install/install.sh | shWindows (PowerShell):
irm https://github.com/naver/d2-coding-font/raw/master/install/install.ps1 | iexEither way, restart the editor or terminal afterwards so it picks the new fonts up.
D2 Coding is based on Nanum Barun Gothic and was optimised for writing code: it improves legibility and the distinction between similar looking characters, while keeping the Latin design in harmony with Hangul. Characters that are easy to confuse while reading code, whether Latin letters, digits, Hangul or punctuation, were reworked to stay distinguishable. It is a monospaced typeface, so line and character spacing stay consistent in any development environment.
Latin letters, digits and similar looking Hangul symbols were drawn to be easy to tell apart while reading code. Hinting was tuned so that the design stays crisp between 8 and 18pt.
The first two rows of the capture are Noto Sans and DejaVu Sans Mono. On the live specimen page the table renders in your own browser's fonts instead, with D2Coding always in the third row.
Hangul is based on Nanum Barun Gothic, so the letterforms stay soft and blend naturally with the Latin design used for code.
The same snippet is rendered live on the specimen page.
Because it is monospaced, code lines up neatly in any development environment. Line and character spacing were tuned through an internal beta test with developers to improve the readability of code.
Sixteen Latin characters, eight Hangul syllables and eight ideographs all end on the same column. The grid above is the one the specimen page can draw over any sample.
1.3.3 maps 19,966 codepoints over 26,188 glyphs: all 11,172 Hangul syllables, Latin and 332 extended characters, 31 control pictures, and 4,620 CJK Unified Ideographs plus 268 compatibility ideographs.
Earlier versions of this README said Hanja was not included and would fall back to another font. That was never true of a released build. 1.0 already carried the same 4,620 ideographs, and they are drawn full width, so they render from D2Coding rather than from a fallback.
A Hangul syllable is exactly twice the advance width of a Latin character, so Korean comments and string literals keep source code, tables and box drawing output aligned. Hanja is full width too.
A separate ligature build adds programming ligatures for common operator sequences such as =>,
!=, <= and -> through the OpenType calt feature. Apart from that OpenType code it is
identical to the standard build: the two share the exact same outlines and metrics.
The original design sources have never been part of this repository. The typeface was drawn for
NAVER by FONTRIX in 2015, and on a project this old, tracking down who still holds the .glyphs or
.ufo files takes time. Rather than leave the project without anything to build from, the UFO
files under sources/ were reverse-engineered from the released binaries with
fontTools and
extractor. They carry the quadratic outlines exactly as
they are in the shipped TTFs, so they are a starting point for further work rather than the
original design data. See sources/README.md for details.
Anyone may use D2 Coding, and anyone may redistribute it under the terms of the SIL Open Font License.
D2 Coding is released under the OFL (SIL Open Font License), an internationally recognised license for open fonts. It places no restriction on use, and redistribution is allowed as long as the license is included. That covers bundling the font with other software, commercial software included.
The full license text is in OFL.txt, and the same text as it was written in the project wiki is in LICENSE.md. Until now it lived only in the wiki, which is what made the license hard to find.