font-subset

font-subset will create a subset of your font, including only the glyphs needed to render the characters you define.

Limitations

These limitations may change in the future. Star https://github.com/jqpe/font-subset for updates!

• Only characters in the Basic Multilingual Plane (BMP) are supported. This means subsetting emoji is often not possible. Popular icon fonts use the E000–F8FF private use area, so subsetting them is possible.
• Only a single font can be processed at a time. If multiple files were dropped, the last will be used.
• All variable axes will be kept as in the original.
• The font is automatically converted to Woff2 format.
• Very large fonts might not work. Tested up to a 2.5MB font. A Woff2 font will need to be decompressed for subsetting, so its actual size might be up to 5x the Woff2 size.
• Modifier symbols currently do not get special treatment, entering ¨ + o to create ö will include both ¨ and ö in the subset input. You can click a character in the input preview grid to exclude it.

Usage

1. Define the charactes you want to keep. You might find it beneficial to start with unicode blocks (see https://unicode.link/blocks for example) and then use the grid to trim the input down.
   1. Do not put U+ or anything else before the Unicode ranges (or a single Unicode codepoint)
   2. 0x00FF = 0xFF, but 0xE000 != 0xe0
   3. Syntax: 0-7f, fffd, 80-ff will include Basic Latin, the Replacement character (�), and Latin-1 Supplement. You can also exclude characters or ranges by prefixing with !: 0-7f, !20-2f, !60 for Basic Latin without punctuation and grave accent.
   4. Control characters are excluded by default.
2. Drop a font.
3. You can now use the editor (supports basic markdown shortcuts like # for a heading) to inspect your font, play with the variations and most importantly, download the subset.

License

Source available for now, Copyright Jasper Nykänen 2025.