OpenType feature and script tags

The ScriptList and FeatureList of the GSUB and GPOS tables name each script, language system and feature with a four-byte tag. Software recognises a feature only by its tag, so the tags are registered: Microsoft keeps the registry as part of the OpenType specification. This page lists every registered feature tag and script tag, and the three feature parameter tables. The language system tags have their own page, and the baseline tags are on the BASE page.

How tags are written

A Tag is four bytes, each a printable ASCII character (0x20 to 0x7E). A shorter name is padded on the right with spaces: the script tag for Lao is the bytes 6C 61 6F 20 (lao ).
Registered feature tags and script tags are four lower-case letters (with a few padded or numbered exceptions, such as yi , dev2 and cv01).
Registered language system tags are three or four capital letters, padded with a space when there are three.
Tags of four capital letters A to Z, with no digits, spaces or punctuation, are left free for font vendors’ private features (for example a vendor’s own PKRN). They are not registered and need not work the same way in other software.
DFLT is the script tag for features that do not depend on the script. As language system tags, dflt and DFLT are reserved and must not appear in a LangSysRecord: the default language system is the Script table’s own defaultLangSysOffset.

Registered feature tags

Each feature is applied by running the lookups its Feature table lists. The lookup types below are the ones the registry recommends; a font may use others. “GSUB 1” means GSUB lookup type 1 (single substitution) and so on: the types are listed on the GSUB and GPOS pages. Which features an application turns on, and in what order, is up to the application and the shaping rules for the script; the feature tag only says what the lookups are for.

Tag	Name	Lookups	What it does
`aalt`	Access All Alternates	GSUB 3, 1	Gives the user every alternate form of a character in one place, such as a glyph palette.
`abvf`	Above-base Forms	GSUB 1	Uses the above-base part of a split vowel (Khmer and other Brahmi-derived scripts).
`abvm`	Above-base Mark Positioning	GPOS 4, 5	Places marks above base glyphs.
`abvs`	Above-base Substitutions	GSUB 4	Replaces a base and the mark above it with a ligature.
`afrc`	Alternative Fractions	GSUB 4	Forms vertical (nut) fractions from digits separated by a slash.
`akhn`	Akhand	GSUB 4	Forms the akhand ligatures of Indic scripts, whatever surrounds them.
`apkn`	Kerning for Alternate Proportional Widths	GPOS 2	Kerns glyphs that are fixed-width by default but have been made proportional by `palt`.
`blwf`	Below-base Forms	GSUB 4	Uses the below-base form of a consonant in Indic conjuncts.
`blwm`	Below-base Mark Positioning	GPOS 4, 5	Places marks below base glyphs.
`blws`	Below-base Substitutions	GSUB 4	Replaces a base and a below-base form with a ligature.
`calt`	Contextual Alternates	GSUB 6	Swaps glyphs for alternates that join or fit better with their neighbours.
`case`	Case-sensitive Forms	GSUB 1 or GPOS 1	Moves punctuation up to suit capitals and lining figures, and changes oldstyle figures to lining.
`ccmp`	Glyph Composition / Decomposition	GSUB 4, 2	Composes several characters into one glyph or splits one glyph into several before other features run.
`cfar`	Conjunct Form After Ro	GSUB 1	Khmer: uses alternate below-base or post-base forms after a subscript Ro.
`chws`	Contextual Half-width Spacing	GPOS 2, 8	Halves the spacing of full-width CJK punctuation according to its neighbours.
`cjct`	Conjunct Forms	GSUB 4	Forms Indic consonant conjuncts, later in shaping than `akhn`.
`clig`	Contextual Ligatures	GSUB chained context	Forms ligatures only where the surrounding glyphs call for them.
`cpct`	Centered CJK Punctuation	GPOS 1	Centres certain punctuation in fonts without separate centred forms.
`cpsp`	Capital Spacing	GPOS 1	Adds space between capitals for all-capital text.
`cswh`	Contextual Swash	GSUB chained context, 3	Uses swash forms where the context suits them.
`curs`	Cursive Positioning	GPOS 3	Joins cursive glyphs by their entry and exit points.
`cv01`–`cv99`	Character Variant 1 to 99	GSUB 3 (or 1)	Selects variants of one character, or a few closely related ones; may carry feature parameters.
`c2pc`	Petite Capitals From Capitals	GSUB 1	Turns capitals into petite capitals.
`c2sc`	Small Capitals From Capitals	GSUB 1	Turns capitals into small capitals.
`dist`	Distances	GPOS 1, 2, contextual	Required spacing between glyphs that the user should not switch off (unlike `kern`).
`dlig`	Discretionary Ligatures	GSUB 4	Decorative ligatures the user may choose.
`dnom`	Denominators	GSUB 1	Uses denominator figures after a slash; applied with `frac`.
`dtls`	Dotless Forms	GSUB 1	Gives dotless forms of mathematical i and j for placing accents (math layout).
`expt`	Expert Forms	GSUB 1	Japanese forms preferred by typographers, extending the JIS78 set.
`falt`	Final Glyph on Line Alternates	GSUB 1	Swaps the last glyph on a line for a wider or narrower form to help justification.
`fin2`	Terminal Forms #2	GSUB 1	Syriac: final Alaph after a non-joining letter other than Dalath and Rish.
`fin3`	Terminal Forms #3	GSUB 5	Syriac: final Alaph after Dalath, Rish or dotless Dalath-Rish.
`fina`	Terminal Forms	GSUB 1 (5, 6, 8)	Uses final forms of joining letters, chosen by Unicode joining type.
`flac`	Flattened Accent Forms	GSUB 1	Flatter accents over tall bases in mathematics.
`frac`	Fractions	GSUB 1, 4	Turns digits separated by a slash into a diagonal fraction.
`fwid`	Full Widths	GSUB 1 or GPOS 1	Sets glyphs on full (usually em) widths.
`half`	Half Forms	GSUB 4	Uses the half forms of Indic consonants.
`haln`	Halant Forms	GSUB 4	Uses the precomposed halant forms of Indic consonants.
`halt`	Alternate Half Widths	GPOS 1	Fits full-width glyphs into half-em widths without changing the glyph.
`hist`	Historical Forms	GSUB 1	Uses historical letter forms, such as long s.
`hkna`	Horizontal Kana Alternates	GSUB 1	Kana designed for horizontal setting only.
`hlig`	Historical Ligatures	GSUB 4	Uses historical ligatures.
`hngl`	Hangul	GSUB 1, 3	Deprecated. Replaced hanja with hangul.
`hojo`	Hojo Kanji Forms	GSUB 1	Uses the JIS X 0212-1990 forms where the JIS X 0213:2004 form is the default.
`hwid`	Half Widths	GSUB 1 or GPOS 1	Sets glyphs on half-em widths.
`init`	Initial Forms	GSUB 1 (5, 6, 8)	Uses initial forms of joining letters.
`isol`	Isolated Forms	GSUB 1 (5, 6, 8)	Uses isolated forms of joining letters.
`ital`	Italics	GSUB 1	Uses the italic glyphs held in the same font (mostly Latin in CJK fonts).
`jalt`	Justification Alternates	GSUB 1	Wider or narrower alternates to help justification.
`jp78`	JIS78 Forms	GSUB 1, 3	Uses the glyph forms of JIS C 6226-1978.
`jp83`	JIS83 Forms	GSUB 1	Uses the glyph forms of JIS X 0208-1983.
`jp90`	JIS90 Forms	GSUB 1	Uses the glyph forms of JIS X 0208-1990.
`jp04`	JIS2004 Forms	GSUB 1	Uses the glyph forms of JIS X 0213:2004.
`kern`	Kerning	GPOS 2, 8	Adjusts the space between glyph pairs in horizontal text.
`lfbd`	Left Bounds	GPOS 1	Aligns line starts by the glyphs’ visible left edge.
`liga`	Standard Ligatures	GSUB 4	Ligatures for normal text, such as ﬁ and ﬂ; usually on by default.
`ljmo`	Leading Jamo Forms	GSUB 1	Hangul: the form of a leading consonant jamo inside a syllable.
`lnum`	Lining Figures	GSUB 1	Uses lining figures.
`locl`	Localized Forms	GSUB 1	Uses the forms preferred for the text’s language, such as Bulgarian Cyrillic.
`ltra`	Left-to-right Alternates	GSUB 1	Variants for left-to-right runs, other than mirrored forms.
`ltrm`	Left-to-right Mirrored Forms	GSUB 1	Mirrored forms for left-to-right runs.
`mark`	Mark Positioning	GPOS 4, 5	Places marks on base glyphs and ligatures.
`med2`	Medial Forms #2	GSUB 1	Syriac: medial Alaph after a joining letter.
`medi`	Medial Forms	GSUB 1 (5, 6, 8)	Uses medial forms of dual-joining letters.
`mgrk`	Mathematical Greek	GSUB 1	Uses the mathematical forms of Greek letters.
`mkmk`	Mark to Mark Positioning	GPOS 6	Places marks on other marks.
`mset`	Mark Positioning via Substitution	GSUB 5	Old Arabic fonts for Windows 95: positions marks by swapping glyphs. Not for new fonts.
`nalt`	Alternate Annotation Forms	GSUB 1, 3	Circled, boxed, bracketed and similar forms.
`nlck`	NLC Kanji Forms	GSUB 1	Uses the kanji forms defined by Japan’s National Language Council in 2000.
`nukt`	Nukta Forms	GSUB 4	Forms Indic consonants with nukta.
`numr`	Numerators	GSUB 1	Uses numerator figures before a slash, and a fraction slash; applied with `frac`.
`onum`	Oldstyle Figures	GSUB 1	Uses oldstyle figures.
`opbd`	Optical Bounds	GPOS 1	Deprecated. Aligned lines by visible edges, through `lfbd` and `rtbd`.
`ordn`	Ordinals	GSUB 6, 4	Uses ordinal letters after figures, and the numero sign for “No”.
`ornm`	Ornaments	GSUB 3, 1	Gives ornaments, either as alternates of the bullet or in place of plain letters.
`palt`	Proportional Alternate Widths	GPOS 1	Fits full-width glyphs onto proportional widths without changing the glyph.
`pcap`	Petite Capitals	GSUB 1	Turns lower-case letters into petite capitals.
`pkna`	Proportional Kana	GSUB 1	Uses proportional-width kana.
`pnum`	Proportional Figures	GSUB 1	Uses figures of individual widths.
`pref`	Pre-base Forms	GSUB 4	Uses the pre-base form of a consonant.
`pres`	Pre-base Substitutions	GSUB 4, 5	Ligatures and vowel variants before the base.
`pstf`	Post-base Forms	GSUB 4	Uses the post-base form of a consonant.
`psts`	Post-base Substitutions	GSUB 4	Replaces a base and a post-base form with a ligature.
`pwid`	Proportional Widths	GSUB 1	Uses proportional glyphs in place of fixed-width ones.
`qwid`	Quarter Widths	GSUB 1 or GPOS 1	Sets glyphs on quarter-em widths.
`rand`	Randomize	GSUB 3	Chooses among alternates at random, to imitate handwriting.
`rclt`	Required Contextual Alternates	GSUB 6	Contextual alternates that the script needs and the user cannot turn off.
`rkrf`	Rakar Forms	GSUB 4	Devanagari and Gujarati: consonant with rakar.
`rlig`	Required Ligatures	GSUB 4	Ligatures the script needs, such as Arabic lam-alef.
`rphf`	Reph Form	GSUB 4	Uses the reph form of Ra plus virama.
`rtbd`	Right Bounds	GPOS 1	Aligns line ends by the glyphs’ visible right edge.
`rtla`	Right-to-left Alternates	GSUB 1	Variants for right-to-left runs, other than mirrored forms.
`rtlm`	Right-to-left Mirrored Forms	GSUB 1	Mirrored forms for right-to-left runs beyond Unicode character mirroring.
`ruby`	Ruby Notation Forms	GSUB 1	Small kana designed for ruby annotations.
`rvrn`	Required Variation Alternates	GSUB 1	In variable fonts, swaps glyphs for some instances through a FeatureVariations table; always applied.
`salt`	Stylistic Alternates	GSUB 1, 3	Purely decorative alternate designs.
`sinf`	Scientific Inferiors	GSUB 1	Small figures below the baseline, as in chemical formulae.
`size`	Optical size	none (GPOS)	Holds the design size and size range as feature parameters; no lookups. Superseded by the STAT table.
`smcp`	Small Capitals	GSUB 1	Turns lower-case letters into small capitals.
`smpl`	Simplified Forms	GSUB 1	Uses simplified Chinese or Japanese forms.
`ss01`–`ss20`	Stylistic Set 1 to 20	GSUB 1 (GPOS possible)	Turns on a designed set of alternates; may carry feature parameters.
`ssty`	Math Script-style Alternates	GSUB 3 (or 1)	Forms for subscripts and superscripts in mathematics; alternate 1 or 2 for the script level.
`stch`	Stretching Glyph Decomposition	GSUB 2	Splits an enclosing mark, such as the Syriac abbreviation mark, into an odd number of pieces that can stretch.
`subs`	Subscript	GSUB 1, then GPOS 1 or 2	Uses subscript forms.
`sups`	Superscript	GSUB 1	Uses superior figures and letters.
`swsh`	Swash	GSUB 1, 3	Uses swash forms.
`titl`	Titling	GSUB 1	Uses forms designed for titles at large sizes.
`tjmo`	Trailing Jamo Forms	GSUB 1	Hangul: the form of a trailing consonant jamo inside a syllable.
`tnam`	Traditional Name Forms	GSUB 1	Traditional kanji forms used in personal names.
`tnum`	Tabular Figures	GSUB 1	Uses figures that all have the same width.
`trad`	Traditional Forms	GSUB 1, 3	Uses traditional Chinese or Japanese forms.
`twid`	Third Widths	GSUB 1 or GPOS 1	Sets glyphs on third-em widths.
`unic`	Unicase	GSUB 1	Maps both cases to a single-case alphabet.
`valt`	Alternate Vertical Metrics	GPOS 1	Centres full-width Latin, Greek and Cyrillic glyphs for vertical setting. Not needed in fonts with `vmtx` or `VORG`.
`vapk`	Kerning for Alternate Proportional Vertical Metrics	GPOS 2	Vertical kerning of glyphs made proportional by `vpal`.
`vatu`	Vattu Variants	GSUB 4	Ligatures of a consonant or half form with a following vattu form.
`vchw`	Vertical Contextual Half-width Spacing	GPOS 2, 8	Vertical counterpart of `chws`.
`vert`	Vertical Alternates	GSUB 1	Forms for upright glyphs in vertical writing, such as rotated brackets.
`vhal`	Alternate Vertical Half Metrics	GPOS 1	Fits full-height glyphs into half-em heights.
`vjmo`	Vowel Jamo Forms	GSUB 1	Hangul: the form of a vowel jamo inside a syllable.
`vkna`	Vertical Kana Alternates	GSUB 1	Kana designed for vertical setting only.
`vkrn`	Vertical Kerning	GPOS 2, 8	Adjusts the space between glyph pairs in vertical text.
`vpal`	Proportional Alternate Vertical Metrics	GPOS 1	Fits full-height glyphs onto proportional heights.
`vrt2`	Vertical Alternates and Rotation	GSUB 1	Pre-rotated glyphs for vertical writing; a superset of `vert`.
`vrtr`	Vertical Alternates for Rotation	GSUB 1	Forms for glyphs that the layout engine will turn sideways in vertical writing.
`zero`	Slashed Zero	GSUB 1	Uses a zero with a slash.

Feature parameter tables

A Feature table’s featureParamsOffset is NULL except for three kinds of feature, which may point to a parameter table. The offset counts from the start of the Feature table. All values are big-endian.

size

The size feature lives in GPOS, lists no lookups, and must have parameters: five uint16 values. Every font of a family that uses it must have it.

Offset	Type	Name	Meaning
`0`	`uint16`	Design size	The point size the font is designed for, in decipoints (720 to the inch, so 100 is 10 points). Never 0. If there is no size range, the other four values are 0.
`2`	`uint16`	Subfamily identifier	An arbitrary number shared by the fonts of one family that differ only in size range. 0 means the rest is to be ignored.
`4`	`uint16`	Subfamily name ID	A name table ID, 256 to 32767, holding the menu name for the subfamily (in English for Windows and Macintosh, and optionally other languages).
`6`	`uint16`	Range start	The small end of the recommended size range, in decipoints, exclusive.
`8`	`uint16`	Range end	The large end of the range, in decipoints, inclusive. Ranges in a subfamily must not overlap.

The registry’s example is Minion Pro Semibold Condensed Subhead, with the values 180, 3, 257, 139, 240: designed for 18 points, subfamily 3 named by name ID 257, and the recommended choice above 13.9 points up to 24 points. Given a requested size, software picks the subfamily member whose range holds it, else the one with the closest range; a size exactly between two ranges takes the larger.

ss01 to ss20

Offset	Type	Name	Meaning
`0`	`uint16`	`version`	0. Later minor versions may add data at the end.
`2`	`uint16`	`uiLabelNameID`	`name` table ID of the label to show for this stylistic set (expected in 256 to 32767).

cv01 to cv99

Offset	Type	Name	Meaning
`0`	`uint16`	`format`	0.
`2`	`uint16`	`featUiLabelNameId`	`name` ID of the feature’s label, or 0 for none.
`4`	`uint16`	`featUiTooltipTextNameId`	`name` ID of tooltip text, or 0.
`6`	`uint16`	`sampleTextNameId`	`name` ID of sample text showing the effect, or 0.
`8`	`uint16`	`numNamedParameters`	How many variants have their own label (may be 0).
`10`	`uint16`	`firstParamUiLabelNameId`	`name` ID of the first variant’s label; the labels use consecutive IDs. 0 if there are none.
`12`	`uint16`	`charCount`	Number of characters listed next (may be 0).
`14`	`uint24` × count	`character`	Unicode scalar values of the characters this feature gives variants for; for information only.

Variant n (counting from 1, the index into the type 3 alternate set) is labelled by name ID firstParamUiLabelNameId + n − 1. The registry recommends labels of at most 25 characters and tooltips of at most 250.

Registered script tags

Script tags mostly follow Unicode scripts, but they are not the same as ISO 15924 codes, and some scripts have two tags. The Indic scripts each have a second tag ending in 2 (dev2, bng2 and so on): a font with that ScriptRecord is shaped with the newer Indic model, which reorders the syllable differently, and a font may hold both. Hiragana and Katakana share kana.

Tag	Script	Tag	Script
`adlm`	Adlam	`ahom`	Ahom
`hluw`	Anatolian Hieroglyphs	`arab`	Arabic
`armn`	Armenian	`avst`	Avestan
`bali`	Balinese	`bamu`	Bamum
`bass`	Bassa Vah	`batk`	Batak
`beng`	Bangla	`bng2`	Bangla v.2
`berf`	Beria Erfe	`bhks`	Bhaiksuki
`bopo`	Bopomofo	`brah`	Brahmi
`brai`	Braille	`bugi`	Buginese
`buhd`	Buhid	`byzm`	Byzantine Music
`cans`	Canadian Syllabics	`cari`	Carian
`aghb`	Caucasian Albanian	`cakm`	Chakma
`cham`	Cham	`cher`	Cherokee
`chrs`	Chorasmian	`hani`	CJK Ideographic
`copt`	Coptic	`cprt`	Cypriot Syllabary
`cpmn`	Cypro-Minoan	`cyrl`	Cyrillic
`DFLT`	Default (no particular script)	`dsrt`	Deseret
`deva`	Devanagari	`dev2`	Devanagari v.2
`diak`	Dives Akuru	`dogr`	Dogra
`dupl`	Duployan	`egyp`	Egyptian Hieroglyphs
`elba`	Elbasan	`elym`	Elymaic
`ethi`	Ethiopic	`gara`	Garay
`geor`	Georgian	`glag`	Glagolitic
`goth`	Gothic	`gran`	Grantha
`grek`	Greek	`gujr`	Gujarati
`gjr2`	Gujarati v.2	`gong`	Gunjala Gondi
`guru`	Gurmukhi	`gur2`	Gurmukhi v.2
`gukh`	Gurung Khema	`hang`	Hangul
`jamo`	Hangul Jamo (not recommended; use `hang`)	`rohg`	Hanifi Rohingya
`hano`	Hanunoo	`hatr`	Hatran
`hebr`	Hebrew	`kana`	Hiragana and Katakana
`armi`	Imperial Aramaic	`phli`	Inscriptional Pahlavi
`prti`	Inscriptional Parthian	`java`	Javanese
`jurc`	Jurchen	`kthi`	Kaithi
`knda`	Kannada	`knd2`	Kannada v.2
`kawi`	Kawi	`kali`	Kayah Li
`khar`	Kharosthi	`kits`	Khitan Small Script
`khmr`	Khmer	`khoj`	Khojki
`sind`	Khudawadi	`krai`	Kirat Rai
`lao`	Lao	`latn`	Latin
`lepc`	Lepcha	`limb`	Limbu
`lina`	Linear A	`linb`	Linear B
`lisu`	Lisu (Fraser)	`lyci`	Lycian
`lydi`	Lydian	`mahj`	Mahajani
`maka`	Makasar	`mlym`	Malayalam
`mlm2`	Malayalam v.2	`mand`	Mandaic
`mani`	Manichaean	`marc`	Marchen
`gonm`	Masaram Gondi	`math`	Mathematical text layout (for math layout engines and the MATH table)
`medf`	Medefaidrin	`mtei`	Meitei Mayek
`mend`	Mende Kikakui	`merc`	Meroitic Cursive
`mero`	Meroitic Hieroglyphs	`plrd`	Miao
`modi`	Modi	`mong`	Mongolian
`mroo`	Mro	`mult`	Multani
`musc`	Musical Symbols	`mymr`	Myanmar (not recommended; use `mym2`)
`mym2`	Myanmar v.2	`nbat`	Nabataean
`nagm`	Nag Mundari	`nand`	Nandinagari
`newa`	Newa	`talu`	New Tai Lue
`nko`	N’Ko	`nshu`	Nüshu
`hmnp`	Nyiakeng Puachue Hmong	`orya`	Odia
`ory2`	Odia v.2	`ogam`	Ogham
`olck`	Ol Chiki	`onao`	Ol Onal
`ital`	Old Italic	`hung`	Old Hungarian
`narb`	Old North Arabian	`perm`	Old Permic
`xpeo`	Old Persian Cuneiform	`sogo`	Old Sogdian
`sarb`	Old South Arabian	`orkh`	Old Turkic (Orkhon Runic)
`ougr`	Old Uyghur	`osge`	Osage
`osma`	Osmanya	`hmng`	Pahawh Hmong
`palm`	Palmyrene	`pauc`	Pau Cin Hau
`phag`	Phags-pa	`phnx`	Phoenician
`pcun`	Proto-Cuneiform	`phlp`	Psalter Pahlavi
`rjng`	Rejang	`runr`	Runic
`samr`	Samaritan	`saur`	Saurashtra
`seal`	Seal (Small Seal)	`shrd`	Sharada
`shaw`	Shavian	`sidd`	Siddham
`sidt`	Sidetic	`sgnw`	Sign Writing
`sinh`	Sinhala	`sogd`	Sogdian
`sora`	Sora Sompeng	`soyo`	Soyombo
`xsux`	Sumero-Akkadian Cuneiform	`sund`	Sundanese
`sunu`	Sunuwar	`sylo`	Syloti Nagri
`syrc`	Syriac	`tglg`	Tagalog
`tagb`	Tagbanwa	`tale`	Tai Le
`lana`	Tai Tham (Lanna)	`tavt`	Tai Viet
`tayo`	Tai Yo	`takr`	Takri
`taml`	Tamil	`tml2`	Tamil v.2
`tnsa`	Tangsa	`tang`	Tangut
`telu`	Telugu	`tel2`	Telugu v.2
`thaa`	Thaana	`thai`	Thai
`tibt`	Tibetan	`tfng`	Tifinagh
`tirh`	Tirhuta	`todr`	Todhri
`tols`	Tolong Siki	`toto`	Toto
`tutg`	Tulu-Tigalari	`ugar`	Ugaritic Cuneiform
`vai`	Vai	`vith`	Vithkuqi
`wcho`	Wancho	`wara`	Warang Citi
`yezi`	Yezidi	`yi`	Yi (bytes `79 69 20 20`)
`zanb`	Zanabazar Square

.OTF File Format – Feature and script tags