Hidden character index

Every codepoint with a page of its own: what it is for, where it comes from, when it is doing a real job, and whether the default pass removes it.

49 characters documented

Grouped by the category the engine reports. Within a category, the option that removes a character is the same for every row.

Zero-width and invisible

Governed by Default pass (always on)

Codepoint	Name	Default pass
U+200B	ZERO WIDTH SPACEA space with no width. It takes up a position in the string, marks a place where a line may break, and draws nothing at all.	Yes
U+2060	WORD JOINERThe opposite of a zero width space: an invisible character that forbids a line break at its position, without adding any width.	Yes
U+00AD	SOFT HYPHENA conditional hyphen. It is invisible until the line breaks at that exact spot, at which point a hyphen appears.	Yes
U+FEFF	ZERO WIDTH NO-BREAK SPACEThe byte order mark, U+FEFF, which editors and error messages usually spell ufeff. At the very start of a file it announces the encoding; anywhere else it is an invisible, non-breaking nothing that Unicode has deprecated for that use since 3.2.	Yes
U+2061	FUNCTION APPLICATIONAn invisible mathematical operator meaning "apply this function to this argument": the silent operator between f and (x).	Yes
U+2062	INVISIBLE TIMESThe invisible multiplication sign: the operator that is meant when you write "2x" and mean "2 times x".	Yes
U+2063	INVISIBLE SEPARATORThe invisible comma, U+2063 INVISIBLE SEPARATOR: the separator between the indices in a subscript like a⁣i⁣j.	Yes
U+2064	INVISIBLE PLUSThe invisible addition sign used between the parts of a mixed fraction: the "and" in "one and a half".	Yes
U+034F	COMBINING GRAPHEME JOINERA zero-width combining mark that exists to change how two characters sort and collate, not how they look. It never renders.	Yes
U+180E	MONGOLIAN VOWEL SEPARATORA Mongolian orthographic control that was reclassified in Unicode 6.3: it used to be a space, and is now a zero-width format character.	Yes
U+2800	BRAILLE PATTERN BLANKThe braille cell with no dots raised. It is a printing character with a real advance width, but nothing is drawn.	Yes
U+3164	HANGUL FILLERA Hangul compatibility character with no shape, used historically to fill an empty slot in a Korean syllable block.	Yes
U+FDD0-U+FDEF	NONCHARACTERSCodepoints Unicode has permanently reserved to never be characters: U+FDD0-U+FDEF plus the last two codepoints of every plane (U+FFFE, U+FFFF, U+1FFFE, and so on).	Yes

Bidirectional overrides

Governed by Default pass (always on)

Codepoint	Name	Default pass
U+202E	RIGHT-TO-LEFT OVERRIDEAn invisible control that forces every following character to be laid out right to left, regardless of what script it belongs to.	Yes
U+202D	LEFT-TO-RIGHT OVERRIDEThe mirror of U+202E: it forces everything after it to be laid out left to right, overriding the natural direction of Hebrew or Arabic.	Yes
U+202A	LEFT-TO-RIGHT EMBEDDINGOpens a left-to-right run inside surrounding text, until a matching POP DIRECTIONAL FORMATTING closes it.	Yes
U+202B	RIGHT-TO-LEFT EMBEDDINGOpens a right-to-left run inside surrounding text, until a matching POP DIRECTIONAL FORMATTING closes it.	Yes
U+202C	POP DIRECTIONAL FORMATTINGCloses the most recent embedding or override. On its own it does nothing; paired, it ends a directional run.	Yes

Bidirectional marks

Governed by Bidi marks --bidi

Codepoint	Name	Default pass
U+200E	LEFT-TO-RIGHT MARKAn invisible, zero-width character with strong left-to-right direction. It resolves ambiguity in mixed text without reordering anything itself.	When free-floating
U+200F	RIGHT-TO-LEFT MARKThe mirror of U+200E: an invisible character with strong right-to-left direction, used to resolve ambiguity in mixed-direction text.	When free-floating
U+061C	ARABIC LETTER MARKAn invisible mark with the direction of an Arabic letter, added in Unicode 6.3 for cases where a right-to-left mark is not quite strong enough.	When free-floating
U+2066-U+2069	BIDIRECTIONAL ISOLATESFour invisible controls (LRI, RLI, FSI and PDI) that wrap a span of text so its direction cannot leak into the text around it. They are the modern replacement for the embedding and override controls.	When free-floating

Joiners, selectors and fillers

Governed by Aggressive -a

Codepoint	Name	Default pass
U+200D	ZERO WIDTH JOINERAn invisible character that asks two neighbours to be drawn as one. It is what turns three separate people into a family emoji and what joins Arabic and Indic letters that would otherwise stand apart.	When free-floating
U+200C	ZERO WIDTH NON-JOINERThe opposite of the joiner: an invisible character that forbids two neighbours from being drawn as one connected form.	When free-floating
U+FE00-U+FE0F	VARIATION SELECTORSSixteen invisible modifiers that pick which of several forms the preceding character should take. VS15 (U+FE0E) asks for the black-and-white text form; VS16 (U+FE0F) asks for the colour emoji form.	When free-floating
U+E0000-U+E007F	TAG CHARACTERSA shadow copy of ASCII. U+E0020-U+E007E mirror the printable ASCII characters, one for one, and render as nothing. U+E007F CANCEL TAG ends a sequence.	When free-floating
U+180B-U+180D	MONGOLIAN FREE VARIATION SELECTORSThree invisible selectors (FVS1 to FVS3) that pick which positional form a Mongolian letter takes when the default shaping is not the one the word needs.	When free-floating
U+17B4-U+17B5	KHMER INHERENT VOWELSTwo invisible Khmer vowel signs, AQ and AA, which spell out a vowel that is normally implied rather than written.	When free-floating
U+115F-U+1160	HANGUL JAMO FILLERSTwo zero-width fillers, U+115F HANGUL CHOSEONG FILLER and U+1160 HANGUL JUNGSEONG FILLER, that stand in for a missing initial consonant or medial vowel when a Korean syllable block is written out in jamo. In everyday terms they are the choseong filler and the jungseong filler.	When free-floating

Private-use characters

Governed by Private-use characters --pua

Codepoint	Name	Default pass
U+E000-U+F8FF	PRIVATE USE AREACodepoints Unicode will never assign a meaning to, reserved for private agreements between a font and the software that uses it. U+E000-U+F8FF plus two large supplementary planes.	No

Exotic spaces

Governed by Spaces -s

Codepoint	Name	Default pass
U+00A0	NO-BREAK SPACEA space that forbids a line break. It looks exactly like a normal space and is a completely different codepoint.	No
U+202F	NARROW NO-BREAK SPACEA thinner no-break space, about half the width of a normal one, used where a full space would look too wide.	No
U+205F	MEDIUM MATHEMATICAL SPACEA space four eighteenths of an em wide, defined for mathematical typesetting: the gap around a binary operator.	No
U+3000	IDEOGRAPHIC SPACEA full-width space, exactly as wide as a CJK character. It is the space key on a Japanese, Chinese or Korean input method in full-width mode.	No
U+1680	OGHAM SPACE MARKThe word separator of the Ogham script. Unusually for a space, it is drawn as a short line in most Ogham fonts, and as blank in everything else.	No
U+2000-U+200A	GENERAL PUNCTUATION SPACESEleven typographic spaces of fixed width: en quad, em quad, en space, em space, three-per-em, four-per-em, six-per-em, figure space, punctuation space, thin space and hair space.	No

Line separators

Governed by Line endings -n

Codepoint	Name	Default pass
U+2028	LINE SEPARATORUnicode's own "new line here" character, distinct from the ASCII line feed. It ends a line without ending a paragraph.	No
U+2029	PARAGRAPH SEPARATORUnicode's paragraph break: the counterpart of U+2028, one level up.	No
U+0085	NEXT LINENEL, a C1 control that means "start a new line". It comes from EBCDIC, where it was the newline character.	No

Control characters

Governed by Control characters --controls

Codepoint	Name	Default pass
U+0000-U+001F	C0 CONTROL CHARACTERSThe 32 ASCII control codes: NUL, BEL, ESC, the file and record separators, and the rest. They are commands to a teletype, not text.	No
U+0080-U+009F	C1 CONTROL CHARACTERSA second block of 32 control codes inherited from ISO 8859. In Unicode they are controls; in Windows-1252 the same byte values are curly quotes, the em dash and the euro sign.	No

Typographic punctuation

Governed by Punctuation --punct

Codepoint	Name	Default pass
U+2013	EN DASHThe medium-length dash used for ranges and some relationships in edited typography.	No
U+2014	EM DASHThe long dash, one em wide. It sets off a parenthetical clause, or marks an interruption where a comma would be too quiet.	No
U+2019	RIGHT SINGLE QUOTATION MARKThe curly closing single quote, and in English the correct apostrophe: don’t, it’s, ’90s.	No
U+2026	HORIZONTAL ELLIPSISThree dots in a single character, with tighter spacing than three periods and no chance of a line break inside it.	No

Lookalike letters

Governed by Lookalike letters --confusables

Codepoint	Name	Default pass
U+0430	CYRILLIC SMALL LETTER AThe Cyrillic letter a. In almost every font it is drawn identically to the Latin letter a, and it is a completely different codepoint.	No
U+043E	CYRILLIC SMALL LETTER OThe Cyrillic letter o, drawn identically to Latin o in every font in common use.	No
U+0435	CYRILLIC SMALL LETTER IEThe Cyrillic letter ie, drawn exactly like a Latin e.	No
U+0441	CYRILLIC SMALL LETTER ESThe Cyrillic letter es, drawn exactly like a Latin c.	No

See the full reference, with the options that govern each category