Language Support

Every language is an isolated definition imported from @tanstack/highlight/languages/<name>. Aliases normalize only when their target definition is registered.

Language matrix

Language	Export	Aliases	Context-aware behavior
Ada	ada	None	Doubled quotes, character literals, case-insensitive keywords
Apache	apache	-	Directives, tags, comments
AppleScript	applescript	osascript	Nested block comments, strings, case-insensitive keywords
AWK	awk	None	Fields, strings, regex patterns following ~, keywords and comments
Batch	batch	bat, dos	REM comments, variables, quoted strings, control keywords
C	c	h	Preprocessor directives, strings, character literals, types
C#	csharp	c#, cs	Verbatim, interpolated, and raw strings with nested quotes, character literals, preprocessor directives, line-leading and parameter attributes, contextual and LINQ keywords
C++	cpp	c++, cc, cxx, hpp, hxx	Raw and prefixed strings, character literals, preprocessor directives, digit separators
Clojure	clojure	clj, edn	Forms, strings, keyword symbols, line comments
CMake	cmake	-	Bracket strings/comments, nested variables, generator expressions
CoffeeScript	coffeescript	coffee, cson	Block comments, triple-quoted strings, keywords and literals
Common Lisp	lisp	commonlisp, cl	Forms, keyword symbols, nested comments, strings
Crystal	crystal	cr	Declarations, quoted strings, line comments
CSS	css	-	Strings and comments protect inner syntax
D	d	None	Nested /+ comments, backtick strings, declarations
Dart	dart	-	Interpolated strings with nested quotes and braces, triple-quoted and raw strings, nested block comments, annotations, named-argument labels, contextual keywords
Diff	diff	patch	Metadata, inserted, and deleted lines
Dockerfile	dockerfile	docker	Common directives, variables, commands
EJS	ejs	-	HTML plus optional JavaScript delegation
Elixir	elixir	ex, exs	Declarations, atoms, quoted and triple-quoted strings
Elm	elm	None	Nested block comments, strings, keywords
Env	env	dotenv	Properties, values, comments
Erlang	erlang	erl	Directives, variables, strings, keywords
F#	fsharp	fs, f#, fsx	Nested comments, verbatim and triple-quoted strings, character literals
Fortran	fortran	f90, f95, f03, f08	Free-form comments, doubled quotes, case-insensitive keywords
GLSL	glsl	None	Preprocessor directives, qualifiers, types, comments
Go	go	golang	Raw strings, runes, comments, declarations
GraphQL	graphql	gql	Block descriptions, variables, directives, schema and operation keywords
Groovy	groovy	None	Declarations, quoted and triple-quoted strings, comments
Haskell	haskell	hs	Nested comments, character literals, strings, keywords
HTML	html	htm, xml, angular-html	Optional JavaScript/TypeScript and CSS delegation
HTTP	http	-	Methods, headers, protocol, paths
INI	ini	cfg	Sections, properties, quoted values, comments
Java properties	properties	java-properties	Keys, simple values, escaped key separators and comments
Java	java	-	Text blocks, character literals, annotations, contextual keywords, digit separators and hex floats
JavaScript	js	javascript, mjs, cjs, js-vue	Templates, interpolation, regex literals
JSON	json	jsonc, json5	Properties, comments, strings, literals
JSX	jsx	-	JavaScript plus contextual JSX tags
Julia	julia	jl	Nested comments, triple-quoted strings, character literals, keywords
Kotlin	kotlin	kt, kts	String templates with nested quotes and braces, raw strings, nested block comments, annotations vs labels, backticked names, contextual soft keywords and accessors
LaTeX	latex	tex	Commands, comments, pipe- and plus-delimited verbatim spans
Less	less	None	Variables, selectors, properties, strings and comments
Lua	lua	-	Level-matched long-bracket strings and comments, string escapes, Luau backtick strings, shebang lines, paren-less calls, LuaJIT number suffixes
Makefile	makefile	make, mk, mak	Targets, assignments, variable references, comments
Markdown	markdown	md	Optional fenced-language delegation
MATLAB	matlab	None	Line and block comments, doubled quotes, common transpose boundaries
Mermaid	mermaid	-	Common diagram declarations and arrows
Nginx	nginx	-	Directives, variables, URLs, comments
Nix	nix	nixos	Indented and quoted strings, comments, keywords
Objective-C	objectivec	objc, obj-c, objective-c, mm	At-prefixed declarations and strings, preprocessor directives
OCaml	ocaml	ml	Nested comments, strings, character literals, type-variable boundaries
Pascal / Delphi	pascal	pas, delphi, dpr	Brace and block comments, doubled apostrophes, case-insensitive keywords
Perl	perl	pl	Sigil and special variables, quote-like operators with nested delimiters, regex vs division, heredocs, POD blocks
PHP	php	-	PHP tags, attributes, quoted strings, heredoc/nowdoc, optional HTML delegation
Plaintext	plaintext	text, txt, -->	Escaping only
PowerShell	powershell	pwsh, ps1, ps	Here-strings, variables, commands, block comments
Protocol Buffers	protobuf	proto	Message and service declarations, scalar types, comments
Python	python	py	Triple strings, prefixes, decorators, comments
R	r	None	Strings, backtick names, literals, comments
Ruby	ruby	rb	Interpolated strings with nested quotes, percent literals, heredocs, regex vs division, symbols and hash keys, block comments
Rust	rust	rs	Nested block comments, raw strings with hash counts, byte and C strings, lifetimes vs character literals, attributes, macros
Scala	scala	sc	Nested comments, quoted and triple-quoted strings, declarations
Scheme	scheme	scm, racket	Comments, strings, forms, literals
SCSS	scss	None	Variables, directives, selectors, properties, strings and comments
Shell	shell	bash, sh, zsh, cmd, console	Heredocs, parameter expansion, comment boundaries
SQL	sql	-	Strings, comments, common SQL clauses
Svelte	svelte	-	Markup plus optional script/style and expression delegation
Swift	swift	-	Interpolated, multi-line, and raw strings with nested quotes, regex literals vs division, nested block comments, contextual keywords, attributes and directives
TOML	toml	-	Strings, comments, tables, properties
TSRX	tsrx	octane	TypeScript, contextual JSX, Octane component shorthand and template directives
TSX	tsx	-	TypeScript, contextual JSX, generic disambiguation
TypeScript	ts	typescript, angular-ts	JavaScript scanner plus TypeScript keywords/types
Verilog / SystemVerilog	verilog	v, sv, svh	Directives, sized numeric literals, keywords and comments
VHDL	vhdl	None	Strings, character literals, case-insensitive keywords, comments
Visual Basic .NET	vbnet	vb	Doubled quotes, apostrophe and REM comments, case-insensitive keywords
Vue	vue	-	Template markup plus optional script/style delegation
WebAssembly text	wat	wasm	Nested comments, named variables, instructions and strings
x86 assembly	x86asm	asm, nasm	Common instructions and registers, labels, directives and comments
YAML	yaml	yml	Comment boundaries and block scalars

Registration matters

import { createHighlighter } from '@tanstack/highlight/core'
import { html } from '@tanstack/highlight/languages/html'

const markupOnly = createHighlighter({ languages: [html] })

markupOnly highlights tags and attributes, but leaves <script> and <style> bodies uncolored. Add the delegated languages explicitly:

import { css } from '@tanstack/highlight/languages/css'
import { js } from '@tanstack/highlight/languages/js'

const withEmbeddings = createHighlighter({
  languages: [html, css, js],
})

Quality boundary

The 40 added definitions from Ada through WebAssembly text provide lexical highlighting for common documentation examples. They share a small scanner and carry only their own vocabulary. They do not include the Highlight.js runtime or its complete grammars.

Expressions inside strings are not tokenized separately in these definitions. Language-specific sigils, heredocs outside PowerShell, regex literals outside AWK's ~ form, macro bodies, and embedded languages are not fully parsed. Fortran covers free-form code; the assembly definition covers common x86 instructions. Existing definitions with deeper interpolation or embedding support retain that behavior.

The existing cmd alias continues to resolve to shell. Use batch, bat, or dos to select the new Windows Batch definition.

Support means useful highlighting for valid code commonly found in documentation. It does not mean compiler conformance or parity with an IDE grammar. The regression suite concentrates on contexts where simple priority regexes are predictably wrong.

Unknown or unregistered language names normalize to the configured fallback, which defaults to plaintext.

See Embedded Languages and Custom Languages for the registry model.