Switch branches/tags
Nothing to show
Find file
Fetching contributors…
Cannot retrieve contributors at this time
793 lines (637 sloc) 29.8 KB

AsciiDoc User Guide

AsciiDoc is a text document format for writing documentation, articles, manuals, books and UNIX man pages. AsciiDoc files can be translated to HTML and DocBook markups using the asciidoc(1) command. AsciiDoc is highly configurable: both the AsciiDoc source file syntax and the backend output markups (which can be almost any type of SGML/XML markup) can be customized and extended by the user.


This is an overly large document, it probably needs to be refactored into a Tutorial, Quick Reference and Formal Reference.

If you’re new to AsciiDoc read this section and the Getting Started section and take a look at the example AsciiDoc (*.txt) source files in the distribution doc directory.

Plain text is the most universal electronic document format, regardless of your computing environment you can always read and write plain text documentation. But for many applications plain text is not the preferred presentation format — HTML and PDF formats are widely used as is the roff man page format. DocBook is a popular documentation markup format which can be translated to HTML, PDF and other presentation formats.

AsciiDoc is a plain text human readable/writable document format that can be translated to DocBook or HTML using the asciidoc(1) command. You can then either use asciidoc(1) generated HTML directly or run asciidoc(1) DocBook output through your favorite DocBook toolchain or use the AsciiDoc a2x(1) toolchain wrapper to produce PDF, EPUB, DVI, LaTeX, PostScript, man page, HTML and text formats.

The AsciiDoc format is a useful presentation format in its own right: AsciiDoc markup is simple, intuitive and as such is easily proofed and edited.

AsciiDoc is light weight: it consists of a single Python script and a bunch of configuration files. Apart from asciidoc(1) and a Python interpreter, no other programs are required to convert AsciiDoc text files to DocBook or HTML. See Example AsciiDoc Documents below.

Text markup conventions tend to be a matter of (often strong) personal preference: if the default syntax is not to your liking you can define your own by editing the text based asciidoc(1) configuration files. You can also create configuration files to translate AsciiDoc documents to almost any SGML/XML markup.

asciidoc(1) comes with a set of configuration files to translate AsciiDoc articles, books and man pages to HTML or DocBook backend formats.

My AsciiDoc Itch

DocBook has emerged as the de facto standard Open Source documentation format. But DocBook is a complex language, the markup is difficult to read and even more difficult to write directly — I found I was spending more time typing markup tags, consulting reference manuals and fixing syntax errors, than I was writing the documentation.

Getting Started

Installing AsciiDoc

See the README and INSTALL files for install prerequisites and procedures. Packagers take a look at Packager Notes.

Example AsciiDoc Documents

The best way to quickly get a feel for AsciiDoc is to view the AsciiDoc web site and/or distributed examples:

  • Take a look at the linked examples on the AsciiDoc web site home page Press the Page Source sidebar menu item to view corresponding AsciiDoc source.

  • Read the *.txt source files in the distribution ./doc directory along with the corresponding HTML and DocBook XML files.

AsciiDoc Document Types

There are three types of AsciiDoc documents: article, book and manpage. All document types share the same AsciiDoc format with some minor variations. If you are familiar with DocBook you will have noticed that AsciiDoc document types correspond to the same-named DocBook document types).

Use the asciidoc(1) -d (--doctype) option to specify the AsciiDoc document type — the default document type is article.

By convention the .txt file extension is used for AsciiDoc document source files.


Used for short documents, articles and general documentation. See the AsciiDoc distribution ./doc/article.txt example.

AsciiDoc defines standard DocBook article frontmatter and backmatter section markup templates (appendix, abstract, bibliography, glossary, index).


Books share the same format as articles; in addition there is the option to add level 0 book part sections.

Book documents will normally be used to produce DocBook output since DocBook processors can automatically generate footnotes, table of contents, list of tables, list of figures, list of examples and indexes.

AsciiDoc defines standard DocBook book frontmatter and backmatter section markup templates (appendix, dedication, preface, bibliography, glossary, index, colophon).

Example book documents

The ./doc/book.txt file in the AsciiDoc distribution.

Multi-part book

The ./doc/book-multi.txt file in the AsciiDoc distribution.


Used to generate roff format UNIX manual pages. AsciiDoc manpage documents observe special header title and section naming conventions — see the Manpage Documents section for details.

AsciiDoc defines the synopsis section markup template to generate the DocBook refsynopsisdiv section.

See also the asciidoc(1) man page source (./doc/asciidoc.1.txt) from the AsciiDoc distribution.

AsciiDoc Backends

The asciidoc(1) command translates an AsciiDoc formatted file to the backend format specified by the -b (--backend) command-line option. asciidoc(1) itself has little intrinsic knowledge of backend formats, all translation rules are contained in customizable cascading configuration files. Backend specific attributes are listed in the Backend Attributes section.

AsciiDoc ships with the following predefined backend output formats:


AsciiDoc generates the following DocBook document types: article, book and refentry (corresponding to the AsciiDoc article, book and manpage document types).

DocBook documents are not designed to be viewed directly. Most Linux distributions come with conversion tools (collectively called a toolchain) for converting DocBook files to presentation formats such as Postscript, HTML, PDF, EPUB, DVI, PostScript, LaTeX, roff (the native man page format), HTMLHelp, JavaHelp and text.

The AsciiDoc Preamble element generates a DocBook book preface.


The default asciidoc(1) backend is xhtml11 — XHTML 1.1 markup styled with CSS2. Output files have a .html extension. xhtml11 document generation is influenced by the following optional attributes (the default behavior is to generate XHTML with no section numbers, embedded CSS and no linked admonition icon images):


AsciiDoc XHTML output is styled using CSS2 stylesheets from the distribution ./stylesheets/ directory.


All browsers have CSS quirks, but Microsoft’s IE6 has so many omissions and errors that the xhtml11-quirks.css stylesheet and xhtml11-quirks.conf configuration files are included during XHTML backend processing to to implement workarounds for IE6. If you don’t use IE6 then the quirks stylesheet and configuration files can be omitted using the --attribute quirks! command-line option.

Default xhtml11 stylesheets:


The main stylesheet.


Tweaks for manpage document type generation.


Stylesheet modifications to work around IE6 browser incompatibilities.

Use the theme attribute to select an alternative set of stylesheets. For example, the command-line option -a theme=foo will use stylesheets foo.css, foo-manpage.css and foo-quirks.css instead of the default stylesheets.

Use the stylesheet attribute to include an additional stylesheet in XHTML documents. For example, the command-line option -a stylesheet=newsletter.css will use stylesheets newsletter.css.


This backend generates plain (unstyled) HTML 4.01 Transitional markup.

Document Structure

An AsciiDoc document consists of a series of block elements starting with an optional document Header, followed by an optional Preamble, followed by zero or more document Sections.

Almost any combination of zero or more elements constitutes a valid AsciiDoc document: documents can range from a single sentence to a multi-part book.

Block Elements

Block elements consist of one or more lines of text and may contain other block elements.

The AsciiDoc block structure can be informally summarized as follows [1]:

Document      ::= (Header?,Preamble?,Section*)
Header        ::= (Title,(AuthorInfo,RevisionInfo?)?)
AuthorInfo    ::= (FirstName,(MiddleName?,LastName)?,EmailAddress?)
RevisionInfo  ::= (RevisionNumber?,RevisionDate,RevisionRemark?)
Preamble      ::= (SectionBody)
Section       ::= (Title,SectionBody?,(Section)*)
SectionBody   ::= ((BlockTitle?,Block)|BlockMacro)+
Block         ::= (Paragraph|DelimitedBlock|List|Table)
List          ::= (BulletedList|NumberedList|LabeledList|CalloutList)
BulletedList  ::= (ListItem)+
NumberedList  ::= (ListItem)+
CalloutList   ::= (ListItem)+
LabeledList   ::= (ListEntry)+
ListEntry     ::= (ListLabel,ListItem)
ListLabel     ::= (ListTerm+)
ListItem      ::= (ItemText,(List|ListParagraph|ListContinuation)*)


  • ? implies zero or one occurrence, + implies one or more occurrences, * implies zero or more occurrences.

  • All block elements are separated by line boundaries.

  • BlockId, AttributeEntry and AttributeList block elements (not shown) can occur almost anywhere.

  • There are a number of document type and backend specific restrictions imposed on the block syntax.

  • The following elements cannot contain blank lines: Header, Title, Paragraph, ItemText.

  • A ListParagraph is a Paragraph with its listelement option set.

  • A ListContinuation is a list continuation element.


The Header contains a document title plus optional authorship and revision information:

  • The Header is optional, but if it is used it must start with a document title.

  • The header can be preceded by comments and attribute entries.

  • Optional Author and Revision information immediately follows the header title.

  • The header can include attribute entries.

  • The document header must be separated from the remainder of the document by one or more blank lines.

Here’s an example AsciiDoc document header:

Writing Documentation using AsciiDoc
Joe Bloggs <>
v2.0, February 2003:
Rewritten for version 2 release.

The author information line contains the author’s name optionally followed by the author’s email address. The author’s name consists of a first name followed by optional middle and last names separated by white space. Multi-word first, middle and last names can be entered in the header author line using the underscore as a word separator. The email address comes last and must be enclosed in angle <> brackets. Author names cannot contain angle <> bracket characters. Here a some examples of author information lines:

Joe Bloggs <>
Joe Bloggs
Vincent Willem van_Gogh

The optional revision information line follows the author information line. The revision information can be one of two formats:

  1. An optional document revision number followed by an optional revision date followed by an optional revision remark:

    • If the revision number is specified it must be followed by a comma.

    • The revision number must contain at least one numeric character.

    • Any non-numeric characters preceding the first numeric character will be dropped.

    • If a revision remark is specified it must be preceded by a colon. The revision remark extends from the colon up to the next blank line or attribute entry and is subject to normal text substitutions.

    • If a revision number or remark has been set but the revision date has not been set then the revision date is set to the value of the docdate attribute.

  2. An RCS/CSV/SVN $Id$ marker (if an $Id$ revision marker is used the header author line can be omitted).

Here a some examples of header revision lines:

v2.0, February 2003
February 2003
v2.0, February 2003: Rewritten for version 2 release.
February 2003: Rewritten for version 2 release.
v2.0,: Rewritten for version 2 release.
:Rewritten for version 2 release.

You can override or set header parameters by passing revnumber, revremark, revdate, email, author, authorinitials, firstname and lastname attributes using the asciidoc(1) -a (--attribute) command-line option. For example:

$ asciidoc -a revdate=2004/07/27 article.txt

The revnumber attribute can be an RCS/CSV/SVN $Id$ marker. Attributes can also be added to the header for substitution in the header template with Attribute Entry elements.

Additional document header information

DocBook defines numerous elements for document meta-data, for example: copyrights, document history and authorship information. The AsciiDoc header syntax provides for basic revision and author information — additional information such as copyrights, document history authorship details can be optionally included from a separate document information file. The document information file can contain any DocBook elements that are allowed inside the DocBook articleinfo and bookinfo elements:

  • The document information file must be in the same directory as the source document and must be named like <docname>-docinfo.xml. For example, if the source document is called mydoc.txt then the document information file would be named mydoc-docinfo.xml.

  • The document information file will only be included in the DocBook output if the docinfo attribute is defined, for example:

    $ asciidoc -a docinfo -b docbook mydoc.txt
  • See the ./doc/article-docinfo.xml example that comes with the AsciiDoc distribution.

  • Similarly AsciiDoc will include a docinfo file named like <docname>-docinfo.html (if it exists) into HTML backend outputs. In the case of HTML outputs however it’s usually easier to put AsciiDoc markup in your source file directly after the header.

DocBook has a rich set of meta data elements, just which of these elements are rendered is DocBook processor dependent.


The Preamble is an optional untitled section body between the document Header and the first Section title.


In addition to the document title (level 0), AsciiDoc supports four section levels: 1 (top) to 4 (bottom). Section levels are delimited by section titles. Sections are translated using configuration file section markup templates. AsciiDoc generates the following intrinsic attributes specifically for use in section markup templates:


The level attribute is the section level number, it is normally just the title level number (1..4). However, if the leveloffset attribute is defined it will be added to the level attribute. The leveloffset attribute is useful for combining documents.


The -n (--section-numbers) command-line option generates the sectnum (section number) attribute. The sectnum attribute is used for section numbers in HTML outputs (DocBook section numbering are handled automatically by the DocBook toolchain commands).

Section markup templates

Section markup templates specify output markup and are defined in AsciiDoc configuration files. Section markup template names are derived as follows (in order of precedence):

  1. From the title’s first positional attribute or template attribute. For example, the following three section titles are functionally equivalent:

    List of Terms
    List of Terms
    List of Terms
  2. When the title text matches a configuration file [specialsections] entry.

  3. If neither of the above the default sect<level> template is used (where <level> is a number from 1 to 4).

In addition to the normal section template names (sect1, sect2, sect3, sect4) AsciiDoc has the following templates for frontmatter, backmatter and other special sections: abstract, preface, colophon, dedication, glossary, bibliography, synopsis, appendix, index. These special section templates generate the corresponding Docbook elements; for HTML outputs they default to the sect1 section template.

Section IDs

If no explicit section ID is specified an ID will be synthesised from the section title. The primary purpose of this feature is to ensure persistence of table of contents links (permalinks): the missing section IDs are generated dynamically by the JavaScript TOC generator after the page is loaded. If you link to a dynamically generated TOC address the page will load but the browser will ignore the (as yet ungenerated) section ID.

The IDs are generated by the following algorithm:

  • Replace all non-alphanumeric title characters with underscores.

  • Strip leading or trailing underscores.

  • Convert to lowercase.

  • Prepend the idprefix attribute (so there’s no possibility of name clashes with existing document IDs). Prepend an underscore if the idprefix attribute is not defined.

  • A numbered suffix (_2, _3 …​) is added if a same named auto-generated section ID exists.

For example the title Jim’s House would generate the ID _jim_s_house.

Section ID synthesis can be disabled by undefining the sectids attribute.

Special Section Titles

AsciiDoc has a mechanism for mapping predefined section titles auto-magically to specific markup templates. For example a title Appendix A: Code Reference will automatically use the appendix section markup template. The mappings from title to template name are specified in [specialsections] sections in the Asciidoc language configuration files (lang-*.conf). Section entries are formatted like:


<title> is a Python regular expression and <template> is the name of a configuration file markup template section. If the <title> matches an AsciiDoc document section title then the backend output is marked up using the <template> markup template (instead of the default sect<level> section template). The {title} attribute value is set to the value of the matched regular expression group named title, if there is no title group {title} defaults to the whole of the AsciiDoc section title. If <template> is blank then any existing entry with the same <title> will be deleted.

Special section titles vs. explicit template names

AsciiDoc has two mechanisms for specifying non-default section markup templates: you can specify the template name explicitly (using the template attribute) or indirectly (using special section titles). Specifying a section template attribute explicitly is preferred. Auto-magical special section titles have the following drawbacks:

  • They are non-obvious, you have to know the exact matching title for each special section on a language by language basis.

  • Section titles are predefined and can only be customised with a configuration change.

  • The implementation is complicated by multiple languages: every special section title has to be defined for each language (in each of the lang-*.conf files).

Specifying special section template names explicitly does add more noise to the source document (the template attribute declaration), but the intention is obvious and the syntax is consistent with other AsciiDoc elements c.f. bibliographic, Q&A and glossary lists.

Special section titles have been deprecated but are retained for backward compatibility.

Inline Elements

Inline document elements are used to format text and to perform various types of text substitution. Inline elements and inline element syntax is defined in the asciidoc(1) configuration files.

Here is a list of AsciiDoc inline elements in the (default) order in which they are processed:

Special characters

These character sequences escape special characters used by the backend markup (typically <, >, and & characters). See [specialcharacters] configuration file sections.


Elements that markup words and phrases; usually for character formatting. See [quotes] configuration file sections.

Special Words

Word or word phrase patterns singled out for markup without the need for further annotation. See [specialwords] configuration file sections.


Each replacement defines a word or word phrase pattern to search for along with corresponding replacement text. See [replacements] configuration file sections.

Attribute references

Document attribute names enclosed in braces are replaced by the corresponding attribute value.

Inline Macros

Inline macros are replaced by the contents of parametrized configuration file sections.

Document Processing

The AsciiDoc source document is read and processed as follows:

  1. The document Header is parsed, header parameter values are substituted into the configuration file [header] template section which is then written to the output file.

  2. Each document Section is processed and its constituent elements translated to the output file.

  3. The configuration file [footer] template section is substituted and written to the output file.

When a block element is encountered asciidoc(1) determines the type of block by checking in the following order (first to last): (section) Titles, BlockMacros, Lists, DelimitedBlocks, Tables, AttributeEntrys, AttributeLists, BlockTitles, Paragraphs.

The default paragraph definition [paradef-default] is last element to be checked.

Knowing the parsing order will help you devise unambiguous macro, list and block syntax rules.

Inline substitutions within block elements are performed in the following default order:

  1. Special characters


  3. Special words

  4. Replacements

  5. Attributes

  6. Inline Macros

  7. Replacements2

The substitutions and substitution order performed on Title, Paragraph and DelimitedBlock elements is determined by configuration file parameters.

Text Formatting

Quoted Text

Words and phrases can be formatted by enclosing inline text with quote characters:

Emphasized text

Word phrases 'enclosed in single quote characters' (acute accents) or _underline characters_ are emphasized.

Strong text

Word phrases *enclosed in asterisk characters* are rendered in a strong font (usually bold).

Monospaced text

Word phrases +enclosed in plus characters+ are rendered in a monospaced font. Word phrases `enclosed in backtick characters` (grave accents) are also rendered in a monospaced font but in this case the enclosed text is rendered literally and is not subject to further expansion (see inline literal).

‘Single quoted text’

Phrases enclosed with a `single grave accent to the left and a single acute accent to the right' are rendered in single quotation marks.

“Double quoted text”

Phrases enclosed with ``two grave accents to the left and two acute accents to the right'' are rendered in quotation marks.

Unquoted text

Placing #hashes around text# does nothing, it is a mechanism to allow inline attributes to be applied to otherwise unformatted text (see example below).

Quoted text behavior
  • Quoting cannot be overlapped.

  • Different quoting types can be nested.

  • To suppress quoted text formatting place a backslash character immediately in front of the leading quote character(s). In the case of ambiguity between escaped and non-escaped text you will need to escape both leading and trailing quotes, in the case of multi-character quotes you may even need to escape individual characters.

Quoted text can be prefixed with an attribute list:

  1. Setting the AsciiDoc role attribute will enclose DocBook markup in a phrase element and HTML markup in a span element. For example this AsciiDoc: [role="foo"]'Hello World' generates this DocBook: <phrase role="foo"><emphasis>Hello World</emphasis></phrase> and this HTML: <span class="foo"><em>Hello World</em></span>

  2. You can set the font color, background color and size using the first three positional attribute arguments (HTML outputs only). The first argument is the text color; the second the background color; the third is the font size. Colors are valid CSS colors and the font size is a number which treated as em units. Here are some examples:

    [red]#Red text#.
    [,yellow]*bold text on a yellow background*.
    [blue,#b0e0e6]+Monospaced blue text on a light blue background+
    [,,2]#Double sized text#.

New quotes can be defined by editing asciidoc(1) configuration files. See the Configuration Files section for details.

Constrained and Unconstrained Quotes

There are actually two types of quotes:

Constrained quotes

Quoted must be bounded by white space or commonly adjoining punctuation characters. These are the most commonly used type of quote.

Unconstrained quotes

Unconstrained quotes have no boundary constraints and can be placed anywhere within inline text. For consistency and to make them easier to remember unconstrained quotes are double-ups of the _, *, + and # constrained quotes:

__unconstrained emphasized text__
**unconstrained strong text**
++unconstrained monospaced text++
##unconstrained unquoted text##

The following example emboldens the letter F:

**F**ile Open...

Superscripts and Subscripts

Put \^carets on either^ side of the text to be superscripted, put \~tildes on either side~ of text to be subscripted. For example, the following line:

e^&#960;i^+1 = 0. H~2~O and x^10^. Some ^super text^
and ~some sub text~

Is rendered like:

eπi+1 = 0. H2O and x10. Some ^super text^ and ~some sub text~

Superscripts and subscripts are implemented as unconstrained quotes and they can be escaped with a leading backslash and prefixed with with an attribute list.

Line Breaks

A plus character preceded by at least one space character at the end of a non-blank line forces a line break. It generates a line break (br) tag for HTML outputs and a custom XML asciidoc-br processing instruction for DocBook outputs. The asciidoc-br processing instruction is handled by a2x(1).

Page Breaks

A line of three or more less-than (<<<) characters will generate a hard page break in DocBook and printed HTML outputs. It uses the CSS page-break-after property for HTML outputs and a custom XML asciidoc-pagebreak processing instruction for DocBook outputs. The asciidoc-pagebreak processing instruction is handled by a2x(1). Hard page breaks are sometimes handy but as a general rule you should let your page processor generate page breaks for you.


A line of three or more apostrophe characters will generate a ruler line. It generates a ruler (hr) tag for HTML outputs and a custom XML asciidoc-hr processing instruction for DocBook outputs. The asciidoc-hr processing instruction is handled by a2x(1).


By default tab characters input files will translated to 8 spaces. Tab expansion is set with the tabsize entry in the configuration file [miscellaneous] section and can be overridden in included files by setting a tabsize attribute in the include macro’s attribute list. For example:


The tab size can also be set using the attribute command-line option, for example --attribute tabsize=4


The following replacements are defined in the default AsciiDoc configuration:

(C) copyright, (TM) trademark, (R) registered trademark,
-- em dash, ... ellipsis, -> right arrow, <- left arrow, => right
double arrow, <= left double arrow.

Which are rendered as:

© copyright, ™ trademark, ® registered trademark, — em dash, …​ ellipsis, → right arrow, ← left arrow, ⇒ right double arrow, ⇐ left double arrow.

You can also include arbitrary entity references in the AsciiDoc source. Examples:

&#x278a; &#182;


➊ ¶

To render a replacement literally escape it with a leading back-slash.

The Configuration Files section explains how to configure your own replacements.

Special Words

Words defined in [specialwords] configuration file sections are automatically marked up without having to be explicitly notated.

The Configuration Files section explains how to add and replace special words.

1. This is a rough structural guide, not a rigorous syntax definition