PubSuite Classic
PubSuite Classic is a professional XML-to-PDF editor. You write your document in a structured XML format, style each element through the sidebar panels, and export a pixel-perfect PDF — with a live preview that matches the output exactly. It includes full mathematics typesetting and five specialised editor modes: Scientific Paper, Magazine Editor, Label Creator, Carousel Creator and Technical Docs.
Getting Started
The start screen has an Engine switch (Classic engine or NextGen engine, the Classic 2 editor) and six cards: Classic Editor plus the five modes. Pick a card to open the editor with that mode's templates and panel. See Editor Modes.
When the editor opens you see three main areas:
XML Editor
Source
Live Preview
PDF output
- Sidebar — tabs for every document setting: Document, Styles, H/F, TOC, Footnotes, Assets, Mapping, Inline, Presets, plus the Math tab and the Mode tab (the current mode's panel and the settings shared by every mode).
- XML Editor — Monaco-powered editor with syntax highlighting. Every change re-renders the preview in real time.
- Live Preview — a canvas rendering that matches the exported PDF exactly. Click any text to edit it in place — paragraphs, table cells and footnotes.
Note: PubSuite Classic uses XML, not Markdown or HTML. The tag names you invent become your style names —
<paragraph>,
<heading1>,
<pullquote> — and you style each one in the Typography panel. Markdown and LaTeX can be imported (see
Technical Docs and
LaTeX Import).
XML Editor
The editor uses Monaco (the engine behind VS Code) with full XML syntax highlighting, bracket matching, and auto-indentation.
Basic Document Structure
<?xml version="1.0"?>
<document>
<heading1>Chapter One</heading1>
<paragraph>Your text goes here, with a formula $e^{i\pi} + 1 = 0$.</paragraph>
<heading2>A Sub-heading</heading2>
<paragraph>More text.</paragraph>
</document>
Inline Markup
| Tag | Result | Example |
<b> | Bold | <b>important</b> |
<i> | Italic | <i>emphasis</i> |
<u> | Underline | <u>link text</u> |
<fn> | Footnote | word<fn>This appears at the bottom.</fn> |
$…$ / <math> | Inline formula | $x^2 + y^2$ |
<cite>, <xref>, <eqref> | Citation, cross-reference, equation reference | <cite key="knuth1984"/> |
<code>, <kbd> | Inline code, key cap | <kbd>Ctrl</kbd> |
<smallcaps>, <nobr> | Small capitals, no line break | <nobr>10 km</nobr> |
Keyboard Shortcuts
| Shortcut | Action |
Ctrl+Z | Undo |
Ctrl+Shift+Z | Redo |
Ctrl+F | Find in editor |
Ctrl+H | Find & Replace |
Ctrl+/ | Toggle comment |
Alt+Shift+F | Format / pretty-print XML |
Ctrl+M | Equation editor — inline formula |
Ctrl+Shift+M | Equation editor — display equation |
Ctrl+Shift+H | Arabic vowel-sign palette |
Double-click a formula | Edit it in the equation editor |
Live Preview & Editing
The preview renders your document exactly as it will appear in the exported PDF — same fonts, same layout engine, same measurements.
- Click to edit in place — click any text to edit it directly without switching to the XML editor. A click goes to the block under the pointer, also in multi-column layouts, next to frames and in a label's second column.
- Select a whole block — move the pointer into the free space where a block's lines start: the page margin for the first column, the gutter between columns for the others. The pointer turns into an arrow pointing at the text — to the right beside a left-to-right block, to the left in the right-hand margin of a right-to-left (Arabic, Hebrew) block, where a light stripe also marks the page's right edge; click to select the whole paragraph, heading or list item. Every column has its own selection bar.
- Table cells — edited one by one: click, type, select inside the cell. Enter adds nothing to a cell.
- Footnotes — edited where they are printed, at the foot of the page. The footnote mark in the text is generated, so typing, Enter and bold/italic land exactly where you click.
- Formulas, citations and references behave as one character: a click puts the caret before or after them, and they disappear only when selected and deleted as a whole. Editing never turns a formula or a footnote into plain text.
- Selection is drawn only over the characters of the element being edited — never over generated numbers, bullets or labels — and always reaches the last selected letter, also in bold, upper-case, letter-spaced or justified text. A paragraph that continues on the next page is highlighted on both pages.
- Section and list numbers are generated: the caret never stops inside "1.2" or a bullet.
- Generated text is lighter while editing — the header and footer, page numbers, section, list and equation numbers, labels such as Chapter 1 or Definition 1, bullets, the table of contents and other generated lists, citations, cross-references and footnote marks are drawn in grey, so the text you type stands out. They change elsewhere (H/F and Doc tabs, Mode tab, the citation list), not by typing in them. Exports are never dimmed.
- Print preview — the printer button in the toolbar, or
Ctrl+Alt+P, shows the pages exactly as they print: nothing grey, no caret, selection or editing handles. Press Esc or the button again to edit.
- Zoom controls — use the
− / + buttons, or Fit to fit the page width (or the spread in the Magazine Editor).
- Page count — shown in the toolbar. Long documents paginate automatically.
- Real-time update — the preview refreshes as you type, after a short debounce delay.
Document Panel
The Doc tab controls the page itself.
Page Size & Margins
Set width and height in mm, pt, or inches. Common presets include A4, A5, Letter, Legal, and book trim sizes. Margins can be set independently on all four sides.
Base Typography
- Base font — default font for any tag without its own style. Built-in: Helvetica, Times-Roman, Courier.
- Base size — default font size in pt.
- Smart quotes — converts
" and ' to typographic quotes.
- Ligatures — enables fi, fl, ff ligature substitutions for uploaded OpenType fonts.
- Top reference — controls whether line leading is measured from cap height, x-height, or full em.
Document Metadata
Title, author, subject, and language. These appear in PDF metadata and are available as {{title}} and {{author}} variables in headers and footers. The document language (Mode tab → Typography) also sets hyphenation, index sorting and the citation language.
Typography & Styles
The Styles tab defines how each XML tag looks. Every tag can have its own style. New tags from the editor modes come with sensible default styles, so you only style them when you want something different.
Style Properties
| Property | Description |
| Font / Size | Font family and size in pt |
| Bold / Italic | Weight and style |
| Color | Hex color for the text |
| Alignment | Left, center, right, or justify |
| Line spacing | Multiplier (1.0 = single, 2.0 = double) |
| Space before / after | Vertical space around the block, in pt. Large type measures its space-before from its cap height |
| First-line indent | Indent for the first line only, in pt |
| Left / Right indent | Block indent from the margin, in pt |
| Hyphenation | Auto-hyphenate long words at line breaks; language, minimum word length and characters before/after the hyphen (Advanced) |
| Advanced Justification | Min / Desired / Max word-space %, and extra letter spacing when justifying (see Justification) |
| Align to grid | The block joins the baseline grid |
| Keep with next | Prevents a page break between this block and the next |
| Uppercase | Forces all text to uppercase |
| Drop cap lines | Number of lines the first character spans |
| Columns | Number of text columns within this block |
| Running title | Text to show in the header/footer for sections using this style |
| Section level / format | Numbering of headings: {1}, {1}.{2}; formats 1, i, a, I, A |
| Border | Width and color of a rule around the block |
| Padding | Space inside the border |
| Background color | Fill color behind the block |
Style Mapping
The Map tab lets you create tag aliases — e.g. map p → paragraph.
Inline Styles
The Inline tab defines styles for inline elements within a parent tag.
Presets
Ready-made document configurations for common use cases. Includes citation standards (Chicago, APA, MLA, Harvard, IEEE) and document structures (Novel, Report, Newsletter, Screenplay, Dissertation).
Apply Modes
- Apply — applies everything: page size, margins, base typography, and all style definitions.
- Page — applies only page size and margins.
- Styles — applies only typography settings.
Preset Configurator
Click + New Preset to build a preset visually using intent cards, page size slider, margin and line-spacing controls, font style buttons, column count, and header/footer settings.
Each editor mode also has complete starter
templates (Mode tab → Template → Load). See
Editor Modes.
Table of Contents
The TOC tab generates an automatic table of contents from your headings. Choose which tags to include, configure per-level styles, and set the leader style (dots, lines, or none).
- By default the TOC is inserted before the first page. Put
<toc/> where you want it instead — for example after a title page — and it is printed there on a page of its own.
- Page numbers are exact: the document is laid out again until every number is stable.
<heading1 numbered="false"> gives an unnumbered heading that still appears in the TOC.
- Lists of figures, tables, listings and algorithms work the same way (see Notes, Glossaries & Index).
- In the PDF every TOC entry is a link, and every heading is a bookmark.
Assets
Custom Fonts
Upload TTF or OTF font files. Once uploaded, the font name is available in the style editor. Ligatures and kerning are applied automatically. Fonts are embedded in the exported PDF. An uploaded font with an OpenType MATH table (Cambria Math, Asana Math, Fira Math, Libertinus Math …) also appears in the Math tab's font list.
Images
<image src="my-photo.jpg" width="120" height="80" />
Width and height are in points. Images are embedded in the PDF. For numbered figures with captions use <figure> (see Figures, Tables & Floats).
Arabic & Right-to-Left
Classic writes Arabic and Hebrew right to left and mixes them with Latin text in the same paragraph. Letter joining, ligatures and mark placement come from the font's own OpenType rules, on screen and in the PDF.
- Amiri is built in — Regular, Bold, Italic and Bold Italic. It loads by itself when the document contains Arabic text or a style uses it, and it is listed in the style editor's font list. Arabic letters, Arabic punctuation (، ؛ ؟) and Arabic-Indic digits typed into a paragraph set in a Latin font are drawn in Amiri automatically; Hebrew uses an uploaded Hebrew font.
- Paragraph direction — set by
dir="rtl" or dir="ltr" on the element, by a language block such as <arabic>, or else by the paragraph's first letter. A right-to-left paragraph is set flush right (left alignment becomes right, the last line of justified text goes to the right) and its indent, border and padding are mirrored.
- Mixed text — words are ordered with the Unicode bidirectional algorithm: Latin words and numbers inside Arabic read left to right in place, brackets are mirrored, and punctuation stays with its sentence.
- Editing — the caret, click position and selection follow the text as drawn; in right-to-left text ← moves forward and → moves back. While you type, the caret of a right-to-left paragraph grows from its right edge.
- PDF — Arabic words are shaped with the embedded font's OpenType features (joining forms, ligatures, mark positioning) and stay selectable and searchable. Spacing around brackets and digits is the font's own.
The vowel-sign tools appear only when the document contains Arabic text:
- Show (Mode tab → Arabic vowel signs) — all vowel signs; none (hamza and madda stay, they are spelling); shadda only; all except case endings; the skeleton without dots; or the bare skeleton (rasm). Set it for the document, for a style (For a style), or for a passage with an attribute such as
tashkeel="none" on any element. Only the drawing changes; the text keeps every mark.
- Spelling conventions — As typed, or As the font's style writes them (naskh puts fathatan on the letter before alef). Attribute:
tashkeel-conventions="style".
- Vowel-sign palette — toolbar button حَ or
Ctrl+Shift+H. Click just after a letter (or select letters); drag the palette by its title bar to move it. The palette shows that letter in its word in blue and lights up the signs it already has. Click a sign to add it, click it again to remove it; a vowel replaces the existing one; Remove marks clears them. ◄ next letter / previous ► step through the word. Marks are always kept in the correct order.
- Below the signs the palette has the letters of the document's language, Normalise letters and Type in Latin letters — see Languages & Letters, which also covers searching with letter variants in Find & Replace.
Smart mark placement, moving marks by hand, the font's tashkeel options, Qurʾānic orthography and the Arabic templates are available in
Classic 2.
AI Assistant
Click the AI button in the toolbar to open the AI drawer. The assistant makes changes to your document settings using natural language.
What the AI can do
- Apply a citation style or document structure preset
- Set page size and margins
- Change base font, size, and line spacing
- Edit style properties for any tag
- Enable headers, footers, and page numbers
- Enable the table of contents
- Find and replace text in the document
Example prompts
Apply Chicago style 17th edition
Set page to A5 with 20mm margins
Make heading1 bold, 18pt, centered with 24pt space before
Enable page numbers in the footer centered
Add a table of contents for heading1 and heading2
PDF Export
Click ↓ PDF in the toolbar. All fonts and images are embedded. The layout engine that generates the PDF is the same one that renders the live preview — output matches exactly.
- Formulas, barcodes and hazard symbols are exported as sharp vector outlines, with an invisible text layer so the PDF can be searched and copied.
- Cross-references, citations, footnote marks and TOC entries are clickable links; headings become bookmarks.
- Optional Tagged PDF (accessibility) and PDF/A-2 (archiving) — see PDF Output Options.
- Label and carousel modes add their own exports (label sheets, PNG slides).
PDF export is available on paid plans. Trial users can use the live preview but cannot export.
Projects
- Save — saves the current XML and all settings as a
.json file: styles, Math tab, editor mode, mode settings, bibliography and shared settings (see Saved Project Data).
- Load — opens a saved project file and restores everything exactly. Older projects load unchanged and get the defaults for new settings.
- New — clears the editor and starts a blank document.
Projects are saved locally as files — they are not stored on the server.
XML Reference
Document Root
Every document must start with <?xml version="1.0"?> and have a single root element. By convention this is <document> but any name works.
Common Block Tags
| Suggested tag | Typical use |
<paragraph> | Body text |
<heading1> | Chapter title |
<heading2> | Section heading |
<blockquote> | Extended quotation |
<pullquote> | Highlighted quote |
<caption> | Image or figure caption |
<bullet> | Bulleted list item |
<numbered> | Numbered list item |
<image> | Embedded image |
<equation> | Numbered display equation |
<figure>, <table> | Numbered figure or table with caption |
<codeblock> | Highlighted code listing |
The tags of every mode work in every document — see the sections under Both editors in the navigation, and General-purpose Tags.
Well-formed XML Rules
- Every opening tag must have a closing tag
- Self-closing tags end with
/>
- Tags must be properly nested — they cannot overlap
- Attribute values must be quoted
- Use
& for &, < for <, > for > (also inside formulas: the alignment & of align is written &)
Lengths in attributes accept mm, cm, in, pt, px or %. A bare number means points; % is relative to the available width.
PubSuite Classic 2 Beta
Classic 2 (the NextGen engine on the start screen) is the advanced editor for multilingual publishing and precision typography. It shapes Arabic with the font's own OpenType rules, lays out Arabic, Hebrew and Latin text together with full bidirectional layout, places Arabic vowel signs intelligently, and corrects collisions between lines — all with its own SVG font engine, so the preview and the export are identical.
Beta: Classic 2 is in active development and available to members. It asks for the beta credentials once per browser session.
Getting Started
The workspace is the same as Classic — sidebar, XML editor, and live preview — and every editor mode and the mathematics support work the same in both (see the Both editors sections). Choosing a card for the other engine on the start screen opens that editor with the mode. While the editor starts, a progress bar shows the step it is on (loading fonts with a percentage, loading the editor, laying out the document).
- Sidebar — Document settings, Styles, H/F, TOC, Footnotes, Assets (with Fonts), Math and Mode tabs.
- XML Editor — Monaco editor, same as Classic.
- Live Preview — SVG-based rendering. Click any text to edit it in place.
- Status bar — shows the typing language (Typing: Latin | العربية) and announces fonts loaded on first use.
XML Editor
The XML editor works identically to Classic — Monaco with syntax highlighting, find & replace, and auto-indentation. The document structure is the same XML format.
Language Tags
Classic 2 adds language block and inline tags for multilingual documents. They are optional — every word is drawn in a font that has its script automatically — but can be written directly or set from the toolbar:
<!-- Block-level: wrap an entire paragraph in a language -->
<arabic>النص العربي هنا</arabic>
<hebrew>טקסט עברי כאן</hebrew>
<english>English paragraph here</english>
<!-- Inline: switch language within a paragraph -->
<paragraph>This contains <ar>كلمة عربية</ar> inline.</paragraph>
Keyboard Shortcuts
| Shortcut | Action |
Ctrl+Z | Undo |
Ctrl+Shift+Z | Redo |
Ctrl+F | Find in editor |
Ctrl+H | Find & Replace |
Alt+Shift+F | Format XML |
Click text | Edit in place in the preview |
Ctrl+M / Ctrl+Shift+M | Equation editor — inline / display |
Ctrl+Shift+F | Switch the typing language (Latin ↔ Arabic) |
Ctrl+Shift+H | Arabic vowel-sign palette |
Alt+arrows | Move the vowel signs above the letter before the caret |
Alt+Shift+arrows | Move the vowel signs below the letter |
Alt+0 | Return the letter's marks to automatic placement |
Home / End | Start / end of the line in reading order (with Ctrl: of the paragraph) |
Live Preview & Editing
The preview shows your document exactly as it will appear in the export. Arabic and Hebrew render right-to-left, and you can click any text to edit it in place.
- Click to edit in place — the click lands in the text under the pointer, in both directions, in columns, beside frames and in label columns. Where two directions meet, the caret goes on the side you clicked.
- Table cells and footnotes — edited where they are printed. Footnote marks are generated and take no room in the text, so the caret is always exact.
- Formulas, citations and references — one character for the caret; double-click a formula (also in a footnote) to edit it.
- Generated text is lighter while editing — the header and footer, page numbers, section, list and equation numbers, labels such as Chapter 1 or Definition 1, bullets, the table of contents and other generated lists, citations, cross-references and footnote marks are drawn in grey, so the text you type stands out. They change elsewhere (H/F and Doc tabs, Mode tab, the citation list), not by typing in them. Exports are never dimmed.
- Print preview — the printer button in the toolbar, or
Ctrl+Alt+P, shows the pages exactly as they print: nothing grey, no caret, selection or editing handles. Press Esc or the button again to edit.
- Select a whole block — click in the free space where the block's lines start (the arrow pointer): left of a left-to-right block, right of a right-to-left one; in columns, the gutter is each column's selection bar.
- Selection — a selection across both directions is drawn as the pieces it covers on screen, never over generated labels or numbers (except when the whole element is selected: then its label, such as Definition 1 and its title, is highlighted too), and is measured from the joined Arabic letters as drawn (kashida included), so it always covers the whole selected word. Blocks continuing on the next page are highlighted on both pages, and a formula broken across two lines is highlighted on both lines.
- Dimming — in the paragraph being edited, when it has two languages, the inactive language is dimmed (on screen only).
- Zoom controls —
− / + buttons and Fit.
- Ligatures — Latin and Arabic ligatures are applied automatically.
Classic 2 uses its own font system. See the
Fonts section for how to select and assign fonts.
Document Panel
Works the same as Classic: page size, margins, base typography, and document metadata. Classic 2 adds:
| Setting | Description |
| Default language | The primary script of the document. With Arabic or Hebrew the whole document reads from the right (see Bilingual & RTL) |
| Base font / Arabic font / Latin font | The default font of each loaded language. In a right-to-left document, Latin words take the Latin font |
| Digits | Western, Arabic-Indic ٠١٢٣ or Persian/Urdu ۰۱۲۳ — everywhere, or only in Arabic text. The text itself keeps 0–9 |
| Justify Arabic with kashida | Justified Arabic lines are filled by lengthening letter joins before word spaces are widened (on by default) |
Typography Features
Each typography role (display, heading, body, etc.) can have OpenType feature settings applied:
| Feature | Tag | Description |
| Standard ligatures | liga | fi, fl, ff, ffi, ffl — on by default |
| Contextual alternates | calt | Context-sensitive glyph variants — on by default |
| Stylistic sets | ss01–ss20 | Designer-defined alternate glyph sets (Amiri: ss01 low baa dot, ss02 alternative meem-alef, ss04 naskh gaf …) |
| Oldstyle figures | onum | Lowercase-style numerals that sit on the baseline |
| Lining figures | lnum | Uniform-height numerals aligned to cap height |
| Tabular figures | tnum | Fixed-width numerals for tables and columns |
Feature availability depends on the selected font. Not all fonts support all features.
Typography & Styles
Style properties work the same as Classic. Classic 2 adds a few properties specific to the SVG engine:
| Property | Description |
| Font / Size | Select from the fonts loaded in Assets → Fonts |
| Arabic font / Latin font … | One field per other loaded language: the font for that language's words in this style. Automatic uses the document's base font for that language |
| Bold / Italic | Uses the bold or italic sibling of the family, loaded automatically the first time it is needed. Arabic has no italic and stays upright |
| Color | Hex color for the text |
| Alignment | Left, center, right, justify. In a right-to-left paragraph "left" means the start side (flush right) |
| Line spacing | Multiplier (1.0 = single, 2.0 = double) |
| Space before / after | Vertical space in pt |
| First-line indent | In pt; reversed for RTL (indents from the right) |
| Columns | Multi-column layout; filled from the right in a right-to-left document |
| Drop cap | Drop cap lines |
| Border / Background | Rule and fill around the block |
| Arabic vowel signs | Show, Placement and Qurʾānic per style (Mode tab → Arabic vowel signs → For a style), see Arabic Vowel Signs |
Fonts
Classic 2 ships professional text families converted from OpenType, each with Regular, Bold, Italic and Bold Italic:
| Family | Scripts | Notes |
| Open Sans | Latin (incl. Vietnamese), Greek, Cyrillic | The default text font; ff, fi, fl, ffi, ffl ligatures |
| STIX Two Text | Latin, Greek, Cyrillic | Times-like book face, contextual f-forms, about 61,000 kerning pairs |
| Latin Modern Roman | Latin, capital Greek | The LaTeX look; ligatures and kerning |
| Liberation Mono | Latin | Monospaced; the code font |
| Amiri (Regular, Bold) | Arabic, Persian, Urdu and other Arabic-script languages, Qurʾānic marks | Classical Naskh for book typesetting, shaped with its own OpenType rules |
The earlier pixel example fonts (BitFont Latin and BitFont Arabic) have been removed. Documents that used them open with Open Sans and Amiri instead — styles, inline tags, headers, footers and footnotes included.
Loading and Assigning Fonts
- Open Sans and Amiri load on first launch. Switch other families on in Assets → Fonts: ticking one style loads all four, and the choice is remembered.
- Bold and italic text uses the sibling fonts automatically; a font a style names but that is not loaded yet is loaded in the background and the page is laid out again.
- Assign fonts per style and per language in the Styles tab, and the document defaults in the Doc tab.
- Characters that no loaded text font has (∎, arrows, ✓ …) are drawn from the math font, on the text's baseline.
- Templates load the fonts they use.
Additional Fonts
Additional fonts can be added to your account. Contact us to request a specific font or script.
Bilingual & RTL
Classic 2 is built for multilingual documents. With fonts for two scripts loaded (for example Open Sans and Amiri) everything below works without extra markup.
Fonts by Script
Every word is drawn in a font that has its script. Arabic typed into a Latin paragraph uses the Arabic font with its joining forms, and Latin inside an Arabic block uses the Latin font — also in the table of contents, headers, footers and footnote labels. No tag is needed.
Paragraph Direction and Word Order
- A language block (
<arabic>, <hebrew>) has the direction of its language; any other paragraph takes the direction of its first word with letters, so it does not flip while you type words of the other language.
- Each line follows the Unicode bidi algorithm: numbers stay left to right inside Arabic, formulas stay whole, punctuation takes the direction of the text around it, and brackets are mirrored in right-to-left text.
- Generated numbers read right to left in Arabic text: citations «[٣ ،٢]», list numbers «.١», section and equation numbers «٢-١».
- Right-to-left paragraphs have their indents, list bullets and numbers on the right; footnotes starting in Arabic have their number and separator on the right.
Right-to-left Documents
With Doc → Default language set to Arabic or Hebrew the whole document reads from the right:
- Every paragraph containing Arabic is right to left, even one starting with a Latin word or code; code blocks stay left to right.
- Header and footer zones are mirrored, columns fill from the right, tables put their first column on the right, and equation numbers stand on the left.
- Generated labels are Arabic: «صفحة» in the default footer, «المحتويات», شكل, جدول, المراجع, مبرهنة, برهان … (anything set by hand wins).
- Note, tip and warning boxes have their coloured bar on the right; slides put the number on the left.
- Reference lists: an Arabic entry is set in Arabic — full names, Arabic commas, «و» before the last author, «وآخرون» for et al., «مج ع ص» for volume, issue and pages.
Digits
Doc → Digits shows numbers — including page, list, section, footnote and equation numbers — as Western, Arabic-Indic or Persian/Urdu digits, everywhere or only in Arabic text ("١٩ مليون" but "COVID-19"). With Arabic-Indic digits the Arabic separators are used too: 12.5 → ١٢٫٥, 1,250 → ١٬٢٥٠, 38% → ٣٨٪.
Kashida Justification
Justified Arabic lines are filled first by lengthening letter joins with tatweel (ـ) at the best point of each word — after س ش ص ض, before a final ة ه د, before a final ا ط ل ك — at most three per word, and only then by widening spaces. The naskh rules are respected: never inside lam-alef, never right before ج ح خ, never between an initial ل and a final ن, and inside runs of teeth only where the denticle rules allow. The text itself is not changed. Turn it off with Doc → Justify Arabic with kashida.
Typing Arabic and RTL Text
- The status bar shows Typing: Latin | العربية. The active language follows the caret;
Ctrl+Shift+F or a click switches it.
- With Arabic active, a Latin keyboard types Arabic in the standard Arabic (101) layout by physical key (Shift for the vowel marks and أ إ آ ، ؟). With Persian as the document's language it types the Persian ISIRI 9147 layout, with Urdu a phonetic Urdu layout. A keyboard already set to Arabic types as it is.
- At a point where two languages meet, the active language decides which side spaces, digits and punctuation belong to.
- Arrow keys move the caret one position left or right on screen, through both directions; past the end of a line they continue on the next line in reading order.
- Backspace deletes the character typed before the caret and Delete the one after it, in every script.
- In an empty right-to-left paragraph the caret starts on the right.
Inline Language Switching
<paragraph>The word <ar>كتاب</ar> means "book" in Arabic.</paragraph>
Typing Arabic in a Latin paragraph wraps it in <ar> by itself; a paragraph that becomes mostly Arabic turns into an <arabic> block, and the <ar> inside it is removed. A language tag is always part of its paragraph, also when it is the paragraph's only content — the caret, vowel-sign moves and letter shapes belong to the paragraph. Documents in which <ar> tags got nested inside each other (<ar><ar>…) are repaired on the next edit of that paragraph: the extra tags go and their settings move to the paragraph.
Spelling, Search and Accessibility
- Spelling checks only words in the dictionary's script, so Arabic words are not marked by a Latin dictionary. Persian has its own dictionary; see Languages & Letters.
- Find & Replace can treat the forms of a letter as one (ي ی ى, ك ک …) and ignore vowel signs — Letters in the Find & Replace window.
- Arabic text is in the PDF's invisible text layer, so it can be searched and copied in reading order.
- In a tagged PDF, Arabic and Hebrew paragraphs get their own language so screen readers switch.
Arabic Templates
Every mode has templates written in Arabic, listed under العربية · Arabic in the template list: scientific article, two-column conference paper and thesis; feature and newsletter; chemical, food, shipping, shelf-edge and address labels (with Arabic merge data); tips and case-study carousels; API reference and user manual. Loading one sets the document to Arabic and right to left, with Amiri as base font and STIX Two Text or Open Sans as Latin font.
Arabic Shaping
Classic 2 shapes Arabic itself with the font's own OpenType layout tables, following the same order and rules as HarfBuzz (the shaping engine of browsers and LibreOffice):
- Letters with marks are composed where the font has the composed letter (alef + hamza → أ), and marks are put in the standard order.
- Each letter gets its joining form — isolated, initial, medial or final. ZWNJ breaks a join; tatweel and ZWJ join.
- Substitutions: required ligatures (lam-alef), contextual alternates, ligatures, and the stylistic sets chosen in the style.
- Positioning: cursive joins, kerning, marks on letters and on other marks.
- Arabic digits and the end-of-ayah sign form a number run (١ draws the number inside the sign).
- Denticles: runs of joined teeth (ب ت ث ن ي …) follow the naskh rules — never more than two low teeth in a row, raised teeth separated by at least two low ones — using the fewest raised teeth that satisfy the rules.
The result was checked against HarfBuzz on 12,000 Arabic, Persian, Urdu and Qurʾānic words: same glyphs, same positions. Letter spacing is never applied inside Arabic words.
Arabic Vowel Signs (Tashkeel)
Vowel signs are characters of the text; the settings only change how they are drawn, in the preview and the export. Find them in Mode tab → Arabic vowel signs (tashkeel).
Show
All vowel signs; none (hamza and madda stay — they are spelling); shadda only; all except case endings (the vowel, tanween or sukun of each word's last letter); the skeleton without dots; or the bare skeleton (rasm) — see Languages & Letters. When marks are hidden, the font's automatic vowels on الله are hidden too.
Spelling conventions: As typed (default), or As the font's style writes them — naskh writes fathatan on the letter before alef (اً → ًا), whatever order it was typed in.
Placement
| Option | What it does |
| Smart (default) | Each mark hugs its letter: it sits a small gap above (or below) the ink actually under it — its own letter, a tall neighbour such as alef, lam or kaf, a tail, or a mark already placed — and never drops below the top of the letter body, so a mark can never be read as a letter. Shadda comes first and the vowel sits on it. A mark that would have to climb over a neighbour moves a little sideways, staying over its own letter (inside a ligature, over its own part) |
| Level per letter group | As smart, then the marks of each connected letter group share one height above and one below — the calm, metal-type look |
| Font's positions, clashes moved apart | The font's anchors, with only collisions resolved |
| Exactly as the font places them | Identical to HarfBuzz |
- Gap above / Gap below — the clearance between a mark and the ink (default 0.06 em; below follows above unless set).
- Kasra below the shadda — Amiri's alternative form (ss05).
- Automatic vowels on الله — on by default; never applied in Persian, Urdu, Ottoman, Pashto or Kurdish text (
lang="fa" …), where ل-ل-ه is not the name of God.
- Line space and move marks away from the line above or below — lines with marks get exactly the room their ink needs (see Microtypography).
The Font's Alternate Shapes
Mode tab → Arabic vowel signs → Font's alternate shapes switches Amiri's stylistic sets on for the whole document: teeth after ر ز و in their alternative form (ss01), meem before alef (ss02), Arabic-style & and @ (ss03), keheh and gaf in the Arabic kaf shape (ss04), stretched joins as plain tatweel instead of the curved kashida (ss07), and the reversed Arabic comma and semicolon (ss08). For one passage use font-sets="ss01 ss04" on any element; for a single letter use the menu that appears while typing.
Qurʾān Mode
- Inline hamza (ء between two letters that join) sits on the joint and the letters stay joined: شَيْءًا, يَءُودُهُۥ.
- Room for crowded marks — where marks crowd, the joint is lengthened once, within the naskh limits (on in Qurʾānic text; turn off with
tashkeel-room="false").
- Sukun — the font's circle or the jazm (head of khāʾ, Cairo style).
- The small letters and signs of the Cairo text (ۥ ۦ ۨ, ۟ ۠, open tanween, small mīm) are drawn as the font has them; a missing sukun is never added.
Per Style and Per Passage
For a style: the same panel sets Show, Placement and Qurʾānic for one style. For a passage: attributes on any element, inner ones winning:
<blockquote quran="true" tashkeel-sukun="jazm">…</blockquote>
<paragraph>… <span tashkeel="none">…</span> …</paragraph>
| Attribute | Values |
tashkeel | all, none, shadda, no_endings, nodots, rasm |
tashkeel-conventions | as-typed, style |
font-sets | the font's stylistic sets, e.g. ss01 ss04 |
tashkeel-placement | smart, group, fix, font |
tashkeel-gap, tashkeel-gap-below | clearance in em, e.g. 0.08 |
quran | true / false |
tashkeel-sukun | font, jazm, circle |
tashkeel-room | true / false |
tashkeel-kasra-below-shadda, tashkeel-allah | true / false |
lang | e.g. fa, ur, ota, ckb, ug, ms-Arab, ps |
Adding Vowel Signs — Step by Step
- Click in the Arabic text just after a letter, or select several letters.
- Open the palette: حَ in the toolbar or
Ctrl+Shift+H. It shows the letter it will change, in blue inside its word (at the start of a word: the letter after the caret), and lights up the signs that letter already has. Drag the palette by its title bar to move it — it stays where you leave it.
- Click a sign to add it; click it again to remove it. A vowel replaces the letter's vowel; shadda, madda, hamza and the Qurʾānic signs switch on and off. With a selection, the sign goes on every letter.
- Use ◄ next letter / previous ► to move along the word without leaving the palette.
- If a sign sits wrong, move it by hand (below).
Palette groups: Vowels, Tanween, Signs (shadda, dagger alef, madda, hamza above/below) and Qurʾān (jazm, rounded zero, open tanween, small mīm). Marks are kept in the correct order, and each change is one undo step. With an Arabic keyboard — or Typing: العربية on a Latin keyboard — the signs can also be typed directly with Shift. The Mode-tab settings only change how signs are drawn, never the text.
Below the signs: the letters of the document's language, Normalise letters and Type in Latin letters — see Languages & Letters.
Letter Variations While Typing
After each letter you type, a small menu appears just above it showing its variations, numbered, and a light blue box marks the letter it works on; it follows the caret as you type. It also opens when you put the caret after a letter (click, arrow keys, Home/End) or click on a letter — then it works on the letter whose ink is under the pointer, also when you click its dots or hamza, which may stand over the next letter or reach the line above, and the caret moves just after that letter. The tatweel added by justification never counts as part of a letter. The toolbar button ك▾ (Alt+L) opens or closes it for the letter at the caret at any time, and the AUTO switch next to it turns automatic opening on and off (remembered in this browser); with AUTO off the menu opens only from the button. For Arabic, its second row has the vowel signs (fatha, damma, kasra, sukun, shadda, tanween, dagger alef, madda, hamza): click one to put it on the marked letter, click it again to remove it; ✕ removes them all. The menu never takes the cursor away from your text — you keep typing where you were. Move the caret with ← → and the menu and its box move to the letter before the cursor. ↑ / ↓ choose a piece of the letter: the letter itself, each of its vowel signs, and the letter's own dots, hamza, madda or the small sign of ك (the two dots of ق, the dot of ن, the hamza of ؤ …, also where the font draws the dots as a glyph of their own, as Amiri does for many letters inside a word) — the box frames the chosen piece and the label names it. The third row moves the chosen piece: Move … with ← ↑ ↓ → and a button showing its offset (click it to put the piece back); with the letter itself chosen, the row moves all its signs Above or Below instead. Step chooses how far one click moves — 1 %, 2 %, 5 % or 10 % of the type size. With a piece chosen, Alt+arrows move it and Alt+0 puts it back. A chosen dot, hamza or other part can also be deleted — the Delete button, or the Delete / Backspace key while the part is chosen (the text is not touched); its offset button then shows deleted, and clicking it or Alt+0 brings the part back. With the letter itself chosen, No dots (Alt+N) draws a dotted letter as its bare body without its dots (nuqta) — ب as ٮ, ج as ح, ش as س, ق as ٯ — and again with them on a second press; hamza, madda and the sign of ك are kept. Select some text and press Alt+N, or click the ٮ button in the toolbar (next to ك▾ and AUTO), to do it for every dotted letter of the selection; press again to bring the dots back. The selection can run over several paragraphs, headings and other elements: drag on from one into the next ones, or click in the first and Shift+click in the last (Shift+click on an element's margin arrow takes that element whole) — the whole stretch is lit and all of it changes at once. Typing or moving the caret then works in the element you started in, as before. Type without dots on the menu's last row (remembered in this browser) makes every dotted letter you type from then on appear without its dots. The letters stay the real letters in the text, so search, copy and spelling still see ب ت ث; only the drawing changes, on screen and in the PDF and image exports. While the pointer is on the menu it stays where it is, so the buttons stay under the pointer while the piece moves; a button acts as soon as it is pressed, so quick repeated clicks all count. Press Alt+1 … Alt+9 (Alt+0 for the tenth), or click, to replace the letter with that variation. The menu stays while you use it; it closes after 8 seconds without typing (never while the pointer is on it), when you type a space or leave the letter, and with Esc. Ctrl+Shift+number also works where your system allows it — on Windows, Ctrl+Shift switches the keyboard layout, so use Alt.
The menu's ▾ button, Alt+V or the toolbar à open the full character table. Its Arabic script tab has the letters, hamza and alef forms, vowel signs, Arabic-Indic and Persian digits, Arabic punctuation, tatweel and the joining controls (ZWNJ, ZWJ), signs such as ﷲ ﷺ , and the extra letters of Persian, Urdu, Ottoman, Kurdish, Uyghur, Jawi and Pashto. It also lists the Qurʾānic signs and every other Arabic-script letter the fonts have. The All accents tab lists every accented Latin letter of the font, by base letter. While you type Arabic it opens on that tab. The Arabic letter forms tab shows every letter in each of its forms (isolated, initial, medial, final), with the font's alternate shapes and the stroke and tail lengths in dots — everything the small menu offers — and inserts the letter in the shape you click.
The cursor never stops between a letter and its vowel signs: the arrow keys step over a letter with its signs as one, so typing never splits them.
- Arabic — auto (the form its neighbours give it), isolated, initial, medial and final, drawn in the font's own shapes — only the forms that keep the word joined where the letter stands (inside a word that is the letter as it is, at the end of a word also the connecting form); forms that cut a word, such as an isolated letter inside it, are in the full table's Arabic letter forms. The chosen form is stored in the text itself (with invisible joining marks, only where a neighbouring letter needs them, and after the letter's vowel signs), so it stays in the PDF and every export and typing on after it keeps the word joined; auto removes it. Letters that join only to the right (ا د ذ ر ز و …) have just isolated and final forms. Next come the font's own alternate shapes of that letter, drawn as they would look — in Amiri, for example, ک and گ in the Arabic kaf shape, or the other form of teeth after ر ز و; the choice applies to that one letter and stays with it when you edit. For letters whose tall stroke or tail reaches into the next line (ا ل ك ط … above, ي ع م ح … below) the menu also offers the length in dots of the font: full, then half-dot steps such as 5↑ 4½↑ 4↑ for a tall stroke (height above the baseline) or 3↓ 2½↓ for a tail (depth below it) — only the steps that letter's shape allows.
- Latin — the accented forms of the letter you typed (e → è é ê ë ē …).
Moving Marks by Hand
Alt+arrows move the marks above the letter before the caret, Alt+Shift+arrows the marks below (0.02 em per step), and Alt+0 returns them to automatic — also with the arrows in the palette. With a sign, dots or hamza chosen in the letter menu, the same keys move only that piece. The moves are stored with the text — signs in the tashkeel-nudge attribute, a letter's dots, hamza or madda in glyph-part — and follow the letter when you edit the text before it. They are in the PDF and image exports too.
Stroke Length of a Letter (ا ل ك ط … and ي ع م …)
Lengths are measured in dots (nuqta) of the font, the calligrapher's unit — in Amiri a dot is 0.132 em; alef is about 5½ dots tall, lam 5½, kaf 6, the letter body (ه) 2½. Put the caret on a letter and open the letter menu or the palette (حَ, Ctrl+Shift+H, section Stroke length of the letter):
- Auto (default) — full length, shortened only where it would touch another line, exactly as much as needed (see Microtypography).
- Full — always full length, never shortened.
- Tall stroke — half-dot steps of its height above the baseline (alef: 5, 4½, 4). At the same step alef and lam end at the same height, so a line stays even.
- Tail — half-dot steps of its depth below the baseline (م: 3½, 3, 2½, 2).
Each letter offers only the steps its shape allows. The shorter letters are not squeezed: the straight part of the stroke gets shorter along its own direction — kaf keeps its slant — while the head, the terminal, the join, a tail's sweep, and hamza or dots keep exactly the font's shape. Every length is a blend between the full letter and its shortest form, as in a variable font. Letters whose stroke cannot get shorter cleanly (the two halves of لا, shallow tails) are not shortened. The choice is saved with the text (glyph-stroke="12:4.5d" for a tall stroke, 2.5b for a tail, both 4.5d+2.5b) and stays with the letter when you edit the text around it. Documents with the earlier 70 % / 45 % lengths keep working. The text itself is unchanged.
Microtypography: Collisions Between Lines
After layout, every glyph of a page — text and formulas — is checked against the line above and below, column by column of its outline. Where ink comes closer than the clearance, the glyph that can give way does:
- A shorter replacement glyph — only the part above the letter body (the tall stroke of ك ل ا ط, Latin ascenders b d f h k l t, accents) or below the baseline (tails of ي ن ح ع م ر و, dots below, Latin descenders g j p q y) is drawn shorter, as much as needed. Amiri ships, for every letter form that can reach another line, its shortest clean form; the replacement is a blend between the letter and that form, exactly as short as the collision needs. Short letters are never squeezed: a tall stroke loses a piece of its straight part along its own direction and its head moves with it, so the stroke width, the terminal and the join stay exactly as the font draws them; a tail gets shallower while its stroke keeps its thickness, and dots and hamza move as a whole. Letters without a clean shorter form keep their shape (then only masking). For other fonts the replacement is derived from the outline.
- Masking — only when a stroke would have to shrink more than the language allows, the rest is cut at the clearance line (never more than a quarter of the letter's height).
- Formulas never change, and vowel signs are never cut: a mark first moves toward its own letter; if it would still touch the other line, that line gets the room it needs.
- The lower letter gives way before the upper one.
- The text is unchanged, and the preview, PNG and PDF show the same result.
Settings: Mode tab → Typography → Fix collisions between lines, Clearance (em) (default 0.12) and Mask what a shorter letter cannot avoid.
To choose the stroke length of a single letter yourself, see Stroke Length of a Letter.
Table of Contents
Works the same as Classic, including <toc/> placement and exact page numbers. An entry whose text starts with Arabic or Hebrew is mirrored: text and indent on the right, leaders, page number on the left. With the document language set to Arabic the default title is «المحتويات».
Assets
Images
Upload PNG, JPG, or JPEG images and reference them with the <image> tag or inside a <figure>. Same syntax as Classic.
Fonts
Switch font families on and off in Assets → Fonts. See Fonts.
AI Assistant
The AI assistant works the same as Classic and can make all the same document setting changes. In Classic 2 it also understands:
- Typography role feature settings (ligatures, stylistic sets, figure styles)
- Master page definitions (margins, backgrounds, headers per page type)
- Language slot assignments
Example prompts
Apply Chicago style 17th edition
Set page to A4 with 25mm margins
Make headings use oldstyle figures
Enable contextual alternates on body text
Create a chapter-opener master page with a 120pt top margin
Set the Arabic font to Amiri
Export
Classic 2 exports to SVG and PDF. The SVG export is a direct output of the rendering engine. The PDF is generated from the SVG output as vector outlines.
- All fonts are embedded in the export — no external dependencies.
- Arabic, Hebrew, and mixed LTR/RTL documents export correctly, with an invisible text layer for searching and copying (Arabic included, in reading order).
- Short replacement glyphs, moved vowel signs and kashida are part of the vector output.
- What you see in the preview is exactly what exports.
- Links, bookmarks, Tagged PDF and PDF/A-2 — see PDF Output Options.
PDF export is available on paid plans.
Projects
Save and load projects the same way as Classic — as .json files that store your XML content, style settings, font assignments, mode settings, bibliography and the tashkeel and microtypography settings. Projects are saved locally and not stored on the server. Projects that used the removed pixel fonts switch to Open Sans and Amiri when opened.
XML Reference
Standard Block Tags
All Classic block tags, and the tags of every mode, work in Classic 2. In addition, Classic 2 recognises language block tags:
| Tag | Script | Direction |
<english> | Latin | LTR |
<arabic> | Arabic | RTL |
<hebrew> | Hebrew | RTL |
<french> | Latin | LTR |
<german> | Latin | LTR |
Inline Language Tags
| Tag | Script |
<ar> | Arabic inline |
<he> | Hebrew inline |
<en> | Latin inline |
<fr> | French inline |
<de> | German inline |
Arabic Attributes
Any element accepts the tashkeel attributes (tashkeel, quran, tashkeel-placement …, see Arabic Vowel Signs) and lang. tashkeel-nudge is written by the editor when you move marks by hand.
Image Tag
<image src="filename.jpg" width="150" height="100" />
Footnote Tag
<paragraph>Text<fn>Footnote.</fn> continues.</paragraph>
Well-formed XML Rules
- Every opening tag must have a closing tag
- Tags must be properly nested — they cannot overlap
- Self-closing tags end with
/>
- Attribute values must be quoted
- Use
& for &, < for <, > for >
Features in both editors Everything below works the same in Classic and Classic 2 unless noted.
Editor Modes
The start screen offers Classic Editor and five modes. Every mode works on both engines and opens with its own templates, styles and Mode-tab panel. The current mode is shown as a badge in the toolbar.
- Changing mode — use ← Menu in the toolbar (or ← Back to menu in the Mode tab). It asks for confirmation and opens the start screen; picking a card starts that mode with a new document. ← Return to the open document goes back without changes. Save first if you need the current document.
- Nothing is locked — every tag from every mode works in any document, and all panels stay available.
- Templates — Mode tab → Template → Load replaces the document with a complete template: all settings are reset first, so nothing is left over. Uploaded images and fonts stay.
| Mode | Templates |
| Scientific Paper | Journal article (A4) · Two-column conference paper (Letter) · Thesis / dissertation |
| Magazine Editor | Feature spread · Newsletter |
| Label Creator | Chemical product (GHS/CLP) · Food product (EU 1169/2011) · Shipping 4 × 6 in · Pallet A5 · Shelf-edge strips · Label strips · Cable wraps · Rack locations · Asset tags · File spines · Address labels |
| Carousel Creator | Tips carousel · Story / case study |
| Technical Docs | API reference · User manual |
Classic 2 additionally offers Arabic versions of the templates (see Bilingual & RTL).
The templates are complete examples about PubSuite and TeX: the scientific ones typeset mathematics with a figure of the Gaussian curve (plot.png, added to Assets with the template — replace it with your own image), the magazine feature is about grids and TeX's line breaking, the newsletter is PubSuite project news, the carousel case study moves a lab from LaTeX templates to PubSuite, the API reference documents an example typesetting service (XML in, PDF out — not a real service), and the user manual explains PubSuite itself. The labels carry example product data.
Settings shared by every mode appear below the mode's panel, in collapsible sections:
| Section | Settings |
| Typography | Document language, hyphenation exceptions, baseline grid; Classic 2: collision fixing |
| Arabic vowel signs | Show, placement, Qurʾān mode … (Classic: Show only) |
| Footnotes & endnotes | Numbering (continuous / per page / per chapter), chapter tag, mark styles, endnotes title |
| Floats | Let figures and tables float; top and bottom shares; space around floats |
| Tables | Grid or booktabs rules; repeat header rows; "continued" label |
| Glossary, acronyms & index | First-use form, titles, page numbers, index columns and letter headings |
| PDF output | Clickable links, Tagged PDF, PDF/A-2 |
Mathematics
Formulas are typeset with TeX's rules and the font's OpenType MATH data, in the preview and as vector outlines in the PDF.
Writing Formulas
| Form | Result |
$x^2$, \(x^2\) or <math>x^2</math> | Inline formula |
$$…$$ or \[…\] inside text | Display-style formula kept in the line |
<equation>…</equation> | Numbered display equation |
<equation*>, <displaymath>, block-level <math> | Unnumbered display equation |
<math><mrow>…</mrow></math> | Presentation MathML |
<eqref ref="id"/> | Equation reference: (3) |
<paragraph>Euler's identity $e^{i\pi} + 1 = 0$ links five constants.</paragraph>
<equation id="eq:gauss">\int_{-\infty}^{\infty} e^{-x^2}\,dx = \sqrt{\pi}</equation>
<equation env="align">
a &= b + c \label{eq:first} \\
d &= e + f \notag
</equation>
Write \$ for a literal dollar sign. In XML, & is written & and < as <.
Equation Attributes
| Attribute | Meaning |
id (or label) | Label for <eqref>, <xref> and \eqref{…} |
env | align, gather, multline, alignat, flalign, eqnarray (and starred forms) |
numbered="false" / nonumber | No number |
tag | Custom label, e.g. tag="*" gives (*) |
align | center or left for this equation |
Supported TeX
- Structures — fractions (
\frac \dfrac \tfrac \cfrac \genfrac), binomials, roots, scripts and primes, \sideset \overset \underset \stackrel \substack \limits \nolimits.
- Operators — big operators with limits, integrals with italic correction, named operators (
\sin \lim \argmax …), \operatorname, \DeclareMathOperator, \bmod \pmod.
- Delimiters —
\left … \middle … \right with all standard delimiters, \big … \Bigg; they stretch with the font's size variants and glyph assemblies.
- Accents — all standard and wide accents, over/under arrows, braces and brackets with labels,
\overline \underline; stretchy arrows \xrightarrow[below]{above} and relatives.
- Environments — matrices (
pmatrix bmatrix vmatrix …), cases, array, aligned gathered split, align gather multline …, rules (\hline \toprule \midrule), \intertext; page breaks between rows of align.
- Per-row control —
\tag \tag* \label \notag \nonumber; \ref and \eqref inside formulas.
- Letter styles —
\mathrm \mathbf \boldsymbol \mathcal \mathscr \mathfrak \mathbb \mathsf \mathtt and more, mapped to Unicode so copied text keeps the style.
- Text and spacing —
\text \textbf \emph \mbox, all TeX spaces, \phantom \smash \raisebox; TeX's spacing table between atom classes.
- Decoration —
\color \textcolor \colorbox \boxed \cancel \bcancel \xcancel \not.
- Symbols — full Greek, AMS relations, arrows and operators; any Unicode math character can be typed directly.
- Macros —
\newcommand \renewcommand \providecommand \def in a formula, or for the whole document in the Math tab.
- Commutative diagrams —
tikzcd (directions, labels, hooks, dashed, bends, shifts …) and amscd CD (@>>>, @VVV, @= …).
- Chemistry (mhchem) —
\ce{2H2 + O2 -> 2H2O}, charges, states, isotopes, reaction arrows with labels; \pu.
- Units (siunitx) —
\num \si \SI \qty \unit \ang, ranges and lists, with exponents, uncertainties, digit grouping and decimal marker.
Script-size alternates (ssty) are used in superscripts and subscripts where the font has them, as LuaLaTeX does.
Line Breaking
- Inline formulas may break after a relation or binary operator at the outer level, as TeX does.
{…} around a formula prevents breaking.
- Display equations wider than the text break automatically at relations (then operators); continuation rows are indented and the number goes on the last row.
Numbering and References
- Numbering: Auto, Number every equation, or No numbers. In
multline only the last row is numbered; in align/gather every row.
- Format:
({n}) by default; {s} is the section counter, so ({s}.{n}) gives (2.3). Number within a tag restarts the counter.
- Position: right or left; a number that doesn't fit moves to its own line.
- References resolve forward too; unknown labels print (??). A reference prefix such as "Eq. " can be set.
Math Tab Settings
| Setting | Values | Default |
| Math font | STIX Two Math, Latin Modern Math (Classic: also uploaded math fonts) | STIX Two Math |
| Letter style | TeX, ISO, French, Upright | TeX |
| Size vs. text / Display size | multipliers | 1.0 |
| Numbering, Format, Number within, Position | see above | Auto, ({n}), none, Right |
| Alignment / Left indent | Centred or Flush left (fleqn); indent in pt | Centred, 24 |
$…$ syntax | on/off | on |
| Colour, Reference prefix, Macros | text | empty |
| Inline breaks / Display breaks | on/off | on |
| Decimal marker, Units with "per", Group digits | siunitx output | point, power, on |
Equation Editor
- Open it with the ∑ toolbar button,
Ctrl+M (inline) or Ctrl+Shift+M (display), or the Math tab buttons.
- Inline / Display switch; for display equations, Layout picks single line,
align, gather or multline.
- A TeX source field, a symbol palette (Structures, Greek, Operators, Relations, Arrows, Fonts, Matrices, Spacing) and a live preview with error messages.
Ctrl+Enter or Insert puts the formula at the cursor; Esc closes.
- Double-click any formula in the preview — also in a footnote — to edit it; Insert replaces exactly that formula.
- In the text a formula behaves as one character, and spell check skips it.
Scientific Paper
Front Matter and Sections
<title>, <subtitle>, <authors>, <affiliation>, <date>, <abstract> and <keywords> come with default styles. Headings with a section level in their style are numbered; give a heading an id to refer to it.
Theorems and Proofs
<theorem> <lemma> <proposition> <corollary> <definition> <remark> <example> <conjecture> <claim> <exercise> <problem> <solution> <assumption> <axiom> <notation> <hypothesis> and <proof>. Attributes: id, name (text in parentheses), label (replaces the word), numbered="false". All share one counter. Theorem bodies are italic (plain style) or upright (Mode tab → Theorem body).
End of proof: every proof ends with a mark — the black square ∎ (the "tombstone" or QED sign). Mode tab → End of proof chooses ∎ solid (default), ▪ small solid, □ open (as LaTeX's amsthm) or none, and Placed puts it after the last word or at the end of the line as LaTeX does (on the left in a right-to-left paragraph).
Cross-references
See <xref ref="sec:method"/>, <xref ref="fig:plot"/> and Theorem <xref ref="thm:cs" fmt="num"/>.
fmt | Output |
auto (default) | Figure 3, Table 1, Section 2.1, Theorem 4, Eq. (5) |
num | 3 |
label | Figure |
page | the target's page number |
full | Figure 3 on page 7 |
<pageref ref="id"/> prints the page of any labelled element; <autoref> equals fmt="auto". Page numbers and forward references are exact — the document is laid out again until every number is stable, like LaTeX's .aux passes.
Algorithms and Code Listings
<algorithm id="alg:euclid" caption="Euclid's algorithm" linenumbers="true">
Input: integers $a, b$
while $b \ne 0$ do
$t \gets b$
$b \gets a \bmod b$ // remainder
$a \gets t$
end while
return $a$
</algorithm>
Leading spaces set the indentation, keywords are bold, // starts an italic comment, and the block is numbered Algorithm 1 with rules like the LaTeX package. <codeblock caption="…" id="lst:x"> is numbered Listing 1.
Mode Settings
Citation style, include all entries, references title, the words Figure / Table / Eq. / Section, caption separator, numbering of figures and tables within a tag, theorem body style. The Insert buttons add an equation, figure, table, theorem, lemma, definition, proof, citation, cross-reference, reference list, abstract, subfigures, algorithm, list of figures or tables, or endnote.
Citations & Bibliography
<cite key="knuth1984"/> → [1] or (Knuth, 1984)
<cite key="knuth1984,lamport1994"/> → [1, 2]
<cite key="knuth1984" page="p. 12"/> → [1, p. 12]
<cite key="knuth1984" prefix="see"/> → [see 1]
<citet key="knuth1984"/> → Knuth [1] or Knuth (1984)
<cite key="knuth1984" mode="author"/> → Knuth
<cite key="knuth1984" mode="year"/> → 1984
Citation Styles
- Built in: Numeric (IEEE), Vancouver, Numeric sorted A–Z, Author–year (APA), Author–date (Chicago), Alphabetic [Knu84]. Same-author same-year works get 1984a, 1984b; citations of one author merge.
- Journal styles (CSL): ACS, AMA, APS, ASA, APA 7th, ACM, BMJ, Cell, Chicago, Elsevier, Harvard, IEEE, ISO-690, MLA, Nature, PLOS, Science, Springer, The Lancet — processed by citeproc-js, the engine of Zotero and Mendeley.
- Upload .csl style adds any style from the Zotero Style Repository; it is saved with the project.
- The citation language follows the document language (27 locales).
- Sort list: as the style says, order of citation, name–year–title, name–title–year, year–name–title, newest first.
- Back-references: entries can end with the pages that cite them (Cited on pages 2, 5.), as links.
Bibliography
- Mode tab → Citations: paste BibTeX, add / edit / delete entries one by one, Export .bib. Every entry type is read; LaTeX accents become Unicode.
<references/> prints the list — only cited entries unless Include all. Attributes: title, notitle, titletag.
- Entries can also be written in the document:
<references><bibitem key="x">Formatted text.</bibitem></references>.
- Numeric styles hang the text after the label; author–year styles use a hanging indent.
Figures, Tables & Floats
Figures and Subfigures
<figure id="fig:plot" layout="row" align="center" placement="tb">
<img src="a.png" width="45%" alt="Before"/>
<img src="b.png" width="45%"/>
<caption>Two panels side by side.</caption>
</figure>
<figure id="fig:panels" cols="2">
<subfigure id="fig:left"><img src="a.png"/><caption>Before.</caption></subfigure>
<subfigure id="fig:right"><img src="b.png"/><caption>After.</caption></subfigure>
<caption>Two states of the system.</caption>
</figure>
A figure with a caption is numbered (Figure 1:). layout is row or column; kind="table" numbers it as a table; short="…" gives a shorter list entry; alt describes the image in a tagged PDF. Subfigures are lettered (a), (b) and referenced as Figure 1b.
Tables
<table id="tab:res" style="booktabs" colalign="l d d" widths="2* 1* 1*">
<caption>Results</caption>
<thead><tr><th>Method</th><th>Mean</th><th>SD</th></tr></thead>
<tr><td>Baseline</td><td>12.5</td><td>0.83</td></tr>
<tr><td>Ours</td><td>9.25</td><td>1.2</td></tr>
<midrule/>
<tr><td colspan="3" align="center">n = 40 per group</td></tr>
</table>
| Attribute / element | Meaning |
<thead>, <tbody>, <tfoot> | Header and body rows |
style="booktabs" | Heavy top and bottom rules, light rule under the header, no vertical lines |
<midrule/> | Extra light rule between rows |
widths | Lengths, % or shares such as 2* |
colalign | l, c, r, j, d (decimal-aligned) |
colspan, align on a cell | Span and alignment per cell |
repeatheader="false" | Don't repeat the header on following pages |
Long tables break between rows, repeat the header and print Table 3 (continued). Cells can contain formulas, citations, references and formatting.
Floats
Numbered figures, tables and algorithms float as in LaTeX, with placement (default htbp):
| Letter | Meaning |
h | Here, if it fits |
t / b | Top / bottom of a page |
p | A float page of its own |
! | Ignore the top and bottom shares |
H | Exactly here, never floats |
<floatbarrier/> places every waiting float; <clearpage/> also starts a new page. Floats stay in place inside columns, frames and slides.
Notes, Glossaries & Index
Endnotes
<endnote>…</endnote> prints a mark and collects the note; <endnotes/> prints the notes collected so far under Notes. Leftover notes are printed at the end. Endnote marks have their own counter and format.
Lists of Figures, Tables, Listings and Algorithms
<listoffigures/>, <listoftables/>, <listoflistings/>, <listofalgorithms/> print captions with page numbers and dot leaders (links in the PDF).
Glossaries and Acronyms
<glossarydefs>
<acronym id="pdf" short="PDF" long="Portable Document Format"/>
<glossentry id="kerning" name="kerning">adjusting the space between two letters</glossentry>
</glossarydefs>
<paragraph>Export as <ac ref="pdf"/>; later <ac ref="pdf"/> is short. Good <gls ref="kerning"/> matters.</paragraph>
<printacronyms/>
<printglossary/>
<ac> prints the full form the first time and the short form after; <acs> <acl> <acf> force short / long / full, <acp> etc. give plurals, <gls> <glspl> <glsdesc> print terms and descriptions. Lists show only used entries (unless all="true"), sorted, with the pages where each is used.
Index
<index term="…"/> or <index>word</index> with makeindex syntax: sort@Display, levels with !, |see{…}, |seealso{…}, |textbf. <autoindex/> prints the index with letter headings, sorted for the document language, merged page ranges, in two columns.
Magazine Editor
Frames
A frame is a box placed at an exact position; the text flows around it.
<frame page="1" x="0" y="0" w="50%" h="60mm" wrap="around" gap="4mm">
<img src="photo.jpg"/>
</frame>
<frame x="55%" y="20mm" w="45%" bg="#f2f4f8" padding="4mm" border="0.5">
<paragraph>Frame text.</paragraph>
</frame>
| Attribute | Meaning | Default |
page | Page number | current |
x, y | Position (% of the content area or page) | from align / current position |
w, h | Size; without h the height fits the content | 40%, content |
align | left, center, right when x is omitted | left |
relative | margin or page (for bleed images) | margin |
wrap | around, top-bottom, none | around |
gap, padding | Space to the text; inner padding | 8 pt, 0 |
bg, border, bordercolor | Background, border width, border colour | none |
Drag a frame in the preview to move it and its corner handle to resize; the values are written back to the XML in millimetres, snapping to the layout grid when Snap is on.
Story Tags and Settings
The Story styles buttons insert <kicker>, <headline>, <deck>, <byline>, <subhead>, <pullquote> and <sidebar>; Column block inserts <columns2>. Settings: grid columns (6), gutter (12 pt), show grid, snap to grid, and facing-page spread view (pages in pairs as in print).
Label Creator
Label Sizes and Sheets
Label size sets the page to one label: 4 × 6 in and 100 × 150 mm shipping, 100 × 50 mm product, address sizes (70 × 37, L7160, 5160), L7165 parcel, 50 × 25 mm asset tag, L7166 food, 70 × 39 mm shelf strip, 100 × 24 mm tape, 25 × 38 mm cable wrap, L4761 spine, A5 pallet, or custom. Sheet layouts place labels on a printed page — one per page for thermal printers, A4 and Avery layouts, or a custom grid.
Barcodes
<barcode type="ean13" value="590123412345" height="18mm" align="center"/>
<barcode type="gs1-128" value="(01)09501101530003(17)261231(10)AB-123"/>
<qrcode value="https://example.com" size="25mm" ecl="Q"/>
Types: code128, gs1-128 (FNC1 added automatically), ean13 and ean8 (check digit calculated or verified), upca, code39, itf, itf14, qr. Attributes: height, width, module, size, ecl, text, fontsize, quiet, align, color. Barcodes can also sit inline in text. All symbologies were verified with standard readers.
Hazard and Handling Symbols
<hazard code="GHS02 GHS07" size="15mm" align="left"/>
<paragraph>Keep away from heat <icon code="keep-away-from-heat"/>.</paragraph>
GHS01–GHS09 pictograms, and this-way-up, fragile, keep-dry, keep-away-from-heat, recyclable, ce — vector paths in preview and PDF.
Data Merge
Mode tab → Data merge → Load .csv (or paste) and Apply CSV. Placeholders work in text, barcode values and attributes:
| Placeholder | Result |
{{Column name}} or <field name="Column name"/> | That column of the current record |
{{serial}}, {{serial:6}} | Running number (padded: 000042) |
{{row}} | Record number |
{{date}} | Today, YYYY-MM-DD |
Values may contain <b>, <i>, <u> (e.g. allergens in bold). Preview record shows any record; Copies per record repeats each one. In the preview a merged value is clicked and selected as one unit — what you edit is the template with its placeholders, so typing never breaks a placeholder.
Export
- ↓ One label per page — a PDF with one page per label, for thermal and roll printers.
- ↓ Print sheets (PDF) — the labels placed on the chosen sheet, with optional crop marks.
The templates follow the rules for each label type (EU CLP/GHS, EU 1169/2011, GS1 logistic label, EU unit prices) and include sample merge data. They are starting points: check the rules for your product and market before printing labels that must comply.
Carousel Creator
<slide><kicker>Topic</kicker><slidetitle>A strong headline</slidetitle><slidesub>Subtitle</slidesub></slide>
<slide><slidebig>01</slidebig><slidetitle>Point title</slidetitle><slidetext>Supporting sentence.</slidetext></slide>
<slide image="photo.jpg" overlay="rgba(0,0,0,0.45)" valign="bottom"><slidetitle>Caption over photo</slidetitle></slide>
- Every
<slide> is a page. Attributes: bg, image (full bleed), overlay, valign, master.
- A footer is added automatically: handle, swipe hint (except on the last slide) and n / N; each part can be turned off.
- Sizes: LinkedIn and Instagram 1080 × 1080 and 1080 × 1350, Stories / Reels 1080 × 1920. Themes: Midnight, Paper, Ocean, Sunset, Mono, Forest.
- Export: ↓ PDF (LinkedIn document), or ↓ PNG slides (.zip) at 1× or 2×.
Technical Docs
Code
<codeblock lang="python" title="example.py" linenumbers="true">
def answer():
return 42
</codeblock>
Syntax highlighting for JavaScript/TypeScript, Python, C, C++, C#, Java, Kotlin, Swift, Go, Rust, PHP, SQL, Bash, CSS, JSON, YAML and XML/HTML; themes Light, Dark and Print. caption numbers the block as a listing. Inline: <code>, <kbd>, <samp>, <tt>, <var> — skipped by spell check and hyphenation.
Callouts and API Endpoints
<warning title="Heads up">Back up your data first.</warning>
<endpoint method="POST" path="/v1/orders">
Creates an order.
<param name="items" type="array" in="body" required="true">Line items.</param>
</endpoint>
Callouts: <note> <tip> <important> <warning> <caution> <danger> <info> <success> or <callout type="…">. Endpoints show the method in colour, the path in monospace and a parameter list.
Markdown Import
Mode tab → Import Markdown, then Append or Replace document. Supported: YAML front matter, headings with {#id}, fenced code, $$ and $ math, bold/italic/strike/code, links and images, footnotes, GitHub callouts (> [!NOTE]), pipe tables, nested and task lists.
Justification & Hyphenation
Justification
Justified text is set the way TeX sets it, and glyphs are never stretched or squeezed:
- Whole-paragraph line breaking (Knuth–Plass): the set of breaks with the most even word spacing wins; uneven neighbouring lines, consecutive hyphens and a hyphen on the last full line are penalised.
- TeX's passes: first without hyphenation, then with it, then with emergency stretch. Most paragraphs need no hyphens at all.
- Word spaces stay between each style's Min %, Desired % and Max %.
- Letter spacing only as a last resort, at most Extra letter spacing when justifying (default 2 %; 0 turns it off).
Hyphenation
- TeX hyphenation patterns for 81 languages (Liang's algorithm, as LaTeX), loaded when a language is first used.
- Language:
lang on a block, else the style's hyphenation language, else the document language.
- Per style: minimum word length and characters before / after the hyphen.
- Exceptions (Mode tab → Typography):
ta-ble marks allowed breaks; a word without hyphens is never broken.
- Formulas, code and URLs are never hyphenated.
Baseline Grid
Mode tab → Typography → Baseline grid puts every body line on a regular grid so facing columns and pages line up. Step (0 = body leading), offset, and grid lines in the preview. Headings, equations and figures keep their position; the text after them resumes on the next grid line. Styles can join with Align to grid; blocks can leave with grid="false".
Languages & Letters
Both editors know the languages written in Arabic script — Arabic, Persian, Urdu, Ottoman Turkish, Sorani Kurdish, Uyghur and Malay in Jawi — besides English and Hebrew. Each language brings its own words for the parts the editor writes for you, its digits, keyboard, letters and spelling. New languages can be added without updating the editor.
The Document's Language
In Classic 2, Doc → Default language lists these languages after Latin, French and German; the Arabic templates choose Arabic. The language sets:
- The words the editor writes — Theorem, Proof, Figure, Table, References, Contents, the page word in the footer, the blank-page note and box titles (Note, Warning …), in that language.
- Digits — Persian and Urdu start with ۰۱۲۳۴۵۶۷۸۹, Arabic with ٠١٢٣٤٥٦٧٨٩ (change it under Doc → Digits).
- The keyboard — typing Arabic script on a Latin keyboard follows the Arabic 101, Persian ISIRI 9147 or a phonetic Urdu layout (Classic 2).
- The font's language forms — for example Urdu digit shapes in Amiri (Classic 2).
A passage in another language takes a lang attribute: <span lang="fa">…</span>. A bibliography entry written in Arabic script is formatted in its language — Arabic with و، وآخرون، مج، ع، ص and «» quotes, Persian, Urdu and Ottoman with their own words. The entry's langid or language field chooses the language when it is not the document's.
Forms of the Same Letter
Unicode writes some letters in more than one way: ي ی ى, ك ک, ه ہ ە, ة ۃ, and a hamza letter or a letter followed by a separate hamza. Text pasted from different sources often mixes them, so a search for one form misses the others.
- Find & Replace → Letters: Exactly as typed; Any form of a letter (ي ی ى · ك ک · ه ہ); or Any form, ignoring vowel signs and hamza (also أ إ آ ا · ة ه, and vowel signs, tatweel and the half space are skipped). Works in the XML source and in the document, for Find, Replace and Replace all.
- Spell check: Persian is in the language list. Words are looked up without their vowel signs and in the language's own letters, so كتاب and کتاب find the same dictionary word.
- Normalise letters (vowel-sign palette حَ): rewrites the letters the language writes another way — Arabic ک → ك and ی → ي inside a word; Persian ي → ی and ك → ک; Urdu also ه → ہ — in the selection or paragraph, or in the whole document. Tick also in pasted text to do it to everything you paste.
Letters of the Language
The palette shows the letters of the document's language that an Arabic keyboard lacks — Persian پ چ ژ گ and the half space, Urdu ٹ ڈ ڑ ں ے, Kurdish ڵ ڕ ۆ ێ ڤ ە, Uyghur ە ۇ ۆ ۈ ې ۋ ڭ — click one to type it at the caret.
Typing Arabic in Latin Letters
In the palette, Type in Latin letters turns a transcription with all its vowels into correctly spelled, fully vowelled Arabic. The preview shows the result; Enter inserts it; untick vowel signs for the letters only.
| Type | For | Example |
b t th j H kh d dh r z s sh S D T Z ` gh f q k l m n h w y | consonants (capitals H S D T Z are ح ص ض ط ظ; ` is ع) | kataba → كَتَبَ |
' | hamza — its seat (أ إ آ ؤ ئ ء) is chosen from the vowels around it | su'aalu → سُؤَالُ, qur'aanu → قُرْآنُ |
aa ii uu | long vowels | kitaabu → كِتَابُ |
| doubled letter | shadda | Hattae → حَتَّى |
-n | tanween | kitaaba-n → كِتَابًا |
ä | tāʾ marbūṭa | madrasätu-n → مَدْرَسَةٌ |
ae at the end | alif maqṣūra | laylae → لَيْلَى |
al-, ash-sh… | the article; sun letters assimilate | ash-shamsu → الشَّمْسُ |
wa- fa- bi- li- ka- | joined to the next word | li l-bayti → لِلْبَيْتِ |
Skeleton Views
Mode tab → Arabic vowel signs → Show: Skeleton without dots draws each letter by its outline only, as early manuscripts did (ب ت ث ن ي inside a word become one tooth ٮ; ف ق become ڡ); Bare skeleton (rasm) also drops hamza and every sign. Only the drawing changes — the text, search and spell check keep the full letters.
Line Breaking
Arabic and Hebrew words are never hyphenated; lines break between words. Uyghur, which writes all its vowels, may break at syllables (Uyghur as the document's language, or lang="ug").
Adding a Language
Languages, scripts and calligraphic styles are modules. A new language is a small file placed in fonts/lang/ and listed in fonts/lang/modules.json — no change to the editor. Pashto is included as an example:
{ "modules": ["pashto.json"] }
A font names its calligraphic style in its main.json ("style": "naskh"), which sets where letters may be stretched, the ligature level and the style's spelling conventions. The full format — including the fields reserved for Indic, Chinese and Japanese scripts — is in docs/LANGUAGE-MODULES.md.
PDF Output Options
Links and Bookmarks
With Clickable cross-references (on by default) cross-references, equation and page references, citations, back-references, footnote and endnote marks, glossary and index page numbers and TOC entries are links; <a href> goes to the web address. The PDF outline lists every heading.
Tagged (Accessible) PDF
Adds a structure tree for screen readers and reflow: headings, paragraphs, captions, lists, tables (with header scope and column spans), formulas (TeX source as alternate text), figures (with alt), code, notes, bibliography entries, TOC and links. Headers, footers and decorations are marked as artifacts; the document language and title are set.
PDF/A-2 Archival
Writes a file for long-term archiving (ISO 19005-2): all fonts embedded (standard fonts replaced by metric-compatible Liberation fonts), sRGB output intent, XMP metadata, invisible text in render mode 3, printable link annotations. Conformance PDF/A-2b, or PDF/A-2a together with Tagged PDF.
The files are not validated with veraPDF inside the app — validate important files before archiving.
LaTeX Import
Scientific Paper → Mode tab → Import LaTeX: choose the .tex file with its .bib and images (PNG or JPG), or paste the source, then Replace document or Append.
- Converted: title block and abstract; parts, chapters and sections (starred too); labels and all reference commands (
\ref \eqref \autoref \cref \pageref …); natbib and biblatex citations and the bibliography style; all equation environments, inline math, tikzcd/CD, \ce and siunitx; figures and subfigures with widths and placement; tables including tabularx, longtable, \multicolumn, booktabs and decimal columns; lists; theorems and proofs; algorithmic/algpseudocode/algorithm2e; verbatim, lstlisting, minted; footnotes and endnotes; glossaries and acronyms; index; lists of figures/tables; macros (\newcommand, \def … expanded); babel/polyglossia language; inline formatting, accents and quotes.
- Images are matched by file name and added to Assets; PDF and EPS images must be converted to PNG or JPG first.
- Replace also replaces the bibliography; Append adds entries.
- A report lists what could not be converted (unknown commands keep their text;
\input files, TikZ pictures and missing images are named).
Saved Project Data
A project file stores, besides the XML and styles:
| Field | Content |
math_settings | Everything in the Math tab |
editor_mode | classic, magazine, scientific, label, carousel or techdocs |
mode_settings | Settings of each mode |
bibliography | Bibliography entries |
pub_settings | Shared settings: baseline grid, hyphenation exceptions, notes, floats, bibliography options and uploaded CSL style, glossary, index, PDF, tables, Arabic vowel signs (tashkeel) and microtypography (microtype) |
Styles also save the hyphenation, justification and grid properties and their tashkeel settings. Older projects load unchanged and get the defaults.
Limitations
- Font expansion and margin kerning are deliberately not used — glyph shapes are never distorted.
- TikZ/PGF drawings (other than
tikz-cd) and picture environments: export them as images.
- LaTeX import:
\input/\include (import each file and append it), \newenvironment definitions (their content is kept), and page-layout packages such as geometry and fancyhdr (set these in the Doc tab).
- Floats inside columns, frames and carousel slides stay in place.
- Boxed blocks beside frames (sidebars, callouts) start below the frame instead of wrapping around it.
- Classic and Arabic: Classic lays out right-to-left and mixed text and shapes Arabic with the font's OpenType rules, so mark placement follows the font; smart placement, hand-moved marks, Qurʾān mode, microtypography, the Arabic keyboard layouts and the Arabic templates are Classic 2 features.
- ل + hamza + ا in Qurʾānic text (لِءَابَآئِهِمْ) is drawn as the لا ligature, as in HarfBuzz.