User guide
RTL Export for Confluence exports pages written in Arabic, Hebrew, Persian or Urdu as a print-ready page, a Word document or a PDF, with every paragraph in the right direction and every letter in its proper shape. This guide covers exporting, the options, what is exported and the known limitations. Questions not answered here are in the FAQ.
Export a page
- Open the page or blog post.
- Select ••• (More actions) → Apps → Export RTL.
- Adjust the options for this export if needed. They start from your site's defaults.
- Select Print / Save as PDF. In the print dialog, choose "Save as PDF" or a printer, and turn off Headers and footers: the export adds its own title and page numbers.
Page numbers and the running title print in Chrome and Edge. Safari and Firefox print the document without them; Download PDF (below) has them in every browser.
Arabic, Persian and Urdu text in the saved PDF can be searched and copied. To copy with the words in reading order, open the PDF in Preview or Firefox: Chrome's built-in PDF viewer copies the words of each line in reverse order.
Export to Word
In the export dialog, select Download Word. The document is created in your browser with the options you chose, and saved to your downloads folder under the page's title. It opens in Microsoft Word and LibreOffice with right-to-left paragraphs, tables with their first column on the right, numbered lists, headers and page numbers. Wait until the dialog has finished loading images: images that have not arrived appear as labelled boxes.
The Arabic-script font of the export (Amiri or Vazirmatn) is embedded in the file, so the document looks the same on computers that do not have it.
Download PDF
In the export dialog, select Download PDF. The PDF is made in your browser by the export's own layout engine, with the options you chose, and saved to your downloads folder under the page's title. Unlike a PDF saved from the print dialog, it has the running title and page numbers in every browser, bookmarks for the headings, page numbers in the table of contents, and links you can click. Its fonts are embedded. As with Word, wait until the dialog has finished loading images.
Text in the PDF can be searched and copied. Firefox and Preview copy right-to-left lines with their words in reading order, and so does Chrome's built-in viewer from Chrome 156; see the known limitations for the exceptions.
Download Word and Download PDF need an active licence for the app; the print view works without one.
Export a page with its child pages
When the page has pages under it, the export dialog shows Include child pages with their number. Select it to export the page and the pages under it as one document:
- Each child page starts on a new page, with its title as a heading. A child page's title is one heading level below its parent page's, and the headings inside it move down to match, so the document keeps the outline of the page tree.
- Each page keeps its own direction: an English page in an Arabic page tree stays left to right.
- Links from one exported page to another jump to that page inside the document.
- Pages are read as you: pages you cannot view are not included. A page that cannot be read keeps its title, with a note in its place.
The dialog shows its progress while it reads the pages; select Cancel to stop. An export takes up to 100 pages, in the order of the page tree.
Cover page and table of contents
Cover page opens the document with a page that shows the title, space, owner, version, date of the last update and date of the export. Page numbers count the cover but do not appear on it.
Table of contents adds the headings, to three levels, after the title (or the cover). It is left out for a page that already has a table of contents macro; in a page tree, the macro lists only its own page's headings, so the export adds its table of contents too. In Word, the table of contents is a field: right-click it and choose Update Field to add page numbers.
Options
| Option | Choices | Effect |
|---|---|---|
| Direction | Detect per paragraph (default), Right to left, Left to right | Detection reads each paragraph's words, so an Arabic sentence that starts with an English name stays right-to-left. The forced settings apply to every paragraph except code. |
| Numerals | Western (123), Arabic-Indic (١٢٣), Persian (۱۲۳) | Applies to numbers in Arabic, Persian and Urdu text, to list numbers and to page numbers. Persian and Urdu pages show Arabic-Indic as Persian digits (۱۲۳). Code, URLs, Hebrew text and numbers after English words keep Western digits. |
| Arabic-script font | Automatic, Naskh (Amiri), Sans (Vazirmatn) | Automatic uses Amiri for Arabic and Urdu, and Vazirmatn for Persian. Text set in either font can be copied and searched in the saved PDF. |
| Paper | A4, A5, US Letter, US Legal | Page size when printing, and of the Word and PDF documents. |
| Calendar | Gregorian, Hijri (Umm al-Qura), Solar Hijri | Calendar for dates in the page. |
| Cover page | On, off (default) | Opens the document with a cover page (see above). |
| Table of contents | On, off (default) | Adds the headings to three levels after the title (see above). |
Administrators set the defaults in Settings → RTL Export.
Supported content
| Content | In the export |
|---|---|
| Paragraphs, headings (with alignment and indentation) | Direction per paragraph; right-aligned when right-to-left |
| Bullet, numbered, task and decision lists | Markers on the right in right-to-left lists; English items align with their marker |
| Tables | First column on the right in right-to-left documents; header row repeats on each printed page; each cell detects its own direction |
| Page layouts | First column on the right in right-to-left documents; a layout whose first column is in English, such as English beside Arabic, keeps the order it has in Confluence |
| Code blocks, inline code | Always left-to-right |
| Panels, quotes, expands | Icon and border on the right in right-to-left text; expands printed open |
| Images | With captions and sizes; images from other pages load too |
| Links | Printed with a numbered list of link addresses at the end; in Word and PDF, links you can click |
| Status, dates, mentions, emoji | As text; dates in the chosen calendar |
| Table of contents macro | Generated from the page headings; in a page tree, from its own page's headings |
| Excerpt, page properties and legacy panel macros | Their content is exported |
Known limitations
- Macros whose output Confluence does not store in the page (for example Jira issue lists, charts, diagrams from other apps) appear as labelled boxes and are listed at the top of the export dialog.
- External images (from other websites) appear as labelled boxes.
- Copying text from a PDF: PDFs that Chrome saves from the print view are not marked right-to-left, so Chrome's built-in viewer can copy their lines with the words in reverse order; Firefox and Preview copy them mostly in reading order. PDFs from Download PDF copy mostly in reading order in Firefox, Preview and Chrome 156 or later, with two known exceptions: Chrome can put English words and numbers inside a right-to-left line out of order ("Jira Software" as "Software Jira"), and Firefox and Preview can copy table rows out of order. Chrome 153 to 155 copies right-to-left lines out of order from these PDFs, as it does from Word's and Chrome's own. In any PDF, letters that carry diacritics (tashkeel) can copy with the mark out of place. Searching for words works.
- Very large pages (over about 4 MB of content) cannot be exported yet.
- In Word documents, page numbers always use Western digits (1, 2, 3), whatever the Numerals option: Word has no setting for Arabic-Indic page numbers, and shows digits as Word's own options say. List numbers and text follow the option. Lists that use Arabic-Indic or Persian digits, or Persian and Urdu letters, are written as text, so Word does not renumber them when you edit.
- Images that wrap text in Confluence sit on their own line in Word documents, and panels have no icon.
- An export takes up to 100 child pages. Pages more than 10 levels below the exported page are not included.
- The table of contents in the print view has no page numbers (browsers cannot print them); in Word, update the field to add them. Word for Mac lays out an updated table of contents left to right.
- Links to the exported page itself, from its child pages, open the page in Confluence rather than jumping to the top of the document.
- Urdu prints in Naskh (Amiri). The open Nastaliq fonts draw letters from several glyphs, so text copied from a browser-saved PDF would lose letters; a Nastaliq option is planned for the PDF export.