Next release will be at end of July/beginning of August

Monday 19th June 2017

We’re in the process of implementing new machine learning (AI) algorithms for our desktop publisher which we hope will help a lot in creating responsive publications from PDF documents. The results we’re seeing so far are very encouraging. We have been working hard on getting these in place since the start of the year under the hood and had a big breakthrough in late April when we took a new approach on how documents can be republished using unsupervised and supervised algorithms.

As we’re in the midst of this we have decided to delay next release until end of July or beginning of August. If you are waiting for something you have reported to be fixed and would like a pre-release then you are welcome to flick us an email and we’ll help you out.

All the best
The FlowPaper Team

FlowPaper 3.0.2 Release Notes

Sunday 14th May 2017

We’re super happy to announce our latest version which includes a number of improvements in both style and general bug fixes. Here’s the list of noteworthy changes this release :

  • Improved cloud service publishing to make sure it doesn’t show previous versions from cache after modifications have been made (Elements)
  • Improved the look for the Elements viewer including adding options for rounded borders options (Desktop Publisher)
  • Fixed an issue where search wasn’t closing properly on touch devices in Elements (Desktop Publisher)
  • Fixed an issue related to positioning of search matches when matching sub-strings (Classic)
  • Fixed an issue where it wasn’t possible to annotate text after performing a search (Classic)
  • Fixed an issue where some CID characters were incorrectly exported in compressed mode (Zine)
  • Fixed an issue where re-importing a PDF and turning on or off flatten/compression could cause fonts to be missing (Desktop Publisher)
  • Fixed an issue where type3 fonts were not rendered in flattened/compressed mode (Zine)
  • Fixed an issue where rotated text with negative angle would get slightly mispositioned in some cases in flatten/compressed mode (Zine)
  • Multiple style template improvements (Zine)
  • Improved viewer size for Zine when toolbar is hidden or positioned in the bottom of the screen (Zine)
  • Improved the ‘Rainforest’ theme in Zine with new icons and better positioning (Desktop Publisher)
  • Switched to using .bin as mime type when exporting in flattened mode to avoid mime type issues in certain web servers (Zine)
  • The publisher is now defaulting to use flatten/compression (Desktop Publisher)
  • The publisher will now warn if the publication contains fonts that are non-unicode when exporting to ePUB (Desktop Publisher)
  • Fixed an issue where printing ranges were not returning the correct pages (Zine,Classic)
  • Fixed an issue where some texts were classified as overlapping in compressed/flattened mode (Zine, Elements)
  • Improved SEO export settings for consistency between Elements & Zine (Desktop Publisher)
  • Removed the PHP output settings as they were redundant with other settings and not compatible with our
    cloud hosting causing confusion (Desktop Publisher)
  • Fixed an issue where the Cordova application wasn’t able to read its template properly (Desktop Publisher)

FlowPaper 3.0.1 Release Notes

Tuesday 11th April 2017

We’re happy to announce our latest version of FlowPaper with a number of great updates. We have been focusing this month at providing some clarity on how FlowPaper can improve search ranking compared to a normal PDF. FlowPaper Elements has had a number of improvements around search indexing and this version greatly improves the way Google and other search engines is able to discover sections and headers in your publications. You can try the new SEO features in Elements by ticking the “Improved SEO” checkbox under behaviour.

Changes in this release

  • In order to improve SEO, Elements is now using h1/h2/h3 tags instead of div tags for headers (Elements)
  • FlowPaper Elements is now detecting and creating a table of contents for publications if they do not have one specified in the PDF (Elements)
  • Improved font for the TOC in the Elements viewer (Desktop Publisher)
  • Improved email detection and unicode output for compressed/flattened documents (Desktop Publisher)
  • Zine now checks the proportions of the available area when deciding on viewing mode. If the viewer is very narrow but high then the viewer will choose to display a single page as opposed to two pages (Zine)
  • Improved indexing (SEO) for Elements publications so that Google and other search engines are now able to index each section separately (Desktop Publisher)
  • The rel=nofollow and rel=noopener tag added to external links in Elements publications (Desktop Publisher)
  • The elements viewer now updates the #page=(page number) name tag in the url when navigating through the publication to make sharing and returning to a page easier (Elements)
  • The Zine viewer is now able to start with TOC open if any is available (Desktop Publisher)
  • Switched to using the new analytics.js tracker for all viewers (Zine, Elements)
  • Improved pre-loading of fonts in flatten/compressed mode for Zine (Desktop Publisher)
  • Improved loading speed further by including partial content for the Elements viewer using the html5 compliant link rel=”import” tag (Desktop Publisher)
  • Improved unicode support for Chinese &B Thai languages (Classic)
  • Fixed an issue where two-fold Elements publications would have their fonts incorrectly adjusted (Desktop Publisher)
  • Fixed an issue where the Elements viewer was attempting to use the computed font style incorrectly causing the wrong font to be used (Elements)
  • Fixed an issue related to editing pages in the elements viewer where the editor would replace a page incorrectly in some scenarios. (Desktop Publisher)
  • Fixed an issue where setting a background image to Zine would cause an incorrect background color to the search abstracts (Zine)
  • Fixed an issue where setting Zoom Interval to 0.4 or higher would cause the viewer to get the incorrect initial scale (Zine)
  • Fixed an issue where self maximized hosted videos were missing the close button (Zine)
  • Fixed an issue where the search abstract would be mispositioned when going into full screen if already visible prior to entering the full screen mode (Zine, Classic)
  • Fixed an issue where the text for the last page would not load properly for the last page in certain pages when using the flatten/compress mode (Zine)
  • Fixed an issue where small PDFs would get unessicariliy copied to the docs/ directory without being needed when using the flatten/compress mode (Desktop Publisher)
  • Fixed an issue where some auto detected links for email addresses would be incorrect (Zine)
  • Fixed an issue where some animations were not being triggered in a timely fashion (Elements)
  • Fixed an issue where some texts where incorrectly clipped when exported to an Elements publication (Elements)
  • Fixed an issue where two-fold publications with uneven total page numbers were missing the last page (Elements & Zine)
  • Fixed an issue where the viewer would not be initialized properly when starting at certain page numbers (Elements)
  • Fixed an issue where it wasn’t possible to use the full range of page numbers for two fold publications when adding new TOC items (Desktop Publisher)

Improve search ranking for PDF content

Wednesday 29th March 2017

This is the first blog post in a series where we are going to explore how Google ranks PDF documents versus web content created using our upcoming version (version 3.0.1) of FlowPaper Elements. We are going to be completely transparent on how we set up our tests and why we think Google prefers using publications created with FlowPaper Elements as opposed to the PDF so that you can test and verify the results yourself and understand why.


Setting up the test

We decided that we wanted to explore how Google ranked a PDF versus the same content published as HTML5 content. So to do this we staged a little test. We created a blog post with a link to 5 different publications. Each publication converted using FlowPaper Elements and with its corresponding PDF document next to it. You can still see the blog post here. It looks like this:

The following assumptions were made around how Google would treat these links:

  1. Google would treat the PDF and the FlowPaper publication equally on a domain name basis since both are hosted under The FlowPaper Elements publication is actually hosted under but Google treats subdomains the same as subdirectories according to themselves. Please see this Youtube link on this.
  2. Both links were added with absolute positions in the blog post to avoid having Google to rank one better than the other if it appears before the other in the layout
  3. Google would treat both equally on a file name basis since both had the same file name


Results & Analysis

We allowed a bit more than a week to pass before starting to collect results. We then decided to do 3 different tests to see how Google found content within these links. Main title, sub titles and body text searches. Below are the results of our findings. Note that we appended “” (in italic) to restrict searches within our own domain in case the same publication would appear elsewhere.


Main title searches

FlowPaper was able to outrank the PDF in every case that we tried for main title. Here are the titles searches we performed for the publications:

  1. 2016 vacation guide (see screen shot)
  2. 2013 annual report (see screen shot)
  3. Yummy St. Patricks Day Guide (see screen shot)
  4. Art & Design Magazine (see screen shot)
  5. Think Issue 12 (see screen shot)

So how does a PDF define a title compared to a title in FlowPaper? Well, PDFs do not contain meta data around things, so a main title in a PDF is just a larger font that typically appears on the first pages of a document. A main title in FlowPaper Elements on the other hand is an actual header tag (typically a H1 tag) as seen in the screen shot below.

Why is this important? Because according to Google, titles do have relevance to the match of where a certain page lives. FlowPaper Elements make sure that headers are real headers and that they match the title of the publication.


Section title searches

Now thats all fine you might say, because the PDFs may or may not have well defined titles in their text content, so thats a relatively easy thing to beat the PDF on. Well how about titles in sections? Titles in sections should rank high in a PDF too, but there is one major difference in how we treat sub titles and how a PDF treats them. Google claims that having too many titles on the same page would considered crud. Since a PDF contains all titles in the same document, its quite natural to think that a header that appears further down in a document would not get the same search relevance as one on the very top. Let’s have a go and see what happens within our test. We performed the following sub title searches (sub titles marked in bold):

  1. 2016VacationGuideIndexTest Leland (see screen shot)
  2. 2013 Annual Report Grants (see screen shot)
  3. St. Patricks Day Whiskey (see screen shot)
  4. Design Build (see screen shot)
  5. think_issue12 Students (see screen shot)

FlowPaper was able to outrank the PDF documents in each and every case of sub title searches we made. How is it possible? Well, just as with the main title, FlowPaper defines each sub section of a publication with a title using a proper HTML5 tag. It also exports each section into its own HTML page and sets the title of the HTML page to correspond with each section as seen in the screen shot below.


Body Text Searches

So far so good, so how about body texts? A PDF and a FlowPaper publication contains the same body text so these shouldn’t rank differently -right? Well there is one major difference we noted briefly in the previous section. A PDF documents contains all body text in one long page per page structure and FlowPaper splits the document into sections. This means that body text in a sub section would appear higher up in a FlowPaper publication than in a PDF. Lets see what results we’re getting. The body text fragment we searched for is marked in bold.

  1. 2016VacationGuideIndexTest as with most things on the beach (see screen shot)
  2. 2013 annual report EJAF and LoveGold (see screen shot)
  3. St. Patrick’s day guide dancing shoes (see screen shot)
  4. Selby’s research (see screen shot)
  5. think_issue12 Thalassaemia (see screen shot)

FlowPaper was able to outrank the PDF in 4 of 5 cases. Number 2 of the tests did not render any Google result from the FlowPaper publication at all (only from the PDF). Whether this was a random fluctuation or why Google for some reason decided to not index that body text is yet unknown. It could be that it will appear in a few weeks.



We have shown that FlowPaper does indeed have the capacity to improve search ranking for your PDF documents while providing accessibility and speed of loading that by far exceeds anything that a normal PDF document can deliver. In our next part we will look at how to avoid getting Google and other search engines to index and save your content using FlowPaper Elements. Keen to get your fingers on our upcoming version? Contact us via email and we’ll send you a pre-release!

Test post for 3.0.1

Thursday 23rd March 2017

This is just a test post we are using while we are doing some testing for version 3.0.1. In the meantime, check these wonderful publications!

FlowPaper 3.0.0 Release Notes

Monday 6th March 2017

We’re happy to announce the latest release of FlowPaper. In the latest year we made big transitions for the product from being a offline desktop publishing suite to a fully integrated service with cloud publishing.

Moving into version 3 is a milestone for the product in its evolution. The current efforts has reached such a maturity that it deserves its own major product version. First out in this release is our desktop publisher, with web server packages for Classic and Zine to follow later on this week.


The past month has mainly been focused on correcting minor but important issues as well as providing further stability around our new compression/flatten format. Here’s a full list of changes:

  • Added a “speech to text” control button that reads out the current pages loud using text to speech synthesis (Elements)
  • Added conversion profile to the import dialogue for Zine
  • Improved the new flatten and compression mode so that fonts are now exported in separate files as opposed to one big base64 encoded style sheet (Desktop Publisher)
  • Improved mouse wheel zoom speed for Zine (Zine)
  • Improved Elements rendering accuracy where subtags where starting in negative x space resulting in incorrectly squished text (Desktop Publisher)
  • Improved animation performance for Elements in slide mode (Elements)
  • Improved performance and reliability in the flatten/compression mode for Zine
  • Increased timeout for custom fonts when using the flatten mode so that fonts are allowed to load properly (Zine)
  • Added a new ‘TouchZoomInterval’ parameter to the Zine viewer which allows you to let touch devices zoom in using double tap or the touch device tool bar in steps (default is 1.5)
    This also lets mobile users zoom deeper than before using the toolbar (Zine)
  • Improved support for Safari 10.x on desktops. Safari 10 and higher now supports user input in full screen (Zine)
  • Added a “?refresh” flag to exported resources to make sure uploaded publications are refreshed if re-uploaded.
    CDN’s usually cache based on resources query string values and adding this flag forces the CDN to use the latest version (Desktop Publisher).
  • Fixed an issue where one of the fallback scripts for splitting the PDF wasn’t found properly (Desktop Publisher)
  • Fixed an issue related to links not being triggered properly on mobile devices in some scenarios (Elements)
  • Fixed an issue where the initial shadow on smaller devices would be incorrect (Elements)
  • Fixed an issue where adding the download button to a publication which had been flattened would result in a incorrect output font file (Zine)
  • Fixed an issue where swiping left or right on a link would cause the viewer to navigate away instead of changing page (Zine)
  • Fixed a regression issue where jQuery was referenced using $ causing compatibility issues (Zine)
  • Fixed an issue where slightly differently sized pages would cause mispositioned texts when using the flatten mode (Zine)
  • Fixed an issue where some publications were using the original PDF file as opposed to the split file when in two fold mode, causing errors to appear in the console (Desktop Publisher)
  • Fixed an issue where edit mode would show the unadjusted two-fold page if flatten and compression was selected for two fold publications
  • Fixed an issue where some two-fold publications would have an unadjusted last page in Zine (Desktop Publisher)
  • Fixed an issue where some flattened publications would unload some of the pages in the page animation incorrectly (Zine)
  • Fixed an issue where old .pdf files were left in the docs directory when switching between using flatten/compression and standard conversion (Desktop Publisher)
  • Fixed an issue where MSIE (all versions) were delaying font loading in a way that slowed the loading of the document when rendering a flattened PDF (Zine)
  • Fixed an issue where the loader animation wasn’t being displayed on all pages properly (Zine)
  • Fixed an issue where the null character code was incorrectly exported from the desktop publisher in flatten mode (Zine)
  • Fixed an issue where some pages would unload incorrectly in HTML4 mode as well as in flattened / compressed mode (Zine)
  • Fixed an issue where loading from cache would sometimes render javascript errors in browser console preventing text selections to be made (Zine, Classic)
  • Fixed an issue where the print quality would be degraded when using the new flatten/compression mode (Zine)
  • Fixed an issue where the flattened/compressed mode would adjust its vertical position incorrectly for pages where the viewer appears further down on the page (Zine)
  • Fixed an issue where ‘Elements’ publications uploaded with ‘Read Only’ mode turned on to our cloud service would fail to render after viewed on the cloud service (Desktop Publisher)
  • Fixed an issue where it wasn’t possible to switch away from ‘PhoneGap Application’ once it had been selected in the desktop publisher (Desktop Publisher)
  • Fixed an issue where loading documents in split mode would sometimes reposition annotations incorrectly due to an incorrect scale being used in re-rendering (Classic)
  • Fixed an issue where rotation would not work properly in HTML4 mode using the classic viewer (Classic)
  • Fixed an issue where the elements viewer would misposition the content if going into full screen after being zoomed in (Elements)
  • Fixed an issue where clicking another element quickly after hovering a corner in Zine would trigger a page to be turned (Zine)
  • Fixed an issue where using the arrow keys (left/right) in the page # input box would cause the page to change unintentionally (Zine)
  • Fixed an issue where injected scripts would cause the elements viewer to fail to load on some CDNs (Desktop Publisher)

How to make PDF documents accessible to screen readers (NVDA, JAWS)

Wednesday 8th February 2017

In this tutorial, we’ll take a quick look at the accessibility features of FlowPaper. FlowPaper Elements has the ability to convert PDF documents into interactive publications and makes sure that the text is readable by the most popular screen readers such as JAWS and NVDA. It does this by converting the PDF document into HTML5 snippets that are readable by the screen reader software. Many PDF documents are fragmented in their texts and Elements also makes sure that words are not “chopped up” into pieces that make them hard to read for screen readers. For example, the word banana may be represented in a PDF in single characters (“b”-“a”-“n”-“a”-“n”-“a”) as opposed to kept as a word, making it hard to listen to when read out by a screen reader.



Getting started

To get started. Download the desktop publisher from our download page and click the “Import PDF” button to import an existing PDF document. You can also choose to select a number of documents if you want the desktop publisher to join them together.


Selecting project type

The  next step is to select project type for our publication. Make sure you select the second project template (“Elements – Flat”) since this is the viewer we want to use. Only the Elements viewer supports screen readers fully at this stage.



Choosing font conversion

Choose a font conversion method that suits your publication. We recommend using the fonts in the PDF per default but you can also choose to import them from a URL or to match the fonts from our font library to different fonts.




Turn on the accessibility features

Now that we have our document imported, it is time to turn on the accessibility features available. Make sure you have switched on “Improved Accessibiltiy” under “Behaviour” on the left hand side of the publisher as well as “Text to Speech” under controls on the right hand side. Turning on the “text to speech” on the right hand side will give you a speech icon in the tool bar of the viewer that you can click on to hear the current page read out using the built in screen reader of the browser.




Publish and Listen

Now that we have our document prepared – its time to publish and listen to the result. Click the “Publish” button in the top right corner to finalize your publication. Click the “View in Browser” button to preview the publication in your preferred web browser. Click the speech button to test each page read out or use your screen reader to read out each open page in the publication.

Convert PDF to EPUB3 with fixed and reflowable layout

Tuesday 7th February 2017

There are a range of tools available for converting PDF documents to ePUB3 on the market today, but very few actually preserve the layout and fonts of your documents in a way which makes them useful for professional publications. In this tutorial we’ll go through the basics of converting a PDF document using the FlowPaper Desktop Publisher.

To get started, simply download the desktop publisher from our download page and create a “free/trial” account to log in. You should see the desktop publisher grid once this is done.


Importing your PDF

Click on “Import PDF” in the top and select the PDF document that you would like to convert when the import PDF dialogue appears.





Choosing ePUB template

ePUB3 supports both fixed and reflowable layouts. Our next step is to choose a layout model that suits our document. The rule of thumb is that graphical or complex publications should typically use the fixed layout. Books and more textual publications should typically use the reflowable layout format. We will use the fixed layout format for this demonstration.





Choose how to import your fonts

The desktop publisher is able to handle fonts in 3 different ways. It can either import the fonts from your PDF, switch to similar fonts or use fonts from a url. The default choice here is to use the fonts in your PDF document. We’ll select “Use fonts in the PDF” for this demonstration. This option preserves all fonts so that they look just like in your PDF document.





Preview and edit the imported result

You can now preview edit your document before exporting it as a ePUB3 publication. The buttons “Preview” and “Edit” in the top right corner allows you to edit and preview each page. You can also add links, images and videos to the publication.





Prepare your final copy

Once you’re happy with how the publication looks, click the “Publish” button in the top right corner. This will take you to the final publishing dialogue which prepares the final ePUB file for you. Finalize by clicking the “Publish” button.


Voila! your ePUB file is ready to be distributed. Here’s a preview of our test ePUB file in Apple iBooks:

FlowPaper 2.5.3 Release Notes

Wednesday 1st February 2017

We’re pleased to announce our newest version of FlowPaper with a number of what we think are great improvements. The focus during January has been to implement a new compression and flatten mode for Zine which is now available in the desktop publisher. This new mode flattens the PDF and makes it possible to zoom to infinite levels on all devices and platforms.

Here are the most important changes for this release:

  • New import dialogue for Zine in the Desktop Publisher that presents a number of new options including the ability to overwrite previously imported PDFs and
    the ability to flatten PDFs file when importing them. The flatten mode reduces their size and improves speed dramatically, similar to how the Elements viewer already works.
    The flatten import option also improves zoom for all mobile devices to infinite levels and stays sharp at all zoom levels.
  • The Zine import mode in the desktop publisher will now ask if you want to overwrite existing publications if a PDF with the same name has previously been imported (Desktop Publisher)
  • Improved rendering accuracy in flip book mode for Elements to avoid pixel rounding when flipping pages (Desktop Publisher)
  • Improved accuracy for Elements in texts that are starting with negative x-adjustments (Desktop Publisher)
  • Improved accuracy for measuring standard font (Helvetica, Times, etc.) widths in Elements publications (Desktop Publisher)
  • Updated the video player for self hosted videos (Zine, Elements)
  • Fixed an issue related to Safari not fitting some pages properly in Zine for differently sized publications (Desktop Publisher)
  • Fixed an issue where two-fold publications were coming through with the incorrect height specified causing added links to be mispositioned
  • Fixed an issue where two-fold publications were not adjusting their TOC (bookmarks table) properly (Zine, Elements)
  • Fixed an issue where two-fold publications were not rendering the correct number of pages in Elements (Desktop Publisher)
  • Fixed a regression issue related to two-fold publications adjusting their sizes incorrectly causing some images to appear in incorrect places (Zine)
  • Fixed a few issues related to the cloud upload sometimes not providing feedback on progress (Desktop Publisher)
  • Fixed an issue where the publisher would use the wrong standard font (Times, Helvetica etc) in certain cases (Elements, Zine)
  • Fixed an issue where the search abstract would unnecessarily shrink the viewer (Zine)
  • Fixed an issue where ‘Disable Mouse Wheel Zooming’ would still prevent the default event from firing (Zine)
  • Fixed an issue where two-fold publications would not adjust their link targets properly in Elements publications (Elements)
  • Fixed an issue where adding annotations to certain publications would cause the text on their subsequent pages to be missing (Zine)
  • Fixed an issue where obfuscated/read-only publications would become corrupted when applying new import settings in the desktop publisher (Elements)

FlowPaper 2.5.2 Release Notes

Tuesday 27th December 2016

This is a smaller follow-up build to version 2.5.1 that contains a few corrections and fancy new shadows for our Zine users. Please find a compacted version of our changes below:

  • Fixed an issue related to textlayers not being enabled on iOS and Android devices for the annotations viewer (CLASSIC)
  • Improved 3D shadows for Zine (Desktop Publisher, web server packages)
  • Fixed an issue where TOC wasn’t reprocessed properly when re-opening publications (Desktop Publisher)
  • Fixed an issue where custom page labels were not being shown properly in the fisheye navigator (Desktop Publisher)
  • Upgraded three.js to version 79 (Zine)
  • Added additional checks for missing PDF files in the Zine conversion process (Desktop Publisher)
  • Corrected various form validations and reported prcessing issues for the Desktop Publisher
  • Fixed an issue where it wasn’t possible to click apply after rearranging the TOC items (Desktop Publisher)
  • Fixed an issue where user selections were causing the browser to jiggle when panning in zoomed in mode for Zine inside the desktop publisher (Desktop Publisher)
  • More cloud control functionality deployed the the commercial account area