How do I use flow to convert a document to a PDF? Can't seem to get this to work?

Nextcloud 19.0.1

Hello Forum,

I’m hoping someone can assist me in how to use PDF Flow to convert a word document (any Only Office document really) into a PDF file. I’m using the PDF Conversion Flow. I’ve created a new configured flow as below:

When a file gets tagged with ‘sop’ the file should be converted to a PDF and Keep the original document.

I have cron setup automatically so it runs every 5 minutes. I assumed that Flow would run every 5 minutes when cron runs. Do I need to enable Flow to run in Cron first?

I read in an issue that Flow does not work on Group Folders or Shared folders (not sure if this is true) so I added a .docx document to one of my Nextcloud document folders. I added a sop tag. I waited more than 5 minutes and confirmed that cron had run in Basic Settings. But my file was not converted to PDF.

How does this work and I what do I need to do to get this setup and working?

Thank you.


Does this mean that no one uses the Flow to convert documents into PDF files?

Thank you.

I tried to use flow with OCR and have no luck either. I do not see where to check what is wrong, nothing in log. This is useless for me.

1 Like

I tried and ran into the same issue. I believe you need to have an instance of LibreOffice running on your server that can do the conversion from document to PDF. I am running Nextcloud on a RaspberryPi and libreoffice isn’t available through an ARM or ARM64 architecture on either snap or docker.

Nextcloud PDF Converter app

This app lets Nextcloud automatically convert documents to PDF. By utilizing the workflow engine it allows Nextcloud administrators to define rules upon which various documents are enqueued to be converted to PDF. Eventually, the conversion happens in a background job by feeding the source file to the found or specific libreoffice or openoffice binary. Depending on the selected behaviour the source file can either be kept or deleted and the resulting PDFs can either be preserved by increasing a number added to the filename or overwritten.

The conversion job is being created when a file was created or updated and also when a system tag was assigned.

Learn more about workflows on


LibreOffice must be installed on the server and the binary must be either detectable by Nextcloud or specified in the config.php as preview_libreoffice_path (cf. the sample config).


This app does not work with either encryption method.

Since LibreOffice is used for conversion, its import filters decide the possibility and quality of conversion. Essentially, Office formats, plain text documents, HTML files but also graphics can be converted. Due to a high number of very custom mime types, by default we feed anything to LibreOffice apart from Audio and Video files.

I use this since years. You really should read the documentation. There it’s clearly stated that you need the LibreOffice binary installed on the server. Without it, it won’t work. Collabora Office or OnlyOffice aren’t used and won’t help with the flow. There’s a ticket since years to utilize them, but it’s not worked on.

In the Original Poster’s defense it took me several hours to figure out that the Automated File to PDF Conversion workflow required an installation of LibreOffice to work. I expected it to work out of the box and when it did not it took quite a long time to realize that LibreOffice was not possible for me to run since I am using Nextcloud on an ARM64 platform.

As a side note it would be beneficial if forum users linked to the documentation that explains the issue they expect the individual troubleshooting to read. Some of us are new to the Nextcloud Platform and have not grown alongside Nextcloud over the years and may not know all of this.

As someone running Nextcloud on a Raspberry Pi 4 I have had to look for documentation on the Nextcloud help forums, the Nextcloud documentation website, the NextcloudPi documentation website, the Nextcloud github pages, the NextcloudPi github pages, as well as scour the web for information on the Nextcloud docker hub page, the NextcloudPi docker hub page, as well as the various individual Nextcloud App github pages.

So when you said “you should read the documentation”, yes, absolutely everyone should read through the documentation of any software or hardware they are using, but it would be helpful to elaborate to users which documentation that might be.


Also in the OP defense, a lot of the flow documentation is misleading.

1 Like

Thanks all for your replies. To close up this post in case anyone else happens by and wants the answer…

I run Nextcloud on Nethserver which is based on CentOS. So as explained previously all I needed to do was to install the libreoffice binaries using this command:

yum install libreoffice

I confirmed libreoffice was installed using this command:

command -v libreoffice

Once installed I setup a Flow task

The team over at Nethserver were awesome at helping me setup Flow:


1 Like