Semi automated document conversion for the Remarkable2

2021-08-22

Update 2024-10-01: Added comment on more methods to export data from the device. Fix broken link for the Perl script. Also, note that the Sync functionality (They call it Connect now), is not free aymore, but subscription based. Without this, the base functionality of the device is highly reduced or, at least, way less convenient.

My wife recently acquired a Remarkable2 for note taking and document reading while studying.

The device itself is quite nice, it's like an e-book but bigger and you can take notes on it. The hand writing works really well, and the bigger screen makes reading PDF files a way better experience than on a Kindle for instance. It can also convert a handwritten document to text. They even enable SSH access to the device and provide the root password.

If I had to put a "but" to it, it will be that the way of sharing documents or the OCR'd content of a handwritten document is quite limited.One can only send the document or the OCR'd text via email. liron at tilde.team tells me that there are more ways of exporting documents from the device. Apparently there's a USB web interface, and one can also play with the API for the cloud service. Also, the only export format is PDF for the handwritten documents and drawings.

This would probably cover 90% of the use case of most of their customers but this was not our case. Wife's uni works with the Microsoft 365 suite (yeah, that, I'm not going to start rambling here about this ...), so she uses it for all things uni related, even her own notes and study documents. This means that she would have to convert her notes to text on the device, send them via email, copy the output on her email client and paste it into the final MS Word document.

I find all this process a little bit convoluted, so I started to think on making it a little bit easier.

The email you get from Remarkable it's just an HTML email with the OCR'd content.

I took a peek at pandoc and it seems that it can convert from html to docx with a simple command: pandoc -f html -t docx -o $destfile $origfile.

Right, this is a first step on the right direction. One cannot avoid the steps of converting to text and send by email, but getting a MS Word document in your inbox and avoid the copy/paste dance.

As I have my own email server, it's trivial to create a dedicated account for this task, and pass any email that comes from Remarkable to a custom script that does the conversion and then sends the newly generated MS Word document as an attachment to my wife's email account.

And that's what I did. You can find my crappy script here. I'll probably modify it and improve it over time, but for now it works for our use.

Have any comments ? Send an email to the comments address.