Can pdfs exported from reMarkable be processed with DT's AI integrations?

[thoresson]

I do a lot of reading on my reMarkable 2. Files are possible to export as pdfs (attached is one example). Are those files possible to process in DT?

There are at least two things I would like to do:

1. Extract all highlighted paragraphs and add them as blockquotes to a markdown file with the same name as the pdf document.
2. OCR all my handwritten annotations and add them to the markdown file as well.

DEVONtechnologies | How to Write a Good AI Prompt.pdf (155.9 KB)

[BLUEFROG]

1. Yes, there is a setting to use blockquotes instead of highlights for annotation summaries to Markdown…
   

   

2. The annotation in your document is not detected as a valid annotation. It’s likely just flattened when exported. You’d have to see if there are any controls for that.

3. OCR is not going to detect handwriting without fairly strict letterforms. External AI may be able to recognize it but that depends on the AI provider and model having vision support.
   In you document, the note was not recognized via text processing but when I specifically set it to vision, it found it. (PS: The annotation it initially reports is one I added in DEVONthink while testing)…

[thoresson]

Yes, sorry, I was a bit too brief: As reMarkable export flattened pdfs, what I really wondered if anyone can think of a way to use an AI model to do some OCR processing to visually find the highlighted paragraphs.

[BLUEFROG]

No worries and as illustrated, a vision-capable model may be able to do it.

And if you have given AI permission to take a screen capture, you could use a prompt telling it to use one, if needed, e.g.,…

What is the handwritten note in this document? Use a screen capture, if needed.