Disclaimer: I don't have a Rocketbook, nor do I want one.

According to https://rocketbookhelp.zendesk.com/hc/en-us/articles/360019987674-How-to-Choose-Your-File-Format, you can at least export your notes. For everything, it's image formats or PDF, but for text specifically…

https://rocketbookhelp.zendesk.com/hc/en-us/articles/360015750993-OCR-Handwriting-Transcription is actually pretty vague.

If you're sending your file to Evernote, OneNote, or Google Drive, you can send as one file (a JPEG attached with a transcription of your text), or two files (one file with your PDF or JPEG scan and one file with your text)

Alright, what file format is the "one file with [my text]"? And the "JPEG attached with a transcription of [my] text"—if is a regular image, where does my text go? Is it displayed in the JPEG file itself?

If you're sending your scans to an email destination, you can have your transcription attached to your email as a .TXT file or embedded within the body of your email

Okay, pretty straightforward.

No native .md output, but .txt is not too hard to turn into an .md file. At first I was thinking you could literally handwrite a markdown file, manually writing the # before a header or the asterisk around italics. But…

Transcriptions are optimized for letters and numbers, not shapes or diagrams. So, if you’re graphing exponential curves, drawing diagrams for a room, or sketching a friend’s portrait, you’ll see a lot of gibberish. We’re working on ways to clean it up, but for now, we’d suggest transcribing just letters and numbers.

Alright, I hope you handle punctuation well, that's a normal and common part of language, but you just said "letters and numbers".

In its current form, our OCR software isn't able to recognize formats and line breaks as they appear in the planning and calendar pages within the Fusion and Panda Planner, and the graph paper pattern in the Rocketbook Matrix causes some issues of its own. For now, we've disabled transcriptions functionality on those pages types.

Not sure if you would get line breaks in that OCR file as long as you are not using those things, or if you just would not get them at all.

Those would be the barrier between you getting anything to help you with turning .txt to markdown or if you would have to do all the conversion by hand.

I also suggest you go to that OCR link to learn more about its limitations, but these are the most relevant.

::: spoiler I am really not qualified to speak about this topic since I don't even know the product but Imma share something anyway

As far as I can tell rocketbook let's you write things in handwriting.
So you would need to convert the pages into text...

And that would probably mean you would have to use some AI model to accomplish this.

You could write a python script which takes some directory and converts it into markdown.

Try this model over on HF, it might be exactly what you need. :::