Ideas on document annotation ergonomics

I drew this mockup with some ideas to improve documents (PDFs, images ecc) annotations.

The premise is that the user references a document in a block with ![ ]( ) and annotations would be children blocks. The rest is described in the mockup itself:

It would be great to be able to copy and paste text and images from the pdf to Logseq with the least amount of mouse movements and keypresses.

Select text and/or images in pdf and press Enter → Text + screenshot (for images or formulas) is automatically copied and pasted into Logseq. (instead of select text, copy, navigate to right spot in .md, put cursor in right spot, paste, then repeat the same for inline images and formulas)

Logseq can keep track of the page range, with an option to keep snippets in text order.



  1. The text cursor is in a block
  2. Use the mouse cursor on the PDF to select text or a rectangle (according to the last used tool)
  3. Right after the selection is completed just start typing and the text will appear in the last block
1 Like

Yes, something like that. The highlight tool in the Zotero note editor also has a button “add to note” right under the color picker. Zotero also mostly keeps the blocks in order, but it doesn’t work on single pages.

One issue that bothers me is mixed text with math, sometimes copying text with formulas leads to huge blocks of newlines interspersed with symbols. I suspect that this might be due to incorrect OCR layers. For text with lots of math, it means one has to copy a line of text, then copy the formula as a picture, then the next line of text etc. Mostly I just take a screenshot of the entire block, which defeats the purpose of editing the text.
I don’t know if there is a better solution to copy text as text but formulas as images.