The whole point of adding an annotation or a note is to insert the full contents — including your own comments — into your document. If you just want to cite the item, you use Add/Edit Citation.

In an upcoming version, Zotero will offer the current page number in the reader when citing an open document (which you could get to by double-clicking the annotation in the items list in Zotero's library tab). It's sounds like that's what you're looking for.

First, thank you for the new Add Annotation feature in the word processor plugins – being able to insert annotations with active citations and automatic page locators is a huge improvement over manual workflows.

I'd like to request more flexibility in what gets inserted, in two layers (this is how Citavi handles it, and it works very well in practice):

1. A global default for what "Add Annotation" inserts
A preference (e.g. "When adding annotations, insert: [Full annotation | Citation only]") that controls the default behavior. Many of us primarily want the pinpoint reference – the citation with the page number – and only occasionally the quoted text.

2. A per-insertion choice in the annotation picker
When selecting annotations in the picker, a checkbox like "Insert citation only" (or a modifier-click), so the choice can be made per insertion and per annotation, overriding the default. This mirrors Citavi's "Insert with options" behavior, where the default is just the reference and you opt into the quotation when you actually need it.

Why this matters for my use case (and I suspect many others'):

In legal and much of humanities writing, the standard citation is a source + page reference in a footnote; the verbatim quote goes into the running text only when the exact wording matters. My highlights in the PDF are markers of relevance while reading – weeks or months later, when I draft the brief or paper, I paraphrase and need the footnote reference, not the pasted quotation.

Currently the only way to get just the reference is Add/Edit Citation, which loses the annotation's automatic page number – I have to open the PDF and find the page manually. The planned improvement (offering the current page of an open document) helps only if the right document happens to be open at that moment; it doesn't help when citing from the annotation picker later on.

Pasting the full annotation and then deleting the body text each time is error-prone: a forgotten deletion accidentally turns a paraphrase into an apparent quotation – a real problem in disciplines where the distinction between direct and indirect citation is critical.

Implementation-wise, this seems very tractable: the picker already knows each annotation's page label, and "citation only" could simply skip the body/comment HTML and insert the citation field with the locator – essentially the same output Add Citation produces, but with the page number sourced from the annotation instead of typed by hand.

---
Related gap: annotations for sources without a full-text file

The same workflow breaks down entirely for print sources. As discussed in the "Annotations for physical items?" thread, annotations can only exist on file attachments (PDF, EPUB, HTML snapshot) – for a print book or a source consulted in a library, there is no way to create an annotation with a page number at all. The suggested alternative, regular notes, lacks exactly the properties that make annotations valuable for citation: a structured, editable page label, colors, types, and above all integration with the word processor plugins.

Two things would help here:

Allow annotations on regular (parent) items with a manually entered page label – a "print annotation" that behaves like a PDF annotation (appears in the annotations tab, is selectable in the Add Annotation picker, carries a locator) but references a physical page of a print source.

At minimum, allow manual page labels where they're currently missing – e.g. EPUB annotations only get page numbers when the EPUB ships embedded pagination, and there's no way to correct or add one manually, which is a problem for e-books with non-matching print pagination.

Together, these two FRs describe one coherent workflow: capture a pinpoint (with page) while reading – whether the source is digital or print – and later cite just the reference, with the quote only where it's actually needed. This is essentially the Citavi model, and it's the main reason many legal scholars in the German-speaking world have stuck with Citavi so far. Making this possible natively would remove a significant adoption barrier.