Could the exported structure be customized?

[Ypot]

I use the clipboard capture, and I paste it to my org.

It looks very handy and a great extension!

Just, could a non-advanced user customize the exported structure?
For example, I take the headline of my notes usually with this format (from this link I would write down this headline):

* Urban, T. [2015-03-24 ma.]. The Procrastination Matrix. /Wait But Why./
https://waitbutwhy.com/2015/03/procrastination-matrix.html

Which would correspond, more or less to APA style:
https://apastyle.apa.org/style-grammar-guidelines/references/examples/blog-post-references

[maxnikulin]

I do not plan adding some kind of template engine to the add-on currently. In the past I tried custom formatters in JabRef and can not say that I found it convenient. In addition, it may be tricky since e.g. when Product schema.org type detected then format of title is completely different: brand, product, price, rating.

You may switch to "object" format (on preview page or permanently in preferences) and to parse JSON structure in Emacs. Be prepared that structure is not settled yet and may be changed.

Another option is to process description list inside Emacs into a format that you prefer, maybe even to a bibtex entry.

Till now, I was concentrated on metadata extraction (I am still not happy with its quality) and interface&internal aspects.

I would prefer to have more detailed entry in my notes than it is necessary to generate concise description (I feel that is necessary, but I postponed such task). Frankly speaking, I have no idea what is the best representation for metadata in Org. I am against hiding them inside properties drawer, but some kind of key-value structure is handy. Maybe some fields should be hidden in a regular drawer.

Certainly, user notes and comments should be tightly associated with captured heading. Currently metadata may consume too much screen space to conveniently read a set of related notes.

I do not mind that sometimes full capture is redundant and concise reference for a couple of lines is preferable. Maybe it is time to add menu to browser action (toolbar icon) instead of launching capture immediately.

Ideally, APA (or other) style entries should be produced by the new org-cite feature.

[maxnikulin]

There is a trick that can save some typing, however it requires setup of org-capture and org-protocol with either native messaging application or with desktop-wide handler of org-protocol: scheme.

(custom-set-variables '(org-capture-templates (quote (
     ("b" "LR Brief" entry
      (file "lr.org")
      "* %:description
[[%:link]]" :empty-lines 1)
))))

Example of result:

* Dimi Lee — Firefox 87 trims HTTP Referrers by default to protect user privacy –  — Mozilla Security…
[[https://blog.mozilla.org/security/2021/03/22/firefox-87-trims-http-referrers-by-default-to-protect-user-privacy/]]

LinkRemark tries to keep reasonable length of titles, so components are truncated. Selected text is ignored with such template, it may be copied from preview & debug page (context menu for add-on button when preview is disabled).

Simplified JSON format that should be more convenient for processing in Emacs or custom native messaging app can be discussed. I have no recipe how to transform it to particular citation style. Example:

[{
"url": "https://blog.mozilla.org/security/2021/03/22/firefox-87-trims-http-referrers-by-default-to-protect-user-privacy/",
"title": "Firefox 87 trims HTTP Referrers by default to protect user privacy – Mozilla Security Blog",
"site_name": "Mozilla Security Blog",
"author": ["Dimi Lee"],
"published_time": "2021-03-22T10:00:53+00:00"
}]

However this particular page have high enough quality of metadata. Usually result will be worse.

• Author and publication date often are not available or not marked as metadata.
• Title, description, site name may be arbitrary mapping between these fields (and author as well)
• Canonical link may have http: protocol instead of https: (so the link will allow provider to inject more ads) or may lead to top site page

Due to these reasons at the current development stage I prefer to have redundant data in my notes.