Mateo

Worth saying before you build a lot of this: keep the queries simple enough that the note is still readable without the plugin.

The appeal of plain-text notes is that they are plain text. A vault where the important content is generated by queries is a vault that becomes a directory of empty pages the day the plugin stops being maintained, or the day you want to read your notes on a device that does not run it.

My rule: queries are for views — dashboards, indexes, task roundups — and never for content. The actual information lives written out in the note. Then a query breaking is an inconvenience rather than a loss.

It also makes the export question much less frightening, which matters more than it seems until the day you want to leave.

The point about encrypted notes not being searchable deserves more weight, because it undermines the reason most people use these apps.

If you encrypt a note, it is opaque to search, to backlinks, to the graph and to every query. So the encrypted notes become a separate silo that does not participate in your system — which is fine if there are three of them and hopeless if there are three hundred.

That is usually the deciding factor. If the sensitive material is a small, self-contained set, per-note encryption or a separate encrypted vault works well. If it is scattered through notes you use daily, protect the whole machine instead and do not try to encrypt selectively.

Test the sync interaction specifically, as suggested above, and test it in the direction that scares you: create a note on the device that does not have the plugin enabled, and see what the device that does makes of it.

That is the asymmetric case and it is where the reported damage comes from. Five minutes with a throwaway note, and then you know rather than hope.