It's built off the [Collate data format](https://github.com/Collateapp/collate-file-format) which is just plaintext markdown files with YAML encoded metadata. One of the underlying principals of Collate is that you should be able to traverse your notes in a file browser and edit them in a text editor manually if for whatever reason Collate doesn't work for you. The current iteration of Collate is exactly that, and I'm looking to build on top of that going forward!
I currently keep notes as Markdown files, with similar key/value pairs at the start of each. I then have a hacky script [1] that extracts all such data from each note with a particular tag, and prints the results as a table. So I have automatically generated tables for my notes on talks (with columns for the date, venue, speaker's name, and talk series), books (title, author, publisher), recipes (title, course, source, whether I've cooked it), etc. without needing any special configuration.
If I were to use Collate, I'd keep notes on talks/books/recipes in separate notebooks, so it would be a useful feature to display a table of all key/value pairs in a particular notebook.