Thanks! Lurk looks cool too -- similar, but more in a complementary way I think in that users have to click out of Apricot to read/watch/listen to the content vs. how Lurk brings the content into an embedded browser. Looking forward to seeing it in action!

Apricot's search is a mix -- for a couple platforms, I do call the search endpoint. For the others (including YouTube), it's postgres full-text search. It's a starting point, but it doesn't do fuzzy search, even in this limited lexical task. "mr beast" is a good example -- try searching for "mrbeast" (no space) and you should see it.

I like your idea for finding content, I think it makes sense for what you're doing with Lurk. Are you worried about response time, or is it more of an interactive tool for Lurk users?