Utilizing Postgres' RLS feature to isolate user data instead of easy-to-forget where-clauses, is such an underrated use case, I really wonder why not more people use it.
If you prefer code over the blog post, I've put the full application example on GitHub. Would love to hear your thoughts.
It is complicated to initially wrap your head around, I agree. That's why I thought a demo application is the best way to show that it doesn't have to be super complicated once understood.
Like many things in PG, they aren't always easy to get right away I remember my first time with it years and years ago.
I built multi-tenant services and never used RLS because it's complicated.