Thanks; I think indeed that running it in a "sidecar" is probably the most pragmatic way to do it. The use case I'm considering it for is a greenfield service, so I have some freedom there. The only thing that's on my mind is the slight added complexity for orchestrating infra for this (so far, the small project is starting off as a nimble modular monolith with one service, PostgreSQL being the only dependency). But it's pretty easy to manage I reckon and for the local dev env I could just add SpiceDB in the already existing docker compose config I'm using for PostgreSQL.