Here are benchmarks: https://pglite.dev/benchmarks

The Nostr Postgres schema is inherently faster than the SQLite one because: GIN index on jsonb allows fast tag queries, and upsert syntax improves replaceable event inserts. This is ignoring the fact it runs in a separate process and uses multiple cores.

So even if the baseline wasn't better, I think it would still be better. But in many cases the baseline is better.

Also I don't need to run it in production, just need to simplify my codebase by removing SQLite support. With pglite I can do that without sacrificing anything. Good developer experience all around.