Server-side read models

It would be cool if we could take the users code and run it 'serverside' (meaning, in the nodes or in sidecars)

• the read models could be in memory, or persisted in one of a small selection of built in stores (stream store, kv store, relational store?) or perhaps persisted in an external store.

• it could result in an API defined by the read model for querying. graphql, blazor?, ...