If someone were to use Prisma Cloud (or self-hosted) would they need to put another server in front of Prisma if they wanted to execute arbitrary code? For instance if I want a certain graphql mutation to first make a call to a 3rd-party api and then perform prisma data mutation based on that 3rd-party response would I need to orchestrate this on a server that sits in front of Prisma or does Prisma have the ability to run arbitrary node code?

By the way, graphcool and prisma are amazing ideas with really great UX execution and open-sourcing Prisma is also very cool so thanks!

EDIT: I think I know the answer now but just a suggestion- you might want to add a bit more contextual information as to how Prisma fits into a conventional stack. I know you guys are probably crazy busy so I imagine more copy/documentation is to come!