In one of my side projects, I'm currently using Apollo Client, but I'm starting to find writing the code to update the cache after a query or mutation rather cumbersome. I'm thinking about switching to Relay, which, if I understand things correctly, handles a lot of that for you, in exchange to some restrictions it imposes on your schema.

So I'd like to hear from more people which client you're using and why. I'm also interested in any tools you're using, like code generators or utility libraries, to reduce boilerplate code.