Repo from talk is available at https://github.com/openweb-nl/kafka-graphql-examples (also created pr to update lacinia docs). It also contains a 'testsuite' to specifically stress the GraphQL endpoints, and the results can now be shown inside the webside instead of just links. Idea is to have comparisons accross different GraphQL implemenstations and/or databases. For example using such a setup with Crux, Neo4J, Drgraph or FaunaDB. Some of these have beter GraphQL support, but no subscriptions at the moment. Also tools like Appolo federation don't support subscriptions. Still thinking about next steps. But if you use kafka with schema registry you already have the types. So it should be possible to create an endpoint with just telling how to reach the kafka cluster, and which topics to connect to GraphQL.