What is GraphQL?

There is one endpoint and a typed schema describing everything available. The client sends a query naming exactly the fields it wants, including nested relations, and gets back that shape and nothing more. Each field is backed by a resolver function on the server, which is what lets a single request assemble data that would otherwise take three round trips.

That flexibility relocates the difficulty to the server. Nested resolvers make it easy to trigger a query per item unless requests are batched deliberately, and because everything is one POST endpoint you lose ordinary HTTP caching and have to add your own. A public GraphQL API also needs query depth and cost limits, since a client can otherwise ask for something extremely expensive by accident.