What is a GraphQL?

A traditional REST API exposes fixed endpoints, each returning a predetermined response shape — if a client needs data that spans multiple endpoints, it typically has to make several separate requests and stitch the results together, or the server has to add new bespoke endpoints for each new client need. GraphQL flips this: the server exposes a single endpoint and a schema describing all the data types and their relationships, and the client sends a query describing precisely which fields it wants, including nested relationships, and gets back exactly that shape in one round trip — nothing more, nothing less. This solves two classic REST pain points: over-fetching (getting a large object back when you only needed two fields) and under-fetching (needing several round trips to assemble what you actually need). The trade-offs are real too — a flexible query language means the server has to guard against expensive or deeply nested queries that could overload it, caching is less straightforward than with REST's predictable URL-based responses, and the schema itself becomes a contract that needs careful versioning as the API evolves.