REST vs GraphQL

01

The idea

In REST, you fetch entire resources (a user, a post) and get all their fields. If you need related data (a user's posts, or a post's comments), you fetch a separate resource. This creates N+1 problems and wasteful over-fetching. GraphQL lets you declare exactly which fields you want across related resources in one query. The server returns only what you asked for.

The tradeoff is complexity. REST is simple to implement and cache. GraphQL requires query parsing and validation, and caching is trickier. For mobile clients with limited bandwidth, GraphQL is often better. For simple, server-driven APIs, REST is often simpler.

02

The maths

Let n be the number of related resources you need to fetch.

• REST: Requires n HTTP requests, one per resource. Total bytes transferred includes unused fields from each response.
• GraphQL: Requires 1 HTTP request. Total bytes transferred is only what was requested.

Latency savings are largest when n is large and each resource has many unused fields. If n=1 and all fields are needed, REST and GraphQL are equivalent.

03

Try it

The widget shows a REST response (multiple requests, waterfall) and a GraphQL response (single query). Check the fields you want and see the difference. Larger selections make GraphQL's advantage clearer.

• REST requires multiple round trips; GraphQL needs one.
• Checking more fields makes the over-fetching in REST more obvious.
• Cache headers work differently: REST is header-based; GraphQL needs app-level caching.

04

Where I used it

05

Easy to get wrong

06

Sources

• GraphQL Officialgraphql.org
• RESTful API Designrestfulapi.net

Compared across projects (2022-2026).