What is API Caching?

API caching stores the response to identical or similar requests for a defined period, so the server does not repeat the same work every time. Product lists, exchange rates, city lists, and public content are common candidates because they are read often and change less frequently.

The cache key usually includes the endpoint, query parameters, language, user segment, and authorization context. When another request arrives with the same key, the response can be returned from cache. A new response is generated when the TTL expires, the source data changes, or a manual invalidation runs.

Technical Approaches

HTTP cache: Browser or intermediary caching through Cache-Control, ETag, and Last-Modified headers
Application cache: API code reads a response from Redis or in-memory storage
CDN cache: Public API responses are stored at edge locations
Stale-while-revalidate: A still-acceptable old response is served while a fresh one is prepared in the background

What to Watch

If cache keys are wrong, one user’s data can be shown to another user. Personalized responses, authorization context, and cart data must be separated carefully. With systems such as Redis, TTLs, invalidation, and memory limits need to be designed together.

API caching can improve performance, but the acceptable freshness window should be a business decision, not only a technical default.