What Is a RESTful API?

A RESTful API (Representational State Transfer Application Programming Interface) is a web service interface built according to the REST architectural style, which defines a set of constraints governing how distributed hypermedia systems should be structured and communicate. Roy Fielding formalized REST in his 2000 doctoral dissertation as a generalization of the architectural principles underlying the World Wide Web. A RESTful API exposes data and functionality as resources, each addressed by a unique Uniform Resource Identifier (URI), and allows clients to interact with those resources using a standardized set of HTTP operations. REST has become the dominant paradigm for designing web APIs because it aligns naturally with the existing infrastructure of the internet and imposes no client-side software dependency beyond a standard HTTP library.

The distinction between REST as an architectural style and HTTP as a transfer protocol is important. REST is not a standard and does not prescribe a single wire format; JSON is the most common representation today, though XML, plain text, and other formats are valid. What makes an API "RESTful" is adherence to the architectural constraints Fielding defined, not the use of any particular format or tooling.

REST Constraints

Fielding defined six constraints that together produce the desirable properties of a RESTful system. The client-server constraint separates user interface concerns from data storage concerns, improving portability and scalability. Statelessness requires that each request from a client contain all the information needed to understand the request; session state is held by the client, not the server, which simplifies server design and improves reliability. The cache constraint allows responses to be labeled as cacheable or non-cacheable, enabling clients and intermediaries to reuse responses and reduce load. A uniform interface, realized through resource identification via URIs and manipulation through standard HTTP verbs, decouples client and server implementations so they can evolve independently. The layered system constraint permits intermediaries such as load balancers, proxies, and gateways to be inserted transparently between client and server. An optional sixth constraint, code on demand, permits servers to deliver executable scripts to clients. Research evaluating these constraints empirically has confirmed that RESTful design rules around HTTP verbs, status codes, and resource naming are the rules practitioners consider most critical for usability and maintainability.

Resource Design and HTTP Methods

The central modeling task in RESTful API design is identifying resources: the nouns that the API exposes. A resource might be a user account, an order, a sensor reading, or a document. Each resource is assigned a stable URI such as /orders/4821, and clients interact with it using the HTTP methods that correspond to standard data operations: GET to retrieve, POST to create, PUT or PATCH to update, and DELETE to remove. The semantics of these methods carry important implications for caching, idempotency, and error handling. GET and PUT are idempotent, meaning that repeating the same request produces the same server state; POST is not. Correct use of HTTP status codes (200, 201, 404, 409, 500, and so on) allows clients to handle responses programmatically without parsing error messages. The IETF has documented RESTful design guidance for constrained IoT environments in draft-irtf-t2trg-rest-iot, extending the principles to low-power embedded systems.

Applications

RESTful APIs have applications in a wide range of systems, including:

• Web and mobile application backends, where REST endpoints expose data to browser and native app clients
• Internet of Things platforms, enabling lightweight communication between sensors, gateways, and cloud services
• Microservices architectures, where independent services expose RESTful interfaces for inter-service communication
• Public data APIs in government, science, and industry, including web services designed for model-driven processes in enterprise integration
• Cloud infrastructure management, where REST APIs control compute, storage, and networking resources programmatically