Definition
REST (Representational State Transfer) is an architectural style for building web services. Clients interact with resources (users, orders, documents) via standard HTTP verbs: GET, POST, PUT/PATCH, and DELETE.
Responses are usually JSON; URLs identify resources (`/users/42`), and status codes communicate success or failure (200, 201, 404, 500).
In simple terms
A REST API is like a restaurant menu with a standard ordering process. You point to an item (URL), say what you want done (HTTP method), and the kitchen returns your order (response body) or explains if something went wrong (status code).
Where you see it
- Mobile apps fetch data from backend REST APIs.
- Stripe and Twilio expose payment and messaging APIs.
- LLM providers (OpenAI, Anthropic) offer REST endpoints for chat completions.
- Goobo Labs datasets and models may be accessed via documented HTTP APIs.
How it works
1.Define resources
Model your domain as nouns: `/articles`, `/datasets`, `/users`.
2.Map HTTP methods
GET reads, POST creates, PUT/PATCH updates, DELETE removes.
3.Return JSON
Serialize data with consistent field names and error shapes.
4.Authenticate
Protect endpoints with API keys, OAuth, or session tokens as needed.
Why it matters
- REST is the lingua franca of web and mobile integration.
- Understanding HTTP APIs is essential before building AI products that call external services.
Often confused
REST requires a specific framework.
REST is a design convention; you can implement it in any language or framework.
Every API must be RESTful.
GraphQL, gRPC, and WebSockets fit different use cases; REST remains the most common for public HTTP APIs.