Endpoint

Simple Definition

An endpoint is a specific web address (URL) that an API exposes so that other software can send requests to it. Think of it as a door into a service — each door leads to a different function.

In AI: when you use an AI API, you send your request to a specific endpoint depending on what you want — generating text, creating an image, or turning speech into text each have their own endpoint.

A Simple Analogy

Think of a hotel with different phone extensions. Extension 1 goes to room service, extension 2 to housekeeping, extension 3 to the front desk. Each extension is an endpoint — a specific address for a specific function. You call the right one depending on what you need.

Common AI Endpoints

Most AI providers structure their endpoints by task:

  • Chat/completions — for generating text responses (e.g., POST /v1/chat/completions)
  • Images — for generating images from text (e.g., POST /v1/images/generations)
  • Audio — for transcribing or generating speech
  • Embeddings — for converting text into vector form for search

Why Endpoints Matter

Understanding endpoints helps you:

  • Know which part of an API to call for a specific task
  • Debug errors (hitting the wrong endpoint is a common mistake)
  • Understand API pricing (different endpoints may have different costs)
  • Read API documentation more easily

Endpoint vs. API

An API is the whole system — like a restaurant. An endpoint is one specific item on the menu. A single API can have many endpoints, each doing something different.

  • API — the broader system that endpoints belong to
  • Function Calling — AI using endpoints to call external tools
  • AI Integration — connecting AI endpoints to your own apps
  • Inference — what happens when the AI processes a request at an endpoint

Continue learning

Explore related guides, tools, workflows, and prompts that help you go deeper into this topic.

See AI terms in action

Browse practical AI workflows that use the concepts in this glossary.

Last updated: