In traditional architecture, front-end and back-end work together in the same computation process. It is impossible to run each of them separately there. In a headless approach, these are physically separated environments, also isolated within the memory of the server.
Suppose a retailer has many channels for interaction with customers, including social networks, instant messengers, smartwatches, smart mirrors, chatbots, etc., then repeating the publication of content in each of these channels is not effective. In professional language, the channels for customer interaction are called “heads”, and the back-end is called the body. When the front-ends’s work is separated from the body, this approach is called headless. To publish the desired content in headless, the front-end interacts with the back-end via the API.