HTTP

in English: Hypertext Transfer Protocol (HTTP / HTTPS)

basicwidespread

HTTP is the common rule set by which a browser or program requests data from a web server and receives a response.

In plain terms

HTTP is the common language of the web. When you enter a web address, your browser requests the page from a remote machine, the server, in this language, and the server responds in the same language.

It is like ordering in a restaurant: you (the browser) place the request with the waiter, the kitchen (the server) prepares it, and you get back the plate, or the answer that "this is not available today". HTTP records how to order and how to respond, so that any guest and any kitchen can understand each other.

HTTPS is the same, but in a sealed envelope: the request and response travel encrypted, so no one can read or alter them en route. Today, almost every website works this way. The web address starts with "https://", and clicking the icon in the address bar lets you view the connection details. Some browsers indicate this with a padlock, while Chrome has shown a neutral settings icon since 2023. In the case of plain HTTP, browsers usually display a "not secure" warning.

Not only humans browse with it. Mobile applications, e-commerce payment systems, and AI agents also access external services via HTTP requests, for example when an agent queries the weather or books an appointment.

How it works

The client (browser, application, or agent) initiates a connection with the server; in the case of HTTPS, it first verifies the server's certificate and establishes an encrypted channel.
It sends the request, which includes the operation type (e.g., GET = "I request this", POST = "I am sending something"), the precise address, and some accompanying data known as headers.
The server processes the request: it retrieves the page, queries the database, or saves the submitted form.
It returns the response with a three-digit status code (200 = OK, 404 = page not found, 500 = server error) and the content itself: text, image, or data.
Every request is independent; the server does not remember the previous one on its own; therefore, cookies and login identifiers are needed so it recognizes you when you click to the next page.

An everyday example

You click a bad link and are greeted by the "404 – Page not found" message: this is an HTTP status code, the server thus informs you that the requested page does not exist. When your online banking address starts with "https://" and the browser does not warn you, it means you are connecting via HTTPS, the encrypted version of HTTP.

Who is affected

everyoneEvery website visit, online purchase, and app usage involves HTTP requests running in the background, even if you do not see them.

developerWebsites, APIs, and AI-agent tool calls all rely on HTTP, so knowledge of requests, headers, and status codes is essential. → Code

bank customer, investorSensitive data, passwords, or card numbers should only be provided over an HTTPS connection, because over plain HTTP these travel unencrypted. → Money

entrepreneur, executiveA corporate website without HTTPS is marked by browsers as "not secure", which harms trust and search engine ranking.

A common misconception

MythIf there is a lock in the address bar, the site is reliable and secure.

In realityA lock only means the connection is encrypted, not that the site operator is trustworthy. Most phishing sites now also use HTTPS.

Limits and risks

Plain HTTP encrypts nothing: on public Wi-Fi, anyone on the same network can eavesdrop on or modify your traffic.
HTTPS hides the content, but usually not which website you are communicating with. Your provider or the operator of your workplace network can mostly still see this. Newer techniques, such as encrypted name resolution, only help partially.
HTTPS protects only the path, not the destination: if the server stores your data poorly or it is hacked, the encrypted connection does not help.
If an AI agent reads the content of foreign websites via HTTP, hidden malicious instructions may be present, so the received response should not be blindly trusted.

Similar, but not the same

URL (web address) – The URL indicates where something is located on the web, while HTTP defines the rules for requesting and receiving it.
Internet – The internet itself is the network connecting machines; HTTP is just one language spoken on this network. The forwarding of emails between mail servers, for example, uses another language, SMTP.
API – The API describes what can be requested from a service, while HTTP is mostly the channel on which the request and response travel.

Related concepts

Broader

Internet, Network protocol, World Wide Web

Narrower

HTTPS, HTTP status code, HTTP request, HTTP header

On the edge

API, URL, REST, Webszerver, Cookie

A brief history

Tim Berners-Lee developed HTTP at CERN as part of the World Wide Web; the first, simple version appeared in 1991. Encrypted HTTPS was introduced by Netscape in 1994, and the current latest major version, HTTP/3, was standardized in 2022.

Where things stand

At the end of 2026, HTTP is the unavoidable foundation of communication between the web and applications, and the overwhelming majority of traffic runs over encrypted HTTPS. The faster HTTP/2 and HTTP/3 versions are widely used, and AI agents, too, typically reach their tools and services over HTTP.

Status: 2026 November

In the news

Appears in 1 Agent Age stories; the curve shows weekly occurrence in the news magazine.

Coinbase unveils AI financial accounts and limited wallets as x402 spreadsSep 9, 2026 · Money

Also known as

HTTP · HTTPS · Hypertext Transfer Protocol · hypertext transfer protocol · HTTP protokoll · HTTP Secure · HTTP/2 · HTTP/3

Sources

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.