HTTP request

basicwidespread

An HTTP request is the message by which a program – browser, application, or AI agent – requests data from a server or sends data to it over the internet.

In plain terms

Imagine you are sitting in a restaurant. You do not go into the kitchen but place your order with the waiter, who then brings out the food. The HTTP request is this order: your computer tells a remote server what it wants, and the server responds.

This happens every time you open a webpage, refresh the news on your phone, or submit a form. Behind the loading of a single page, dozens of such requests often run: separate ones for the text, the images, and the fonts.

In the AI world, this is important because chatbots and agents reach the outside world this way. When an agent checks the weather, books an appointment, or views a website, it sends HTTP requests in the background. When you write to an AI application, your query also travels as an HTTP request to the server running the model.

How it works

Your program (the client) addresses the request: it specifies the URL, which shows which server and what within it the request is for.
It specifies a method, meaning what it wants to do: GET retrieves something, POST sends data, PUT creates or replaces something, DELETE removes.
It appends headers containing metadata: the expected response format, the program sending the request, and often an identification key verifying authorization.
If it sends data (for example, a filled form or a question intended for the AI), that data goes into the body of the request.
The server processes the request and returns a response with a status code: 200 means OK, 404 means the page does not exist, and 500 indicates a server error.

An everyday example

You type the address of an online store into your browser and press Enter: your browser sends an HTTP request to the store's server, which returns the page. If you see the '404 – page not found' message, that is a status code received in response to an HTTP request.

Who is affected

everyoneBehind every website visit, app update, and question asked to AI, HTTP requests run, even if you do not see them.

developerConnecting applications, APIs, and AI agents almost always involves writing, debugging, and securing HTTP requests. → Code

entrepreneur, executiveThe price and capacity of online services often depend on the number and size of requests. For AI services, this typically refers to the amount of text processed (tokens), making it a cost and planning issue.

A common misconception

MythHTTP requests are sent only by the browser when you open a webpage.

In realityMobile applications, smart devices, background programs, and AI agents continuously send HTTP requests, even without human clicks.

Limits and risks

Plain HTTP is unencrypted: only data sent over HTTPS (addresses starting with "https://") cannot be read by others en route.
A request does not guarantee a response: the server may be overloaded, slow, or unreachable, and providers often limit how many requests you can send within a given time.
If an agent independently sends requests on your behalf, these can trigger real operations (purchases, deletions, messaging), so pay attention to what access you grant it.
Identity keys and passwords traveling in requests are valuable: if they leak, others can send requests in your name and at your expense.

Similar, but not the same

API call – An API call means one program uses another program's service; this is most often done via an HTTP request, but the HTTP request is only the delivery method.
HTTP response – A request is the client's message to the server, and the response is the server's reply to it, including a status code and the requested content.
URL – The URL is only the address where the request is sent; the request itself is the full message including the method, headers, and any data.

Related concepts

Broader

HTTP, kliens–szerver modell, internetes protokoll

Narrower

GET request, POST request, HTTP header, status code

On the edge

API, API key, HTTPS, endpoint, webhook

A brief history

Tim Berners-Lee developed HTTP at CERN as part of the World Wide Web between 1989 and 1991. The first version described in an RFC document, HTTP/1.0, appeared in 1996, the widely used HTTP/1.1 in 1997, HTTP/2 in 2015, and HTTP/3 became a standard in 2022.

Where things stand

The HTTP request is a decades-old, ubiquitous basic operation of the web. At the end of 2026, with the spread of AI agents, a growing share of it originates not from human clicks but from machine-to-machine traffic.

Status: 2026 November

In the news

Appears in 64 Agent Age stories; the curve shows weekly occurrence in the news magazine.

Hungarian trade sources: GEO largely builds on SEO, with takeaways from a Vienna conferenceNov 5, 2026 · Marketing
Bug report alleges Claude Code writes user's email address into User-Agent headerNov 5, 2026 · Code
South Korea to build physical AI and robotics hub in Jeonbuk as $15 billion AI budget takes shapeNov 3, 2026 · Ipar
Hungarian court makes new EU referral on press publications and LLM chatbotsNov 3, 2026 · Jog
Copilot Cowork generated 1,211 pull requests in a single dayNov 2, 2026 · Code
Athenahealth launches tool to let patients ask Anthropic's Claude about their health recordsNov 2, 2026 · Health
GitHub adds prompt-injection scanning to AI agent pull requestsOct 28, 2026 · Code
GitHub launches Copilot Fusion, which picks and routes models for each taskOct 26, 2026 · Code

Also known as

HTTP request · HTTP request · HTTP request · web request · request · HTTPS request · GET/POST request · query

Sources

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