You type an address, press Enter, and the page is there before you have time to wonder how. The trip — from your keyboard to a computer that may sit on another continent, and back — usually takes less than a second. It is not magic: it is four ideas working together, an address book, a request, a set of files, and a program that assembles them into the page you see.
Computers do not understand names like poptanilabs.com. They understand numbers: IP addresses. So the first job is translation, and it belongs to the Domain Name System, or DNS — the internet's phone book.
Your browser first checks its cache, where DNS answers are stored for a while. If the answer is not there, it asks a DNS resolver, usually run by your internet provider or a public service, which walks a chain of servers: root servers point to the servers for the ending (.com), which point to the site's own authoritative server, which hands back the IP address.
Sources: Splunk — What Is DNS?, MDN — How the Web works
With the address in hand, your browser sends an HTTP request — Hypertext Transfer Protocol, the language of the web. In plain terms: "GET the main page, please, from this host." HTTP is a client-server protocol: the browser always starts the conversation, and the server replies.
If all is well, the server answers "200 OK" and sends the page's files. A missing page gives the familiar 404; a forbidden one gives a 403. The first file back is the HTML document itself, and as your browser reads it, it spots references to stylesheets, scripts and images and requests each of those too. One page load is really a small bundle of requests.
Sources: MDN — Overview of HTTP, MDN — Client-server overview
Receiving files is only half the job. The browser must now build something you can look at, from three ingredients: HTML for structure (headings, paragraphs, links), CSS for looks (colours, spacing, layout), and JavaScript for behaviour (menus that open, buttons that respond, games that play).
The browser reads the HTML, applies the CSS, runs the JavaScript, and paints the result on your screen. Modern browsers run whole applications this way, not just documents — which is why one can host everything from a calculator to a poker table. Our recent look at AI agent browsers shows that idea being pushed toward browsers that act on your behalf rather than just display pages.
A server can answer in two broad ways. A static site serves files that already exist, exactly as stored: the same HTML goes to every visitor. A dynamic site builds the page on demand — pulling your account details from a database, or assembling a feed just for you. Most of the web mixes the two: pre-made files for the page's shell, fresh data slotted in where it matters.
Static is simpler, faster and harder to break — ideal for any site publishing the same content to everyone. Dynamic is what you need the moment a page must differ from one visitor to the next.
Source: Hostinger — What is a website?
This site is a good specimen: about as simple as a website gets. Poptani Labs is a static HTML site hosted on Cloudflare Pages, a service built for exactly that: it takes a folder of HTML, CSS and JavaScript files and serves them to visitors. Publishing a change means uploading the new files with a deploy script called pages_deploy.py. There is no database, no login, and no page built differently per visitor.
There is exactly one exception. The Birthday Cosmos feature, which shows what the sky looked like on your special day, needs fresh NASA data each time — so it has a single small server-side function, the APOD API, that fetches it on demand. Everything else is a file, served as-is — even the Poker Night game, cards, chips, opponents and all, runs entirely as files in your browser once they arrive.
Source: Cloudflare Pages docs — Static HTML
Here is the anecdote the early web deserves. On July 18, 1992, Silvano de Gennaro, a CERN developer, photographed Les Horribles Cernettes — a CERN parody band that sang doo-wop songs about physics — backstage at the CERN Hardronic Festival. Tim Berners-Lee, then building the web at CERN, asked for a few scanned photos to try out image support on his new system.
The picture that went up was tiny, roughly 120 by 50 pixels, the size of a postage stamp, and took about a minute to load.Scientific images had, strictly speaking, already travelled over networks. What made the Cernettes' photo special is that it was the first picture published on the web purely for fun — the first band photo, posted because a colleague asked, not because an experiment required it. The web's first image was a joke between coworkers, and the internet has barely changed since.
Sources: Wikipedia — Les Horribles Cernettes, The Marginalian — July 18, 1992: The First Photo Uploaded to the Web
A website is not a place. It is files, an address, and a set of agreements — DNS for finding things, HTTP for asking for them — that let any browser fetch and assemble them into a page. That simplicity is the whole trick: anyone who can write a file and put it on a server can publish.
Next time a page appears before you finish blinking, you will know the four stops it made: a name became a number, a request went out, files came back, and your browser built the page. Nothing magic — just a very good postal system.
More in Built in the Lab · Related: Poker Night: the game that died after every flop, AI agent browsers: reality check