We rebuilt a Yahoo-style web directory in 2026. Not ironically, not as a joke, and not as a nostalgia project that would live for a weekend and die. We built it properly — categories, descriptions, a submit form, a random button — and put it at wl.pt. Then we watched what happened.
What happened was that people used it. Not millions of people. Not a viral moment. But thousands of clicks on the top links, steady browsing, actual submissions from strangers who found a site they thought deserved to be listed. Real usage from real humans navigating a curated list of links, sorted into categories, with no algorithm deciding what they should see.
That surprised us. It probably shouldn't have.
What the web used to be
In 1994, Roger was a journalist in Lisbon. His editor at Nova Gente magazine gave him an assignment: write about this new thing called the Internet. To write about it, he had to use it. To use it, he needed to find things on it.
There was no Google. Google wouldn't exist for another four years. There was no search engine at all, really — not yet. The way you found things was through directories and link pages: hand-curated lists of websites, organised by someone who had actually visited them. In 1995, Yahoo, AltaVista, and Excite would change the way people found things online, but in 1994, when Roger was researching his article, none of that existed yet. If you found something good, you saved it — because if you didn't, it was gone. There was no way to find it again.
By 1995, Yahoo had become the most famous directory on the web. Not a search engine — a directory. A hand-organised list of websites, sorted into categories by actual humans who visited each site and decided where it belonged.
You didn't search the web in 1994. You browsed it. You opened a category — Arts, Science, Computers, Entertainment — and every category was a small door into a new world. You scrolled through a list of links that someone had chosen to include. Each link had a short description. You clicked one, read it, went back, clicked another. It was slow, deliberate, and surprisingly satisfying. You discovered the internet at a human pace.
The web felt finite. You could reach the bottom of a category. You could feel like you'd seen what was there. That feeling — of completeness, of having explored a space — is something the modern web has completely destroyed.
What the web became
Today, the web is infinite and exhausting. Every platform is a feed. Every feed is algorithmic. Every algorithm is optimised for engagement, which means it's optimised for outrage, anxiety, and addiction. You never reach the bottom because there is no bottom. You never finish because you're not meant to finish. You're meant to keep scrolling.
Search engines don't help as much as they used to. Google in 2026 is a wall of ads, sponsored results, AI-generated summaries, and SEO-optimised content farms. The actual useful result is buried somewhere on page one, sandwiched between an ad for something you searched for last week and a "People also ask" box that answers questions nobody asked.
Social media killed the homepage. RSS died a slow death. Bookmarks became pointless when browsers synced everything and nothing. The web didn't get worse, exactly — it got bigger, faster, and louder, and in the process it lost the thing that made it feel like a place worth exploring.
It lost curation.
Why we built World Links
We didn't set out to make a statement. We set out to scratch an itch.
We travel full-time in a motorhome across Europe. We're constantly discovering useful websites — local services, government portals, cultural institutions, camping resources, news sources in languages we speak. And every time we found a good one, we'd bookmark it, lose the bookmark, search for it again three months later, and fail to find it because the search results had changed.
What we wanted was simple: a page where good links lived, organised by category, with a short description of what each site was. Something we could browse when we wanted to find something useful. Something we could share with other people who might want the same thing.
So we built it — an homage to 1990s Yahoo!, back when browsing meant exploring and being surprised on every click. Vanilla PHP, a MariaDB database, clean HTML, no JavaScript frameworks. Categories inspired by the old Yahoo directory: Arts and Humanities, Science, Computers and Internet, Education, Entertainment, Business and Economy, News and Media, Recreation and Sport, Health, Society and Culture, Travel and Tourism. Each link hand-reviewed, hand-categorised, and described in a sentence or two.
We added a few touches that the old Yahoo didn't have: a random site button that sends you somewhere unexpected, a submit form so anyone can suggest a link, click tracking so we can see what people actually find useful, and both Portuguese and English versions because that's who we are.
The whole thing took less than a week to build. It lives at wl.pt, it runs on our own servers, and it costs us nothing beyond the domain registration.
What we learned
The first thing we learned is that people actually want this. Not everyone, not most people, but a meaningful number of humans who land on a curated directory will browse it. They'll click the categories. They'll follow the links. They'll hit the random button. They'll come back.
The top links on World Links have accumulated thousands of clicks. SAPO, the Portuguese portal, leads the board. Público and Jornal de Notícias — two of Portugal's most important newspapers — are close behind. The Portal das Finanças, the Portuguese tax authority, is in the top five. GitHub is up there too.
These aren't obscure discoveries. People know these sites exist. They could type the URL directly. They could Google them. But they clicked through a directory instead. Why?
We think it's because the directory provides something the modern web has forgotten: context. When you see a link in a category, next to other links in the same category, with a human-written description, you understand what it is and why it matters. You're not being sold to. You're not being tracked. You're just being shown something that someone thought was worth your time.
That's curation. That's what Yahoo did in 1995. And it still works in 2026.
The random button
Our favourite feature is the simplest one: the random button. Click it and you go to a site you didn't choose, didn't search for, didn't get recommended by an algorithm. You just... go somewhere.
This is the purest expression of what the early web felt like. You followed a link because it was there, not because a machine decided you should see it. Sometimes it was boring. Sometimes it was fascinating. The point wasn't efficiency — it was discovery.
The random button gets more clicks than we expected. People like being surprised. People like not knowing what comes next. In a web that's optimised to show you exactly what you want before you even know you want it, randomness is radical.
Curation is not scalable, and that's the point
The obvious criticism of a web directory is that it doesn't scale. Yahoo's directory collapsed under its own weight as the web grew from thousands of sites to millions to billions. No team of humans can categorise billions of web pages. That's why search engines won.
But we don't have billions of links. We have 83. And that's fine.
World Links isn't trying to index the internet. It's trying to be a good starting point — a place where you can find a handful of genuinely useful, carefully chosen sites in a dozen categories. It's a reading list for the web, not a replacement for Google.
The fact that it's small is a feature, not a limitation. When you open a category and see five links, you don't feel overwhelmed. You feel invited. You can actually look at all five. You can make a choice based on what sounds interesting, not based on what an algorithm ranked highest.
Curation doesn't scale. That's exactly why it's valuable. The things that scale — feeds, algorithms, AI-generated content — are drowning us. The things that don't scale — a human choosing a link and writing a sentence about it — are the things we actually trust.
The old web wasn't better, but it was different
We're not nostalgic for dial-up modems and 56k connections. We're not pretending the 1990s web was some golden age of information. It was slow, ugly, unreliable, and most of the content was terrible.
But it had a quality that the modern web has lost: it was browseable. You could wander. You could stumble onto something you weren't looking for. You could reach the end of a page and feel like you'd finished. The web was a library, not a slot machine.
World Links is our small attempt to bring back that feeling. Not the technology — the technology of 2026 is vastly better. But the experience. The feeling of opening a door, looking around, finding something good, and closing the door satisfied.
We don't think the directory will replace Google. We don't think it needs to. We think 83 well-chosen links, organised by humans, with a random button and no tracking, is a small act of resistance against a web that has forgotten what browsing means.
And based on the thousands of clicks, we're not the only ones who feel that way.
Build your own
If this resonates with you, here's the honest advice: build your own. Not a clone of World Links — your own version, for your own community, with your own taste.
A web directory is one of the simplest things you can build. A database of links, a few categories, a page that lists them. No frameworks, no APIs, no complex infrastructure. A weekend project that might last for years.
The web needs more curated spaces. More small, intentional collections of links chosen by humans who care. More random buttons. More finite pages you can actually finish reading.
The 1990s directory didn't teach us anything new. It reminded us of something old: that the best links on the internet are the ones someone chose for you, not the ones a machine decided you should see.