
What Is the Dark Web, Really?
Few terms in tech get mangled as consistently as “the dark web.” News segments use it as shorthand for anything illicit online, which makes it sound like some vast hidden continent of the internet. In reality it’s a small, specific, and well-understood piece of network infrastructure — and most of what people mean when they say “hidden from Google” isn’t the dark web at all.
Three layers, not one
The surface web is everything indexed by ordinary search engines — the pages Google, Bing, and friends have crawled and cached. Best guesses put this at a small fraction of all web content.
The deep web is just… the rest of it. Your bank’s login-gated statements page, a company’s internal wiki, a paywalled journal article, a Google Doc that isn’t publicly shared. Nothing sinister — it’s simply not crawlable or not meant to be public. Almost everyone uses the deep web daily without a second thought.
The dark web is a much narrower category: sites that are only reachable through an overlay network built for anonymity, most commonly Tor. These addresses don’t resolve through the ordinary DNS system at all — a .onion address is meaningless to a standard browser and only resolves inside the Tor network itself.
Why the network exists
Tor was developed with U.S. Naval Research Laboratory funding in the mid-1990s, with the explicit goal of protecting government communications by mixing them in with ordinary traffic from everyone else. It’s maintained today by the nonprofit Tor Project, and it’s used by:
- Journalists communicating with sources under authoritarian governments
- Whistleblowers submitting documents (SecureDrop, used by most major newsrooms, runs over Tor)
- Ordinary people in countries with heavy internet censorship or surveillance
- Researchers and security professionals
- Yes, also people running or browsing illicit marketplaces — same as any communication tool ever invented
The anonymity Tor provides is the point, and it’s dual-use by nature: the same property that protects a dissident protects a criminal. That’s not a flaw unique to Tor — it applies to postal mail, cash, and encrypted messaging too.
What it actually looks like
Practically, an onion site is just a normal web server — often the exact same nginx-and-static-files setup as any other site — that’s only reachable through a Tor circuit rather than a public IP address. No special dark magic, no different HTML, no different HTTP. The distinguishing feature is entirely at the network layer: how a visitor’s connection gets there, and how little either side can learn about the other.
That’s really the whole story. The next post digs into how the routing that makes this possible actually works.