In its broadest definition, the deep web is anything on the Internet that you can’t find on Google—however, this doesn’t mean that your email inbox is on the same level as hidden online markets. The biggest problem in shining a public spotlight on the deep web is the lack of common language describing it, and confusion over the border where the normal internet “ends” and the deep web “begins”.

Most search engines track web pages by following links, which means anything requiring a password or special search term to access won’t be displayed; by this definition well over 90 percent of pages on the Internet are part of the deep web. However, these databases, message boards and archives don’t form most people’s ideas of the “underground” internet.

The parts of the deep web most commonly associated with anonymous and illegal activities are special networks designed to keep users anonymous. The most popular is the free-to-download Tor, which itself is only one method of hiding everyday internet use from any tracking attempts.

Tor encodes any outgoing messages with multiple layers, and then sends it through a randomly selected network of volunteers’ computers. Each stop through a volunteer unencrypts one layer and reveals the next destination. The only trackable link in this chain is between the last volunteer computer (the “exit node”) and the message’s destination, and it is nearly impossible to follow this link back to the original computer.

Anonymity networks themselves are legal, because there is no law preventing encrypting internet use. However, the networks enable hidden services: revealed classified documents, counterfeiting services for money and papers, and illegal markets for drugs or weapons. These sites are accessed via randomly generated, 16-character names ending in .onion (for Tor’s “onion routing” encryption), and since they are rarely linked between each other, they are found only on dedicated non-deep web pages on websites like Pastebin, Reddit, and Twitter.