FundamentalsUpdated April 24, 2026 · 7 min read

Dark web vs deep web

The deep web is your inbox. The dark web is a tiny anonymity-focused slice on top of Tor. Here's the honest comparison — what each is, what lives there, and why mixing them up matters.

Three terms get used almost interchangeably and shouldn't be: surface web, deep web, and dark web. Confusing them is how you end up with headlines claiming "96% of the internet is the dark web," which is off by several orders of magnitude.

The short version: the surface web is what Google indexes; the deep web is everything else; the dark web is a tiny, specifically anonymous slice of the deep web.

The one-paragraph version

Surface web — public pages a normal search engine can find. Wikipedia, news homepages, product listings.
Deep web — everything that isn't indexed. Your email inbox, your bank dashboard, internal corporate wikis, paywalled archives, draft Google Docs. The vast majority of the internet by volume.
Dark web — a small subset of the deep web that you can only reach with anonymity software like Tor. Addresses end in .onion and don't exist on the public DNS.

Surface web

The surface web is the part of the internet a search engine can discover, crawl and serve to you. It includes everything Google, Bing, DuckDuckGo or any other public engine has in its index. Pages here are:

Reachable from a public link.
Allowed by the site's robots.txt to be crawled.
Not behind a login or paywall.

By volume, the surface web is small. Most of the content you read regularly — news, blogs, Wikipedia, software documentation — lives here, but it's a thin layer over a much larger amount of private content underneath.

Deep web

The deep web is everything a search engine can't reach. That sounds dramatic until you realise it's mostly your normal life on the internet:

Your Gmail inbox, your iMessage archive, your DMs.
Your bank dashboard, your tax filings, your health portal.
Corporate wikis, internal Jira tickets, private GitHub repositories.
Subscription archives — JSTOR, the Financial Times paywall, a Netflix watch history.
Database-backed pages that only exist when you submit a form — flight searches, real-estate filters, library catalogues.

None of this is sinister, and none of it requires special software to reach. You just need permission — usually a login. The deep web is enormous because the surface web is tiny by comparison.

Two things the deep web is not:

It's not the dark web. Reading your inbox isn't "going on the deep web" in any meaningful sense. The phrase is mostly used to inflate cybersecurity sales decks.
It's not anonymous. Your bank knows exactly who you are. The deep web is private, not anonymous — those are different properties.

Dark web

The dark web is a small, deliberately hidden slice of the deep web. To reach it you need software designed for anonymity — most commonly the Tor Browser, but also I2P or Freenet. Sites here have addresses ending in .onion instead of .com, and the addresses themselves are cryptographic public keys rather than human-chosen names.

What lives there is much less exotic than the headlines suggest:

Newsrooms running censorship-resistant mirrors — the BBC, the New York Times, ProPublica, Deutsche Welle.
Privacy services like Proton Mail, Mullvad VPN and Riseup.
Whistleblowing platforms — SecureDrop and GlobaLeaks instances run by major newsrooms.
Search engines for hidden services, such as Ahmia, Torch and DuckDuckGo's onion mirror.
Forums, wikis and software mirrors for projects that prefer their users not to be tracked.

Yes, illegal markets and abusive content exist too — that's the part the news covers. It's a minority of the traffic, and the serious directories (this one included) don't list it.

Quick comparison

Indexed by Google? Surface: yes. Deep: no. Dark: no.
Needs special software? Surface: no. Deep: no. Dark: yes (Tor or similar).
Addresses look like… Surface and deep: example.com. Dark: 56-character-hash.onion.
What it's mostly used for — Surface: public content. Deep: private content (your accounts). Dark: anonymous access (journalism, privacy, censorship circumvention).
Anonymous? Surface: no. Deep: no — your bank knows you. Dark: yes, by design.

Why the distinction matters

When people conflate the deep web and the dark web, two problems follow. First, the deep web sounds scary when it isn't — it's your inbox. Second, the dark web sounds vast when it isn't — it's a small specialised network that does one thing well (anonymous access). Mixing the two makes both harder to talk about accurately.

The honest version is boring: most of the internet is private (deep web), a thin layer is public (surface web), and a much thinner layer is deliberately anonymous (dark web). Each is built for a different purpose, and none of them is mysterious once you know what they're called.

Where to go from here

If you'd like to actually visit the dark web, the step-by-step guide covers setup on desktop and mobile. If you want a deeper look at the network itself, read what Tor is and how it works. And if you're after addresses, the dark.cfd directory only lists services we can verify against the organisations behind them.

Frequently asked questions

Is the deep web the same as the dark web?

No. The deep web is everything search engines don't index — your inbox, your bank dashboard, internal corporate systems. The dark web is a tiny subset that requires anonymity software like Tor to reach. By volume, the deep web is most of the internet; the dark web is a rounding error.

How big is the dark web compared to the deep web?

Estimates vary, but most credible figures put active onion services in the tens of thousands at any given time. The deep web — every private email, every paid database, every internal wiki — is orders of magnitude larger. The popular '96% of the internet is the dark web' claim conflates the two.

Is the dark web real or just a movie thing?

It's real and very mundane. Most of it is newsrooms publishing censorship-resistant mirrors, privacy services, whistleblowing inboxes, forums and search engines. The headlines focus on the small criminal slice because it's newsworthy, not because it's representative.

Do I need to access the deep web through Tor?

No. You're already on the deep web every time you log into Gmail or your bank — that content just isn't indexed by Google. The deep web doesn't need special software. The dark web does.

What makes the dark web different from the regular internet?

Three things: addresses don't exist on the public DNS (they end in .onion and are derived from cryptographic keys), traffic is routed through layers of encryption so the server never learns your IP, and you can verify you're talking to the real operator because the address is, literally, their public key. The regular internet doesn't give you any of those guarantees.