Internet Archaeology: Visit Great Old Websites Online
Internet archaeology made easy: travel back to the web of 1999โ2022, see what Yahoo, GeoCities and early Baidu looked like, and learn to read Wayback Machine snapshots.
Every website you use today has a past, and most of it has been deleted, redesigned or quietly shut down. Internet archaeology is the hobby of digging that past back up: opening a page as it looked in 1999, finding out which sites ruled the web in 2003, and noticing what has vanished since. You don't need special software. This guide shows you where to dig, how to read what you find, and how to start in about ten seconds.
What is internet archaeology?
Unlike a ruined temple, a web page leaves no physical trace when it disappears. The only reason we can still look at old sites is that someone saved copies. The biggest collection is the Wayback Machine, run by the non-profit Internet Archive, which stores timestamped snapshots of pages from many years back.
An internet archaeologist usually asks one of three questions:
- What did this site look like then? (Design, layout, slogans, features.)
- What was popular then? (Rankings, portals, search engines, communities.)
- What happened to it? (Merged, sold, abandoned, or still alive in a different form.)
Method 1: Jump to a random year with the Internet Time Machine
If you don't know where to start, start with luck. The Internet Time Machine covers 1999โ2022 across four language webs: English, Japanese, Simplified Chinese and Traditional Chinese.
- Open the tool and press the big Warp! button (or hit the space bar).
- It picks a random site and year from a weighted pool, then opens that year's archived page in a new window.
- A landing card appears below with the year, the domain, a short description and its rank that year. Use See it then to reopen the snapshot, Visit now to compare with today, or Warp again to jump somewhere else.
You can narrow the pool with the language chips and the era slider, so you can say "only Japanese sites between 2003 and 2008" and let chance do the rest.
If your browser blocks the new window, the landing card shows a notice. Tap See it then to open it manually.
Browse instead of warping
Prefer a guided tour? Switch to browsing mode: pick a language, then a year from the year strip. You get two lists:
- Fun sites of the year: a hand-picked selection (nine are shown at random each visit) with a short blurb per site.
- Top 20 of the year: the ranking for that language and year.
Years without a trustworthy ranking are crossed out in the strip, and the Top 20 block simply doesn't appear for them. Tap any site to see its rank history across years and its list of archived snapshots.
Tip: the keyboard shortcuts are Space for Warp and the left/right arrow keys to change year.
Method 2: Look up any site in the Internet Museum
When you already know which site you want, use Internet Museum (web-archive). Type a domain such as example.com, then choose:
| Button | What it does |
|---|---|
| First Snapshot | Opens the oldest archived version, the site's earliest known look |
| Latest Snapshot | Opens the most recent archived version |
| Popular Sites | One-tap examples if you just want to try it |
The snapshots open inside a retro Windows 98 style browser on win98.si, so the old page is shown in an old-fashioned window. The tool itself is only a shortcut: it stores nothing and relies on the Wayback Machine for the actual archive.
What the web looked like: a few real data points
The Time Machine's ranking data comes from different sources by period, and it is worth knowing which, because that changes what the numbers mean:
- 2003โ2022: Alexa historical rankings (language lists in the early years, country lists later).
- 1999โ2002: earlier period data from other analytics sources, not Alexa. For English 1999โ2002 the entries come from group-level (property) figures mapped to main domains, so they are not directly comparable with later single-domain lists.
With that caveat, some snapshots from the data:
| Year | English-web list starts with | What it tells you |
|---|---|---|
| 1999 | yahoo.com, msn.com, aol.com, microsoft.com, netscape.com, geocities.com | Portals and personal homepages dominate |
| 2003 | yahoo.com, msn.com, google.com, passport.net, ebay.com | Google has arrived, portals still lead |
| 2010 | google.com, facebook.com, yahoo.com, youtube.com, wikipedia.org, myspace.com | Social and video era; MySpace still on the list |
| 2022 | google.com, youtube.com, amazon.com, yahoo.com, facebook.com, zoom.us | Platform consolidation, plus a pandemic-era newcomer |
The other languages tell different stories. The 2003 Japanese list includes yahoo.co.jp, rakuten.co.jp, infoseek.co.jp and geocities.co.jp, and the 2010 Japanese list adds fc2.com, ameblo.jp and mixi.jp. The 2010 Chinese national list has baidu.com, qq.com and taobao.com at the top. Each language web had its own heroes.
How to read an archived page
Old snapshots are rarely perfect. Knowing why helps you avoid false conclusions.
- A snapshot is a photograph, not the live site. The crawler saved the page at one moment. Different parts of a page (images, scripts, stylesheets) may have been captured on different days.
- Missing images and broken layout are normal. If a file was not captured, the browser shows a gap.
- Dynamic features rarely work. Search boxes, logins and anything that needed a server will not respond.
- The date matters. Check the timestamp before quoting anything. Pages changed constantly.
- Not everything was archived. Some sites blocked crawlers or were too deep to crawl, so absence of a snapshot proves nothing.
Be careful with links. Click-throughs from old pages may lead to domains that have since been taken over by other owners. Treat unknown destinations with normal caution.
Ideas for your first dig
- Your first search engine. Find which engine you used as a kid and open its 1999โ2003 snapshot.
- Your hometown web. Set the language chip to your language and browse the Top 20 of the year you were born.
- Rise and fall. Open a site's detail view and read its rank history. Look for the year it peaked and the year it dropped off.
- Then and now. Use See it then, then Visit now, and compare the two windows side by side.
- Fun sites over rankings. Rankings show what was big; the editor's picks show what was strange and delightful.
Frequently asked questions
What is the Wayback Machine?
It is a free web archive run by the Internet Archive, a non-profit. It stores dated snapshots of web pages so you can see how a site looked in the past.
How do I see an old version of a website?
Enter the domain in the Internet Museum and pick First Snapshot or Latest Snapshot. For a random trip, use the Internet Time Machine and press Warp!.
Which years does the Internet Time Machine cover?
1999 to 2022, across English, Japanese, Simplified Chinese and Traditional Chinese. Some language-year combinations have no trustworthy ranking and are shown crossed out in the year strip.
Why does an archived page look broken?
Archived pages are saved piece by piece. Images, scripts or styles can be missing, and server-dependent features such as search or login cannot work.
Does the tool store my data?
The Internet Museum is only a quick shortcut to the Wayback Machine and stores nothing on its side. Snapshots themselves are served by the Internet Archive.
A site from the list won't open. What should I do?
The data is updated and pruned over time, and old domains change hands or disappear. Use the feedback option in the site's menu to report it.
Wrapping up
Internet archaeology costs nothing and takes a minute to start. Press Warp! in the Internet Time Machine for a random trip, or look up a specific site in the Internet Museum. Every snapshot is a small piece of how people once lived online, and the more you open, the stranger and more familiar the past becomes.