How-to · Published October 17, 2026 · 9 min read

How to See What a Website Looked Like in the Past (Wayback + Automated Archiving)

Quick answer

To see a website's history and what it looked like in the past, open web.archive.org, paste the address and pick a date from the calendar. Blue dots mark captures where the page loaded normally; green dots are redirects. If the Wayback Machine has nothing, try archive.today or a national web archive. For your own sites, keep a dated screenshot of the website yourself so its history never depends on a third party.

1996the year the Internet Archive started collecting the web

2024Google removed cached pages from search

1 URLpattern opens any archived date directly

0guarantees that a third party archived your page

The Wayback Machine, step by step

Go to web.archive.org and paste the full address, for example example.com/pricing.
Pick a year in the timeline at the top. The bar height shows how many captures exist.
Pick a day in the calendar. Hover over a dot to see the exact capture times. Blue dots are captures that returned a normal page (2xx), green ones are redirects (3xx), orange are client errors (4xx) and red are server errors (5xx), so start with blue.
Browse inside the archive. Links on the archived page open other archived pages from about the same time.
Compare dates with the "Changes" view to see what was added or removed between two captures.

URL tricks that save clicks

You want	Type this
All captures of a page	`web.archive.org/web/*/example.com/pricing`
The capture closest to a date	`web.archive.org/web/20150101/https://example.com`
Every archived URL under a site	`web.archive.org/web//example.com/`
The raw archived file without the toolbar	add `id_` after the timestamp: `web.archive.org/web/20150101000457id_/http://www.example.com/`

The timestamp format is YYYYMMDDhhmmss. You can shorten it (2015 or 20150101) and the Wayback Machine redirects to the nearest capture it has.

Save a page today, so it exists tomorrow

At web.archive.org/save you can ask the Wayback Machine to capture a page right now (Save Page Now). It is useful before a page is likely to change: terms of service, a price, a public statement. The capture gets its own permanent archive URL that you can cite. Logged-in users get extra options, such as capturing outlinks.

Other archives to try

Archive	Good for
archive.today (archive.ph)	On-demand snapshots, including a screenshot of the page as rendered
National web archives (for example the UK Web Archive, the Library of Congress web archives)	Government, news and culturally important sites in that country
The site's own changelog, press page or GitHub history	Official record of what changed and when
Search Console (for your own site)	How Google last crawled and rendered a page

What happened to Google cache?

For years, "Cached" next to a search result was the quickest way to see a recent copy of a page. Google removed that link in early 2024 and the cache: search operator stopped working later in 2024. Google now links to the Wayback Machine from the "About this result" panel. Older tutorials that still recommend Google cache are out of date.

Query the archive with the CDX API

When you need the history of many pages, the Wayback Machine's CDX server returns a list of captures as JSON. This request returned real results when we ran it:

request

https://web.archive.org/cdx/search/cdx?url=example.com&output=json&limit=3&from=2015

response (shortened)

[["urlkey","timestamp","original","mimetype","statuscode","digest","length"], ["com,example)/","20150101000457","http://www.example.com/","warc/revisit","-","B2LTWWPU...","525"], ...]

python

import requests def captures(url, year_from="2015", limit=20): r = requests.get( "https://web.archive.org/cdx/search/cdx", params={"url": url, "output": "json", "from": year_from, "limit": limit, "filter": "statuscode:200", "collapse": "digest"}, timeout=30, ) rows = r.json() header, data = rows[0], rows[1:] for row in data: item = dict(zip(header, row)) yield f"https://web.archive.org/web/{item['timestamp']}/{item['original']}" for link in captures("example.com/pricing"): print(link)

collapse=digest skips captures whose content did not change, and filter=statuscode:200 keeps only successful ones. The archive rate-limits heavy use (we got HTTP 429 from its availability endpoint during testing), so add pauses and cache results.

Why the snapshot you need is missing or broken

Symptom	Reason
No captures at all	The page was never crawled, was behind a login, or the site asked to be excluded
Page loads without styles or images	Stylesheets and images are archived separately and were not captured at that time
App shows a spinner forever	Single-page apps load data from APIs that the archive did not store
Big gaps between dates	Crawling frequency depends on how popular and linked the page is

A screenshot avoids the missing-file problem because it stores the rendered page as one image. It can still record a spinner, empty image slots or a half-loaded page if it is taken too early, so wait for the content you care about before capturing and check the result.

Archive your own pages automatically

If the history matters (pricing, terms, landing pages), keep your own archive. A daily job that saves a dated full-page screenshot covers many marketing and product needs. For regulated content, check your record-keeping obligations first: they may require specific formats, retention periods or capture frequencies that a daily image does not meet.

daily-archive.sh

#!/usr/bin/env bash set -euo pipefail day=$(date -u +%F) mkdir -p "archive/$day" for url in https://example.com/ https://example.com/pricing https://example.com/terms; do name=$(echo "$url" | sed -E 's#^https?://##; s#[/:?&=]#_#g') curl -sf --get "https://api.snapshotflow.com/screenshot" \ -H "X-Api-Key: $SNAPSHOTFLOW_API_KEY" \ --data-urlencode "url=$url" \ --data-urlencode "full_page=true" \ --data-urlencode "format=webp" \ --data-urlencode "cache=false" \ -o "archive/$day/$name.webp" done

Run it from cron or a scheduled CI job and store the folder in object storage. To get alerted when something changes instead of browsing folders, compare captures with a visual diff, as described in website change detection. If you need captures to hold up as evidence, record hashes too; see screenshots as evidence.

Rights and archived pages

An archived copy is still the original owner's work. Looking at it and using it for personal research is generally fine; republishing screenshots of archived pages follows the same copyright rules as live pages. Archived pages can also contain personal data, so treat them with the same care. The details are in is it legal to screenshot a website.

Takeaways

The Wayback Machine is the first stop; use /web/*/ and dated URLs to jump straight in.
Google cache is gone; old tutorials that rely on it no longer work.
The CDX API lists captures as JSON for bulk research.
Archives have gaps; for pages that matter, keep your own dated screenshots.

Sources

FAQ

How can I see what a website looked like years ago?

Open web.archive.org, paste the address and pick a date in the calendar. The Wayback Machine has stored billions of pages since the 1990s. If it has no copy, try archive.today or a national web archive.

Does Google still show cached versions of pages?

No. Google removed the cached-page link from search results in early 2024 and the cache: search operator stopped working later that year. Google points to the Wayback Machine instead, in the About this result panel.

Why does an old page in the Wayback Machine look broken?

Archives store a page's files separately. If images, stylesheets or scripts were not captured at the time, or the site was excluded, the page shows up unstyled or with gaps.

How do I make sure my own website history is preserved?

Do not rely on third-party archives alone. Schedule your own full-page screenshots of key pages, store them with the date, and optionally submit pages to the Wayback Machine with Save Page Now.

Can I reuse screenshots of archived pages?

Archived copies are still the original owner's content. The same copyright rules apply as for a live page, so personal research is usually fine but publishing needs permission or a legal exception.

Never be missing a page's history again

Schedule full-page captures of your key pages, store them by date, and compare any two days with a visual diff. The free tier gives you 300 screenshots, no credit card required.

Create free account Change detection guide Read API docs