How to see when a website was last updated

A date printed on the page

The quickest method is the one the page prints itself: a line reading Last updated or Last modified with a date beside it, usually under the headline or in the footer. It is the first place to look, because it is already there.

Its honest limit: that date is written by whoever published the page, and nothing checks it. A shared template can reprint a fresh date over text nobody has touched, and a date typed by hand can sit unchanged for years.

The Last-Modified response header

Every page arrives as an HTTP response, and a response can carry a header named Last-Modified holding the time the server believes the page was last written. It is readable without a browser: a request for the headers only, such as curl -I, prints Last-Modified when the server sends it.

Its honest limit: many servers never send that header at all, and where it is sent it is the time the file or the page output was last written on the server. A site that rebuilds on a schedule moves that time on every build, even when not one word changed, and a page built fresh for each request can report the time of your own visit. It is a server time, not an edit time, and it says nothing about which words moved.

The lastmod value in sitemap.xml

Many sites publish a sitemap.xml, a list of the addresses on the site with an optional date on each entry named lastmod. It is meant to be the date that address last changed, and it is easy to open: it usually sits at the site address followed by /sitemap.xml.

Its honest limit: the value is written by the site's own publishing software, so it is only as good as that site's habits. Some sitemaps stamp every address with the current date on each build, which says nothing about the text, and some carry values so stale that a real edit never shows up. A lastmod is a claim a site makes about itself.

Archived copies in the Wayback Machine

The Internet Archive keeps dated copies of public pages at web.archive.org. Opening the same address in two archived copies shows what the readable text said on the two dates the archive happened to save it, so the copy carries its own evidence of the day it was taken, with no printed line to trust.

Its honest limit: the archive saves a page when its own crawler reaches it, or when somebody asks it to save one. A snapshot therefore gives you the day the archive looked, not the day of the last edit: a page edited on a Monday and saved a month later looks untouched for that month. A page the archive has never visited has no copies to compare.

The date Google shows on a result

Google and other search engines often print a date beside a result, which makes this the fastest method of the five: the date is already on the screen in front of you, next to the link you were going to click anyway.

Its honest limit: the engine works that date out from whatever it has, which may be the day it first saw the page, the day it last crawled it, or a date it read off the page itself. None of those is the edit date, and the same address can show one date in one engine and another date in the next. Read it as a hint, not as a record.

To know the next change, watch the page

None of the five methods tells you when a page will change next. Four of them are the site's own word about its own past, and a result date is a guess. To learn about the next edit, keep a copy of the readable text and compare it on a schedule, which is what a monitor does.

Point a monitor at the page and the next change is emailed to you with the added lines marked + and the removed lines marked -, and the same change stays in the monitor's history with a side-by-side diff. A page that only repaints a clock, a date or a cookie banner does not count, because those lines are stripped before two versions are compared, so an alert means the words really moved.

That is the difference between dating a page and watching it: one tells you a number you have to interpret, and the other tells you what changed and when.

Try a page before you sign up

The preview tool on the home page (/) needs no account. Paste a public page address, add a selector if you want one, and read the exact text that would be watched. Up to 20 previews a day are allowed from one address.

The pricing page (/pricing) lists the plans, the interval each one allows and what each one costs. It reads the same plan configuration the server enforces, so the numbers match.