Website Change Monitor

Why show you the text first

Every change monitor asks you to trust that it sees what you see. The commonest way that trust breaks is a page that draws its content with JavaScript: the server sends an almost empty shell, the monitor compares that shell to itself forever, and you find out weeks later that the thing you were watching changed and nobody told you.

So this shows the extracted text up front and says plainly when a page looks like an empty shell. It is the one check worth doing before setting up any monitor, on this site or anyone else's.

What is ignored, and what is not

Raw HTML comparison fires constantly, because a page's code changes far more often than its content. Stripped before anything is compared:

• Scripts, styles and SVG. None of it is visible and all of it changes on deploys.
• Every attribute. CSRF tokens, cache-busting query strings, generated element ids and analytics hooks all live in attributes and all change per request.
• Whitespace. Reformatting a source file is not a change to the page.

Then a line only counts as new if no removed line matches it once digits are blanked. That single test separates the two things that look identical to a percentage:

The cost of that rule, stated plainly: it cannot tell a price from a clock, so a price change reads as noise. Watching numbers needs a switch this does not have yet.

What it will not do

• Pages behind a login are not supported.
• Pages that render with JavaScript are read as the server sends them, which usually means no text at all. The preview tells you when that is the case.
• Responses over 2 MB, and anything that is not HTML, are refused rather than half-read.
• Ten tools that do offer recurring monitoring today, with every price read from their own pricing pages.