What can I compare between two URLs?

Rendered text, the served HTML, the metadata block, response headers and every outbound link. The summary tells you which of those changed at all, so you can go straight to the one that did.

Can I diff staging against production?

Yes, provided both are reachable from the public internet. Both pages are fetched at the same moment, so you are measuring the difference between the two URLs rather than the difference between two points in time.

Why does the diff miss content I can see in my browser?

Because the comparison uses the HTML the server sends, without running JavaScript. Anything a page assembles client-side is not included — which is also what search crawlers see, so a large gap here is worth knowing about.

Web Diffserver-backed

Fetch two pages at once and see exactly what differs — text, markup, metadata or headers.

Enter two URLs to compare them.

Both are fetched at the same moment, so what you see is a difference between the pages rather than a difference between two points in time.

Runs on R3X infrastructure. Both URLs are fetched by an R3X server and returned to your browser, where the diff is computed. Neither page is stored.

Reference

What gets compared

View	Compares
Text	The readable text of each page, with script, style and markup stripped. The right view for 'did the copy change?'
HTML	The served markup, one tag per line. Catches attribute and structural changes the text view cannot see.
Metadata	Title, description, canonical, H1s and the social tags, field by field.
Headers	Every response header from both sides, aligned by name.
Links	Every resolved href on each page, sorted. Finds navigation and footer changes quickly.

Both pages are fetched at the same moment

The two requests go out together. That matters: comparing a page you loaded a minute ago against one you loaded now measures the passage of time as much as the difference between the two URLs. Fetching in parallel means a rotating advertisement or a timestamp in the footer is the only such noise you have to reason about.

It also means neither page is rendered. Anything a page assembles with JavaScript after load is not part of the comparison, in the same way it is not part of what a crawler sees.

How the diff is computed

Myers' algorithm over lines, in your browser — the server only fetches the two documents and hands them over. There is a ceiling on the edit distance: two pages with almost nothing in common produce a diff that is longer than either document and useful to nobody, so past that point the tool says the pages are unrelated and shows the first 200 lines of each instead of pretending to align them.

Restrictions

Both URLs must be public. Comparing staging against production works when staging is reachable from the internet; it does not work for localhost or an address on your private network, which R3X refuses on purpose. Each page is downloaded up to 1 MB — enough for the markup of essentially any real page, and the tool says so when it has to stop early.

Questions

What can I compare between two URLs?: Rendered text, the served HTML, the metadata block, response headers and every outbound link. The summary tells you which of those changed at all, so you can go straight to the one that did.
Can I diff staging against production?: Yes, provided both are reachable from the public internet. Both pages are fetched at the same moment, so you are measuring the difference between the two URLs rather than the difference between two points in time.
Why does the diff miss content I can see in my browser?: Because the comparison uses the HTML the server sends, without running JavaScript. Anything a page assembles client-side is not included — which is also what search crawlers see, so a large gap here is worth knowing about.