Live change tracking, SEO devtools, hreflang validation and page-one ownership

The headline of this release is change tracking, which shows you how a URL's important SEO attributes have changed since you last visited it. Every visit is recorded on your own machine – the history is stored locally and never leaves your system – and builds into a picture of how the sites you work on and watch are changing over time.

Open the 'Changes' tab in the on-page analysis panels to see each date a page was visited across a maximum 24-month window, which can be customised based on your storage requirements. 'Timeline' and 'Compare' give you two views of that same dataset: a running list of changes over time, or a side-by-side look at the specific dates you care about. Say the open graph image for a page has been swapped since your last visit. The structured output lets you copy the old og:image URL straight out and compare it against the new one, so you can see what the site in question has been working on.

To browse everything you've tracked, click the three vertical dots in the top right of the app (beside 'Ask AI') and choose History > Show all history, or press ⌘Y – you'll get every page stored locally along with the number of changes recorded between visits. Change tracking is configured under 'Settings' in the bottom left corner, then in the 'Browser' section, where you can set how long history is kept, clear it, and exclude specific URLs and parameters. We exclude parameters such as srsltid automatically by default, since it's appended to every URL from search when auto-tagging is enabled in Google Merchant Center.

SEO devtools

Our Chrome DevTools integration is now being replaced with our own version, named SEO devtools. Because the feature set for Chrome DevTools is largely irrelevant to SEO (at least not directly impactful), we have narrowed in on the most impactful application of devtools for SEO. Every entry point you already use – the toolbar button, the keyboard shortcut, right-click → Inspect – now opens the SEO drawer instead, and the old right-sidebar Alerts panel has been retired into it with no findings lost. Our new SEO-native tools drawer includes:

• Elements: read the served HTML and the live DOM side by side, pick any element on the page and land on its DOM node (shadow DOM and iframe content are flagged), and diff any two captures – served HTML per user agent, initial DOM, current DOM – with differences highlighted in place. Minified HTML is pretty-printed automatically, and the DOM tree loads lazily rather than stopping at an arbitrary node cap.
• Network: every request with status, redirect hops, timing, and initiator, including traffic from cross-origin iframes and workers. The current page's redirect chain – with indexing directives served on intermediate hops and any hop disallowed by robots.txt – is one click away from anywhere in the drawer.
• Robots evaluation: test requests against robots.txt as a chosen search or AI bot, see exactly which rule blocks a request, and re-render the page without the blocked resources to see what that bot's render loses.
• Response body search: answer "which file contains this canonical, noindex, or string" without leaving the app.
• Console and Storage: read-only views of load-time errors and warnings, cookies, and site storage – no code execution, no editing.
• Issues: every rule-engine finding for the current page with severity, category, why it matters, and a recommendation, folded together with console and HTTP trace findings. Each piece of evidence jumps straight to the thing itself – the DOM node, the request, the comparison, the console entry.
• Handoff: HAR export and copy-as-cURL, so a finding can go to a developer without anyone needing Chrome DevTools.

Two things are worth calling out beyond the feature list. First, everything the drawer can show, the in-app AI will be able to read through its own tools (in our next release): the agent can walk a redirect chain, search response bodies, evaluate robots access as a chosen bot, compare fetches and renders, and read console, storage, and issues with their evidence – the same diagnosis you could perform yourself, without driving the UI. Drawer capabilities and agent tools now ship together rather than one trailing the other. Second, this is a clean cut: Chrome DevTools and our SEO capture both rely on the Chrome DevTools Protocol and cannot attach to the same page reliably, so leaving DevTools in place would have meant that opening it could silently corrupt the SEO analysis you were looking at.

Hreflang validation

SERP Lens has always shown the hreflang declarations it finds on a page, but seeing a list of alternates tells you nothing about whether those alternates actually work. Hreflang validation closes that gap: the declarations for the current page are visible up front, and when you choose to run a check, we fetch each alternate and tell you which declarations are healthy, which are missing a reciprocal link back, which point at unreachable targets, and which conflict with the alternate's own canonical. Nothing is requested automatically on page load – validation runs only when you ask for it – and you can watch progress as it goes, cancel partway through without losing the results already returned, and rerun the check when a page's declarations change.

Declarations are read from the rendered DOM as well as the HTML source, so sites that inject their alternates client-side get complete results rather than a false empty set, and header-delivered alternates now reach the panel too. Issues are summarised above healthy results so large alternate sets stay readable, each one naming the affected language/region and target URL with a short explanation of what needs attention. To access the new hreflang validation tool, you can find it within the 'Indexability' tab in the default 'Indexing' sub-tab. Simply click the 'Check X URLs' button to run the analysis.

Who owns page 1

Run a search in classic Google results, and the SERP Features panel now carries a 'Who owns page 1' card, which aggregates the results page into a ranking of the domains actually controlling it. Each row gives a domain, its highest position, how many listings it holds, and the total pixels those listings occupy – measured from the rendered result blocks on the page in front of you, not estimated from rank-slot heights – and you can switch the ranking between listing count and pixel footprint depending on whether presence or real estate is the question. If a domain has been set up as a project within SERP Lens, the folder icon now displays as assigned to entities within the tooling.