Audit every page. See exactly what to fix.

A page audit becomes a record

Sukat SEO already told you what was wrong with a page. Now it tells you what changed on it, and when — every scan compared with the one before it, grouped by day, and any two versions comparable.

Added

• A change log for every page. Each scan is compared with the one before it, and only the fields that actually moved are recorded. A page audited weekly keeps a readable history instead of fifty identical snapshots.
• Day-by-day grouping. Updates are grouped under the day they happened, newest open and the rest folded with a count, so a page with a hundred entries opens as a short list rather than a wall.
• Compare any two versions. Pick two points in a page’s history and copy the difference between them. The copy control stays in view while you choose, however long the list.
• 7, 28 or 90 days on any page. The Top queries card takes a range, so a page’s recent movement is one click away instead of a trip to Search Console.
• Sortable Top queries. Every column sorts — query, clicks, impressions, CTR and position. The first click puts the best values first, a second reverses it, and sorting reaches the whole query list rather than the ten rows on screen.
• Search for the homepage by name. Typing home or homepage finds the site root. Searching / matched every path and filtered nothing, which left the homepage as the one page no search could isolate. This works everywhere pages are listed.
• Page weight is now a finding. A page carrying more than 2 MB of images is flagged for fixing rather than review — that is more than a whole page is meant to weigh, images and everything else together.
• Findings link to what they counted. “10 images over 100 KB” opens those ten images, filtered to that page.
• Re-audit a single page from its report. No need to run a whole site to check one fix — and it says what happened: “Score 91 → 95. Recorded in Updates”, or “No change since the last audit”.
• A seat list for the licence. Five Chrome profiles, showing which is which and when each was last used. The owner can read their own licence key in Settings, masked until asked for, and stored on that device only.

Changed

• Search Console drift stays out of the way. Clicks and impressions move every day; a title moves rarely. Copied audit cards list what changed on the page and leave the traffic figures to the Search Console section above them.
• Reconnecting happens where the problem is. An expired connection now says so and offers to reconnect, on the report itself, rather than a refresh that cannot work or a trip to another surface.
• Optimise acts on what you can see. Filter a page’s images and the button offers exactly the rows in front of you. Opening a page from a filtered list shows that page’s real totals, not the subset the filter left.
• Giving up a seat takes effect immediately on the profile that gave it up. An install that predates Pro returns to what it had rather than dropping to the free caps, and renaming follows ownership — name your own; the owner can name any.
• Licence key masking. The key on Settings reads 83A1-XXXX-XXXX-XXXX — the first four characters, then a fixed placeholder, instead of showing the last four as well.
• Clearer wording throughout: >100 KB instead of “heavy”, Licence ID for the reference that is not your key, Lifetime in place of a renewal date that never applied. Settings is reachable from the extension menu, not only from the site report.

Fixed

• Search no longer forces every page group open. On the Images and Links tabs, a search expanded all matching groups at once, burying a short result list under hundreds of rows — and the chevron stopped working, so a group could not be closed while anything was in the search box.
• The locked site report no longer shows controls it cannot use. On the free plan, Refresh, Print and Export CSV stayed in the header of the gated report alongside a freshness marker with nothing to describe. The gated page now keeps its title and nothing else.
• A scan returns to the top of the report, so a change summary is not left off-screen above. Turning on run notifications asks once rather than twice.

Under the hood

• Tightened which origins are allowed to talk to the extension.
• Audit data still never leaves the machine. Search Console figures are fetched with your own Google sign-in and stored locally; nothing is uploaded, and no account is required to use the extension.

Whole-site work, and a Pro tier

The Search tab puts your Google Search performance beside the audit for the whole property, not just the page you are on — and Pro adds the whole-site work: the site report, the structure map, unlimited bulk runs and your own logo on every export.

Added

• The Search tab. Whole-property clicks and impressions over time on a dual-axis chart, top queries and top pages side by side, device and country splits, and which query lands on which page. Every query row carries the landing page’s audit score, so a query ranking on a weak page is visible without leaving the tab. Hovering the chart reads out that day’s figures.
• Profiles and licence. One licence covers five Chrome profiles. See which profiles are using it, rename them, hand ownership to another profile, and free one up at any time. The first profile activated owns the licence and is the only one that can remove others — so sharing a key doesn’t hand over the power to evict you. A profile unused for 90 days frees itself.
• White-label exports. Your logo, agency name and colour on every card, report and structure map you send a client. A brand with a name but no logo gets its initial as a mark rather than ours. Per-site brand profiles, so each client’s reports carry their own identity.
• Set aside. A finding you have judged and decided not to act on can be set aside. It leaves the score and the counts, but stays visible in its own tab with an undo — so “not doing this” is a recorded decision rather than a finding you keep re-reading.
• Pages with no inbound links. The Overview says how many audited pages nothing else links to, and the wording changes with coverage: below two thirds of the sitemap it says so rather than reporting a number that is mostly an artefact of what has not been scanned.
• Structure map, exported. The whole depth map copies to the clipboard as a PNG at twice screen resolution, branded, with the site’s own name and figures in the header.
• Storage and data, in one place. A Settings page showing what the extension has stored, every audited site with its size and score, backup and restore per site, your licence, and your brand profiles.
• Image resize. An upscaled image — one served far larger than it displays — can be handed to the Sukat compressor at the size it actually renders at, in one click. Nothing is uploaded: the compressor fetches the image itself from its own address.
• Bulk Search Console sync. Sync Search Console figures for many pages in one run rather than a page at a time.

Changed

• Everything free stays free. Every install that already had audits keeps unlimited sites, unlimited bulk runs, re-auditing, the site report and the structure map, permanently. Pro adds white-label exports, the Search tab, bulk Search Console sync and one-click image optimise — none of which existed before, so nothing has been taken away from anyone.
• The date range is a list, not a calendar. Two clicks set a range, with a reset and a cancel, and it dismisses on an outside click.
• Search Console figures now cover the range you picked. The top-queries list under the trend used to be summed from whatever each page last synced, so it never moved when you changed the range. It now comes from the same window as the chart above it.
• The trend stops where the data stops. Google finalises Search Console figures two to three days late. The chart used to fill those days with zeros, which drew a cliff at the right-hand edge and pulled the totals below what Search Console itself reports. The window now ends on the last day Google actually reported, and the range chips say “90 days” rather than “3 months” — because ninety days is what is fetched.

Fixed

• A page with no position data no longer reads as an error. One day missing its average position turned the whole figure into NaN.
• The structure map export was missing a row. The header’s page, orphan and cluster counts were read with a selector that matched nothing, so every exported map went out without them.
• Copy buttons confirm what they did. The tick and its colour survive the redraws that happen while a report is loading, instead of vanishing before they are seen.

Every goal page measured separately

With more than one goal marked, distance is no longer collapsed to whichever goal happens to be nearest — and long reports page properly instead of showing only the worst twelve.

Added

• When a site has more than one goal page. Marking a page as a goal measures how many in-content clicks every other page is from it. With several goals marked, that number was the distance to whichever goal happened to be nearest — so a page one link away from Contact looked healthy even when nothing on the site led from it to the service you actually sell. The structure map now measures each goal separately.
• Reach by goal says how much of the site can get to each goal at all, one line per goal with a bar for the share. For a clinic whose specialism is one page and whose contact form is another, these usually differ sharply, and the gap is the finding.
• How far each page gets puts the whole site in one bar: how many pages reach every goal, some of them, or none. Six worst cases said nothing about the shape of the work.
• Closest to fixing lists the pages that already reach something, ranked by how few links close the gap, with the steps to each goal in its own column. Pages that reach nothing at all are a separate list, because the fix is different in kind — those need a route built, not a link added. All of this appears only when more than one goal page is marked.
• Pages the sitemap does not list. The Audited count includes pages the sitemap never mentions — scanned by hand, reached through a redirect, or a URL variant — which made it read higher than the sitemap pane’s figure with nothing to explain the difference. The sitemap pane now says how many, and a chip lists them. A page worth auditing that the sitemap omits is worth knowing about.

Changed

• Reading a long report. The page list opens on every audited page rather than the twelve worst, with a chip for each score band and one for the whole set. Long lists are paged, twenty at a time, worst first across the whole set rather than only within a page — so page one is always the work.
• The Images and Links tabs page the same way, twenty site pages at a time, and both now say how many pages their findings are spread across.

Fixed

• An image with no source, or one waiting on lazy loading, was reported as broken. Nothing had been fetched, so there was no test to fail. Only images that actually failed to load are flagged now.
• A page selection made on one site survived a switch to another, where the panel showed the new site while the run would have gone to the old one’s pages. Selections are cleared when the site changes and reconciled against the current list, so the counts can no longer disagree with themselves.
• The picker’s header wrapped mid-phrase in a narrow panel.
• Column headings in the goal lists sat a few pixels off the columns they labelled.

A map of how the site actually links together

A full-tab view of the site’s real structure — built from the links written into the page copy, not the ones in the menu — plus a bulk audit that works through the sitemap a batch at a time.

Added

• Site structure map. A new full-tab view draws how a site actually links together. The homepage sits on the left and every page fans out to the right, one column per click from home, with the connections between them drawn as they exist in the page copy. Pages four or more clicks from the homepage are flagged, and each page shows how many pages it links out to and how many clicks it gets in Search Console — so a hub earning nothing is obvious at a glance.
• Menu links are excluded by default. Left in, every page sits one click from every other and the picture says nothing: a menu listing ten service pages makes all ten look like top-level sections. The map keeps the links written into the body of each page, which is what tells a search engine how the site is really organised. One toggle brings the menu links back when you want to see what a crawler can reach.
• Orphans, told apart from menu-only pages. Two lists sit alongside the map: pages nothing links to at all, and pages linked only from the menu — reachable, but with no editorial link pointing at them anywhere on the site. They are different problems and they are now distinguished.
• Goal pages, and how far every page is from one. Mark the page a site exists to drive people to — a booking form, an enquiry page — and every other page is measured against it: how many in-content clicks away it sits, or whether there is no route at all. The distance ignores menu links on purpose, because a contact link in the footer makes every page one click from the goal, which is true for a crawler and useless for a reader. What is counted is the route someone reading the page could actually follow. Pages with traffic and no route are ranked by the clicks at stake, so the fix — one in-content link from each — is ordered by what it is worth.
• Reading the map. It scales to fit with zoom controls or a fit-to-window button, drags to pan, and exports the whole tree as an image at twice screen resolution for sending on. Findings float in a corner panel that collapses out of the way and stays put while the map moves behind it.
• Bulk audit. Audit a run of pages from the sitemap in one go rather than one at a time. Choose a batch — ten, twenty-five, fifty, or everything — and the pages are scanned in sequence in a background tab that never steals focus. Progress, the page being scanned, what has finished and what is still queued are all visible while it runs. It pauses and resumes where it left off, and pages that fail are listed with the reason rather than silently skipped.
• What a bulk run leaves you with. Every page from the run is listed with its score and its fix and review counts. Click any of them to open that page’s audit; a page that failed opens in a tab so you can see what is actually there. The next batch is one click away, and there is a way back to the setup screen when you want to change the size.

Changed

• Link checking is switched off during a bulk run. Each page would otherwise fetch up to two hundred internal links, which is the slowest part of an audit and a great deal of traffic at someone else’s server. Scanning a page on its own still checks them. Versions created in a bulk run are marked in the page’s timeline, and the link fields record that they weren’t checked rather than reporting zero broken links — a scan that didn’t look shouldn’t read like a clean bill of health.

Fixed

• A repeat restore counts correctly. Records identical to what was already saved are now counted as unchanged rather than overwritten, so restoring the same file twice reports what it actually did.

A redirected page keeps its audit

An audited page that starts redirecting no longer drops out of the list, and the search snippet gains checks for duplicate title and description tags.

Added

• Redirected pages keep their audit. When a page starts redirecting somewhere else, the audit you already ran on it stays put. The page keeps its row, its score and every finding from before the redirect, and gains a status badge showing where it now points. Scan the destination and Sukat SEO works out which of your audited pages redirect to it, then marks them. The redirect is recorded in the page’s history too, so a chain appearing, changing status code or going away reads as a dated event on the timeline alongside every other change.
• Duplicate title and meta description tags. Two checks were added to the search snippet: exactly one title tag, and exactly one meta description. Both are common on sites where more than one plugin or theme writes head tags — and both are near-invisible from the page itself, because the browser silently uses the first and ignores the rest. The count reads the page as rendered, so a second tag injected by script after load is caught as well. Titles inside inline graphics aren’t counted, and a description tag is matched whatever its capitalisation — which is exactly the case where a second one usually hides.
• Restore tells you what it did. A summary appears when a restore finishes: audits added, updated or already identical, history versions gained, screenshots added, and what was merged. Restoring the same file twice reports everything as unchanged rather than claiming work it didn’t do, and the panel waits for you rather than reloading out from under the result.

Changed

• Menus opening from the report header now point back at the button that opened them, and the text size controls adopted the same treatment as the rest of the header menus.

Fixed

• Restoring a backup no longer discards history. Restoring used to overwrite a page’s timeline with the backup’s copy of it, so every scan recorded since the backup was taken vanished. Both timelines are now merged: entries are interleaved by date, and versions saying the same thing collapse into one, the same way repeated scans of an unchanged page already behave. Nothing recorded is lost.
• Screenshots restore safely alongside them. Pictures from a backup taken on another machine used to land on top of unrelated local ones; they now take their own place in the gallery.

The route the page took to load

Redirect chains are reported hop by hop, and a change log above the check list shows what moved since the last audit of that page.

Added

• Redirect chains. Sukat SEO now reports how the audited page was actually reached. If the URL redirected before the page loaded, every hop appears with its status code, along with the final URL the browser landed on. The panel flags the cases that matter: a temporary redirect where a permanent one was probably intended, a trailing-slash hop, a final URL that differs from the one requested, and chains long enough to be worth collapsing. The chain can be copied or exported alongside the rest of the audit. The page itself only shows where the browser ended up, never the route it took, so this is information no on-page check could surface before.
• Changes since the last audit. A change log sits above the check list. After a re-scan it shows what moved since the previous audit of that page — which checks changed, in which direction, and what the values were before and after. Rows are filterable, and selecting one jumps to the check it belongs to.

Search Console says when the sign-in has expired

Refresh no longer fails silently, the panel can’t stick on “Connecting…”, and a link opportunity copies the URL rather than a full HTML tag.

Changed

• Copy link replaces Copy <a> tag. The panel card and the on-page popup now copy the destination URL on its own. The full anchor only helped when pasting into raw HTML; in a block editor or a CMS link dialog it arrives as literal text, and what those ask for is the URL. It does mean the anchor text is no longer carried along, so pasting into raw HTML means writing the tag yourself.
• Screenshot capture and the gallery moved into the action row, alongside notes, to-do and delete, instead of sitting up in the score panel.
• Text size moved into a menu behind a button in the header’s right corner — smaller, reset, larger, with the reset row doubling as the current level. It stays open while you step through sizes, and closes on the button, an outside click, or Escape.
• Change-log rows jump to the section they describe. The log collapses, the checks return, and the card is expanded, scrolled to and flashed. A row whose section isn’t on the board isn’t clickable, so it can’t lead nowhere. One trade-off: those rows are buttons now, so you can no longer drag-select a title out of the log.

Fixed

• Refresh reported nothing when the sign-in had expired. The failure was shown as a passing message that the next redraw wiped, and nothing was recorded, so the report card and the collapsed card showed no problem at all — pressing refresh simply looked inert. The bar, the card banner and the collapsed card now report the same state and offer Sign in, and a refresh that succeeds clears a previously recorded error rather than leaving it up.
• The panel could stick on “Connecting…” with nothing to press. The connect that runs on navigation had no time limit, so a request that never settled left every surface reading Connecting. It now gives up after 20 seconds. A successful automatic connect also used to skip the figures refresh that follows it, so the numbers arrived a beat late for no reason.
• Link-opportunity buttons were invisible on some sites. The popup is added to the page you are auditing, so anything it doesn’t set explicitly is inherited from that site — Ignore and Close took the site’s own button colour, which on a white popup meant white on white. Every property that can swallow a label is now set outright, and the same hardening went to the on-page arrows.
• Jumping to a card landed it behind the sticky header. The card you had just jumped to was the one thing you couldn’t see. The scroll now accounts for the header’s measured height, leaving the card just below it.
• Two escaping fixes from the pre-submission security review. An image row’s format value and the on-page popup’s target URL were both inserted without escaping. Neither was reachable in this build — the extension’s own content policy blocked one and browsers encode the character that would have triggered the other — but the popup goes into someone else’s page, where that policy does not apply, so neither now depends on those defences.

Request indexing opens the right Google account

On a machine with several Google accounts signed in, the link to Search Console used to land in whichever one Chrome treats as default.

Changed

• Switching accounts feels faster. The panel used to keep showing the old connection prompt, still naming the account you had just replaced, until all the new figures had loaded. It now updates as soon as the property is matched, with the figures filling in behind it.
• The report header no longer has a re-scan button. Re-scanning is reached from the site view, or from the notice that appears when a saved audit no longer matches the live page.

Fixed

• Request indexing opened the wrong Google account. The link to Search Console’s inspection tool was opened without saying which account it was for, so Chrome used its default — which on a multi-account machine meant landing in a Search Console that doesn’t have the site. The link now names the account the site is connected with.
• Backing out of the account picker no longer signs you out. Starting an account switch and then closing the chooser used to leave you signed out of the account that had been working, after which every other site reported an expired sign-in. Cancelling now leaves the existing connection alone.

Every image measured directly, and schema you can actually locate

Image sizes no longer depend on what the page reports about its own resources, so CDN-hosted files are real numbers rather than gaps — and every schema entity now says where on the page it came from.

Added

• Structured data tells you where it came from. Finding a schema block is often harder than fixing it: a page can carry four or five JSON-LD blocks, none of them visible, emitted by a plugin, a theme or a tag manager. Every validated entity now carries a Found in section — which block it belongs to and whether that block sits in the head or the body, the region of the page it sits in with the chain of containers leading to it, the line number in the page source with an expandable view of the surrounding lines, and the SEO plugin responsible where one can be identified.
• Whether it was in the server’s response, or added by a script. If a block wasn’t in the HTML the server returned, searching the codebase will never find it — it came from a tag manager or client-side code, and the card says so.
• Show on page. A button reveals the block in place on the page itself. Microdata and RDFa, which do attach to visible elements, can be scrolled to and highlighted directly.
• Every image row shows its measurements. File size, format, and both dimension pairs — the file’s own size and the size it is displayed at — now appear on every image, labelled so there’s no ambiguity about which is which. Previously these only appeared on images that already had a problem, so a clean image told you nothing about itself.

Changed

• The image audit measures every file directly. Sukat SEO used to read file sizes from what the page reported about its own resources. That works for images on the same domain, but most sites serve images from a CDN, and a CDN usually doesn’t expose size information back to the page — so page weight was a minimum rather than a figure, and the over-100KB count covered only the handful that could be read. Page weight, that count, and every per-image size are now real numbers regardless of where the file is hosted.
• Formats are read from the file, not the filename. A CDN that converts images on the fly will serve WebP from an address ending in .png. The audit reports what the browser actually received.
• The same file used twice on a page counts once, rather than once per placement, which used to inflate the image count and double its contribution to page weight. A full-size image linked from the thumbnail that displays it is likewise one image, not two.
• Lazy-loading placeholders are no longer counted as images. Many sites put a 1-pixel transparent placeholder in an image slot until the real file loads. Those were listed as their own image — unmeasurable, and flagged as upscaled, because one pixel stretched to fill a container looks like extreme upscaling.
• Background images and gallery links are checked for dimensions. These are never decoded by the page itself, so they couldn’t previously be judged upscaled or oversized. They are now measured like ordinary images.
• The statistics and the list they describe can no longer disagree. The count at the top of the card and the images the filter returns come from the same data.
• Results match Sukat Inspector. The two extensions audit images with the same rules and report the same figures for the same page.
• Scans of image-heavy pages take a little longer. Measuring directly means any image whose host won’t report its size on request has to be downloaded to be weighed, which also uses more data than before. Audits saved by earlier versions were produced under the old counting rules and refresh automatically the first time you open them.

Fixed

• A page could be reported as a duplicate of itself. Title and description uniqueness checks compared page addresses as plain text, so the same page stored with a trailing slash, a hash or a stray parameter failed to recognise itself and appeared in its own duplicate list.
• A warning with nothing to show for it. The structured data card could report an issue to review while every entity showed as valid. The finding — entities not linked to each other by ID — was real but never displayed. It now appears with an explanation.
• Search Console connection state is honest on every surface. A signed-in account with no verified property for the site now says so, names the account and offers to switch, rather than showing a Connect button that fails the same way each time. An expired sign-in is reported wherever figures appear and survives reopening the panel, instead of keeping a “synced 7 hours ago” label with no working credential behind it. Cached figures are labelled as being from the last sync rather than presented as current, and controls that could only fail have been removed in favour of the action that resolves the situation.
• Entities start collapsed and open one at a time, so a page with several schema types stays readable.

Backups are now per-site

Backing up used to fail outright on installs with a lot of saved data. One site at a time, as a compressed archive, removes the limit that caused it.

Changed

• One site at a time, as a compressed archive. Everything on the device — every site, every audit, every screenshot — used to be packed into a single file, and past a certain size that file could not be built at all; the only report was that the backup could not be created. Back up a site from its row in the Data panel, or use the footer button to back up the site in the current tab.
• Backups no longer contain your Google account token. Search Console credentials are excluded, so a backup file can be moved or stored without carrying account access with it.
• Internal linking moved into the Links card rather than having a card of its own. The suggestions stay out of the pass/fail result for the page, because they concern other pages on the site, not anything wrong with the page being audited.

Fixed

• Restoring a site no longer affects your other sites. Notes and to-dos for sites not in the backup are left alone.
• Restore explains what went wrong. An incomplete download, a damaged archive and the wrong file type are reported separately instead of as a single generic failure. Backups made with earlier versions still restore.
• Dismissed link suggestions can be brought back. Dismissing every suggestion on a page previously left no way to reach them again; the dismissed list is now always accessible.

Edit a task without retyping it

Fix a typo in place instead of deleting the task and losing its due date with it — and a pasted paragraph no longer takes over the list.

Added

• Tasks can be edited. Until now a typo meant deleting the task and retyping it, losing its due date and page tags with it. An edit control appears on the task when you hover it, and the text opens in place — @-tagging keeps working while you edit, and existing page tags survive. Enter saves, Shift+Enter adds a line, Esc cancels. Clearing the field cancels rather than deleting; deleting has its own button.

Changed

• A long task no longer takes over the list. Pasted text clamps to three lines with a Show more that appears only when there’s more to see, so every task keeps a predictable height and the list stays scannable.

Fixed

• Pasted paragraphs no longer run together. A line break was dropped with nothing in its place, so “…the fact.” and “Every on-page…” collided into one word. Breaks now become spaces.

What Google sees, next to what the page says

The Search Console card now reports Google’s own index record for the page — including when Google has quietly picked a different canonical.

Added

• What Google sees. Alongside what the page declares, the card shows what Google’s index says: when it was last crawled and by which crawler, whether crawling and indexing are allowed, whether the fetch succeeded, which sitemaps list it, which pages Google thinks link to it, and which rich results it detected. Collapsible, open by default.
• Canonical disagreements are surfaced. When Google has chosen a different URL as the canonical version of a page, the card says so and names the URL it picked. This can’t be found by checking the page — the canonical tag is valid and self-referencing; Google simply overruled it, and the page is being folded into another as a duplicate with its clicks attributed elsewhere. Trailing-slash and protocol differences are ignored, so only genuine disagreements are flagged.
• The pages Google thinks link here can be expanded and opened in a new tab. This is Google’s own list, so it won’t match the Links card — it includes external pages and links Google knows of but hasn’t recrawled.

Changed

• Scanning a page also refreshes its Search Console figures, so the numbers beside the audit are from the same moment as the audit. It runs in the background and doesn’t hold up the report.

Every image on the page, and any Google account

The image audit stops missing CSS backgrounds and lazy-loaded files, and Search Console sign-in no longer follows whichever account your Chrome profile happens to use.

Added

• Search the images on a page. The Images card has a search box matching alt text or filename. It works alongside the stat filters rather than replacing them, so you can select Oversized and search within just those — the count tells you how many of the filtered set matched.
• Two more image stats: page weight and broken images. Weight is the combined size of every image that could be measured, with a + when some couldn’t be. Broken counts images the browser tried and failed to load.

Changed

• You can switch which Google account Search Console uses. Sign-in no longer follows whichever account your Chrome profile happens to use, so a site whose property lives under a different account is reachable without switching Chrome profiles. Each connected site remembers the account it was connected under, and a site belonging to an account you aren’t currently signed in as shows Connect rather than appearing connected and quietly failing to sync.
• The image audit counts every image on the page, not just <img> tags. CSS background images, lazy-loaded files that haven’t swapped in yet, <picture> sources the browser didn’t pick, and full-size files behind gallery links are all included — matching what Sukat Inspector reports for the same page. A hero image set as a CSS background used to be invisible here.

Fixed

• The image stats no longer claim a page is clean when they couldn’t measure it. Some hosts don’t expose file sizes to the page, in which case >100KB: 0 meant “nothing measured”, not “nothing heavy”. The card now says how many images it couldn’t weigh.
• Connecting Search Console tells you when it didn’t work. If the connected account has no verified property for the site you’re on, the card says so and names the account instead of appearing to do nothing. It also gives the two ways forward: have that account added to the property, or authorise the account that already owns it.
• A failed sync no longer reports success. The “synced” time was stamped even when the request failed, so a page could show week-old figures while claiming to be up to date. A sync that fails now says why.

Search the links, and a readable overdue reminder

Filter any link list by anchor text or URL, and read the daily reminder as rows rather than a wall of pasted links.

Added

• Search the links on a page. The Links card has a search box matching anchor text or URL, filtering whichever list you have open — internal, external, broken and the rest. A count shows how many of the list matched, and the query carries over when you switch between lists.

Changed

• The daily overdue reminder is readable again. Tasks ran together into one block of text, mostly made of pasted links. Each task now appears as its own row with the site it belongs to, and links are left out of the summary — they’re a reference, not the thing to do.

Fixed

• A to-do containing a long URL no longer runs out of its row. Pasted links had no wrap opportunity, so the text spilled past the edge of the panel. Long strings now break where they have to, and a page tag at full width no longer overhangs by a couple of pixels.

A report opens at the top, with every section closed

Sections no longer expand themselves, so a page with findings lands on the overview instead of halfway down someone else’s scroll position.

Changed

• Every section starts closed. Sections used to expand whenever they had something to report, so a page with several findings loaded already scrolled into the middle of expanded content. You now land on the Search Console card with an overview of every section’s status, and open the ones you want.

Fixed

• The content-scope notice no longer wraps into a narrow column. On a page with no <main> or <article>, the explanation beside the warning icon was squeezed into a thin strip; it now uses the full width of the panel.

Query opportunities from your Search Console data

The query list becomes findings: queries just off page one, queries the page never mentions, gaps in the title, clicks under what the position should earn.

Added

• Query opportunities. Where Search Console has enough data for a page, the card turns its query list into findings — queries sitting just off page one, queries the page never mentions, gaps in the title, clicks well under what the position should earn, and pages of yours competing for the same query. Collapsed by default, with a count on the header. How much evidence a query needs before it’s used is adjustable, and a page with too little traffic says so rather than showing nothing.

Changed

• The Top queries list pages five at a time, so the findings below it stay in view.
• Long values no longer overflow their row. A page tag with a long path is capped and truncates, with the full path on hover.
• A finished page tag reads as struck-through text rather than a live link, so the tags left to do are the ones that stand out.

Fixed

• Changing the query threshold or paging the query list no longer flashes the whole card.

Every to-do in one list, and a daily overdue summary

A cross-site to-do view in the data panel, and one morning notification covering everything still overdue.

Added

• All your to-dos in one list. The data panel has a Sites / To-dos switch. The to-do view gathers every open task across every site you’ve audited, grouped into overdue and upcoming, with the site and page each one belongs to. Tapping a task opens the page it’s tagged to.
• A daily reminder for overdue to-dos. A task that passed its due date used to be announced once and never again. Every morning at 8am, anything still open and overdue is summarised in a single notification — one message listing the tasks, not one per task. It stops as soon as the task is marked done, and the existing due-time and 30-minute reminders are unchanged.

Fixed

• Saved page URLs scroll instead of being cut off. A long path in the saved-pages list used to end in an ellipsis; it now scrolls sideways like the other long values, and works with a mouse wheel as well as a trackpad.

Copy-paste canonical tags, and safer screenshot deletes

The missing-canonical warning now hands you the exact tag to paste, and deleting a screenshot asks first.

Added

• The canonical warning writes the tag for you. When a page has no canonical, the finding now shows the exact <link rel="canonical"> line to paste — built from the page you’re on, with a copy button. Tracking parameters and #fragments are stripped first, while parameters that genuinely change the page, like ?page=2, are kept.

Changed

• Deleting a saved screenshot now asks first. Both the gallery and the full-screen viewer confirm before removing anything, and the prompt names the capture time so you can tell which shot it is.

Fixed

• Paging through saved screenshots no longer jumps twice. Moving to the next or previous shot used to shift the image once into a fitted layout and again into full-width, with the previous shot lingering while the new one loaded. Each page turn is now a single change, the outgoing shot holds until the next is ready, and a tall screenshot opens at the top rather than keeping the last one’s scroll position.

Pan a zoomed screenshot, and sideways scrolling for a plain mouse

Small, hands-on fixes: pan around a zoomed capture, scroll the sideways rows with an ordinary mouse wheel, and a steadier data panel.

Added

• Pan tool in the annotator. At high zoom, drag the screenshot to move around it instead of relying on scrolling. One drag moves both directions, which matters on a wheel or trackpad that only drives one of them.

Fixed

• Sideways scrolling now works with an ordinary mouse. The values that scroll horizontally — schema @ids, scan-history badges, Search Console pills, site details and the date filters — used to respond only to a trackpad, because a mouse wheel produces no sideways movement. A vertical wheel now scrolls them, and hands back to the page once the end is reached.
• Scrolling a zoomed screenshot no longer escapes to the page behind it part-way through a gesture.
• The data panel no longer scrolls as a whole. The title and totals stay put, only the site list scrolls, and the paging controls are pinned to the bottom so they’re reachable without scrolling to the end of the list.

Favicons and a date filter for the sites list

The audited-sites list gets easier to scan — each site by its favicon, and a date filter to narrow it down.

Added

• Favicons in the sites list. Each audited site now shows its favicon, so rows are recognisable at a glance, with the audit score tucked into the icon’s corner. Sites with no cached icon fall back to a lettered tile.
• Filter the sites list by date. Narrow the list to the last 24 hours, 7, 30 or 90 days, this year, or any time — by when each site was last audited. It works alongside the domain filter.

Fixed

• The domain filter no longer disappears when you have ten or fewer audited sites.

The data panel, a screenshot annotator, and saved page text

A big release: a data panel that breaks down everything Sukat has stored per site, a screenshot annotator with crop and save-as-new, saved page text with persistent region highlighting, and a round of History and layout fixes.

Added

• See what the extension has stored. A new button in the panel header opens a breakdown of everything Sukat has saved on your device — the total size, how it splits across audits, scan history, screenshots and notes, and how much of your browser’s storage allowance that uses.
• Every audited site in one place. The same panel lists each site you’ve audited, with its average score, page count, screenshot count, last scan and total size — until now it only ever showed the site in your current tab, so an old client could sit there taking up space with no way to spot it. Click any row to open that site in a new tab and the panel follows along, or delete it to clear every audit, screenshot, note, to-do and cached sitemap for that site. A filter and paging keep the list usable once you’ve audited a lot of domains, with your current site pinned to the top.
• Crop in the annotator. Crop a screenshot before you save it. Marks you’ve already drawn move with the image, and anything left outside the crop is removed — Enter applies, Esc cancels, and it undoes like any other edit.
• Save an annotation as a new screenshot. Keep the original instead of overwriting it — the Save button now offers both, and defaults to keeping the original.
• Page through saved screenshots. Previous and next arrows — and the arrow keys — move through everything saved for the current page without a trip back to the gallery, with a position counter to keep your place.
• Save the text of a page. Keep the text of a page, or a selected region, with a scan and read it again later from the History timeline.
• Re-highlight a saved region on the page. When you audit a selected region, Sukat remembers the exact spot it covered and keeps it across re-scans. Each saved region in History gains a small grid icon next to View — click it and Sukat paints a green overlay over that spot on the live page, scrolls it into view, and fades it out after a moment. It lands accurately while the page is unchanged; a different window width or edited content can move where it points.

Changed

• Long values scroll instead of being cut off. Structured-data type names and @id values, scan-history badges, Search Console pill strips and the site details now scroll sideways when they don’t fit, with a soft fade marking where there’s more to see, rather than being clipped with an ellipsis.
• Deleting an annotation mark is now undoable.
• Version deletion asks first. Deleting a saved version now opens a confirmation dialog instead of a two-click arm.
• Region re-selection no longer clutters History. Selecting a region only creates a new history version when its word count differs from the current one, so re-selecting the same area adds nothing new.
• Removing saved page text asks first. Unsaving text you kept with a version opens a quick confirm — Cancel, or a red Remove — instead of deleting on a single click. Saving stays one click; only removal is guarded.
• One clear place to save page text. Saving and removing live solely in the text viewer’s own Save this text button, reached from the View icon — no duplicate control on the word-count line.
• The History header keeps its buttons in place. The date and status badges scroll sideways when they overflow, so the copy and delete buttons stay pinned on the right rather than dropping to a second line, and the annotation palette and image stage scroll cleanly without a visible scroll track.

Fixed

• Redaction no longer stutters on large screenshots. The blur tool stopped re-reading the whole canvas on every drag, so pulling a redaction box across a big capture stays smooth.
• FAQ text is counted again. FAQ answers and the content below them were missing from the word count and the saved page text on some pages.
• Saved-text icons appear against the right version. They sometimes failed to show against the version the scan had actually landed on.
• The region-highlight icon no longer overflows its row.
• History sections stay put after a delete. They could collapse unexpectedly once a version was deleted.
• The Save button in the text viewer now works. It was silently failing on every click and saving nothing; page text persists again.
• The full-size screenshot preview opens above the Saved-screenshots list again. A recent layering fix had pushed it behind the modal.
• The screenshot-count badge is no longer clipped. The green count on the Saved-screenshots button had its corner cut off by the collapsible score summary; it shows in full.
• Ignored findings get their width back. A dismissed finding was cramming its text, buttons, the IGNORED tag and Undo onto one line. The View and Save buttons are now hidden on ignored rows so the text flows full-width, and the tag, Undo and the remaining buttons line up as an even row.

Annotate screenshots, collapse the score, sort by index status

Mark up your screenshots, give the score summary a rest, and find the pages Google hasn’t indexed.

Added

• Annotate your screenshots. Open any saved capture and draw on it — arrows, boxes, circles, lines, freehand, highlighter and text. There’s a pixelate tool for covering client names, API keys and anything else you’d rather not share, and it destroys those pixels rather than hiding them. Move a mark after you’ve placed it, undo anything, and zoom in when a long capture needs a precise arrow. Saving keeps the screenshot at its original size.
• Collapse the score summary. Tap the chevron under the panel header and the score, grade and stats fold away, giving your queries and category findings the room instead. The category filters shrink to icons at the same time. Sukat remembers the choice, so the panel opens the way you left it.
• Sort by index status. Sukat can now order your audited pages by whether Google has indexed them, putting the ones it hasn’t at the top. Sorting by audit score is there too, in its own section — so you can rank by Sukat’s own findings even on a site you haven’t connected to Search Console.
• Open any image full size. Thumbnails in the Images tab now open the real file in a new tab, so you can check a compression result without hunting for the URL.
• Recently scanned pages come first. The audited list is ordered by when you last scanned each page, with the page you’re on still pinned to the top.

Fixed

• Pages marked “Indexed, not submitted in sitemap” were shown as not indexed. They are indexed — Google simply didn’t find them through your sitemap. Some pages will change status after this update; the new one is correct.
• Hovering an image row sometimes highlighted a different image. On sites that lazy-load images or serve them through a resizing CDN, the highlight could land on the wrong one. It now tracks the right image even when the page rewrites its own URLs.
• Long schema @id values broke the schema tab layout. Entity rows stay one line each, and nothing gets pushed off the edge.

Changed

• System notifications are now optional. Sukat asks the first time you set a due date, rather than at install — so updating never interrupts you, and you only see the request if you use reminders. Turn them on later from any dated task. The toolbar count and the on-page reminder work either way, with no permission at all.

Task due dates and reminders

Give any task a deadline and Sukat reminds you — on the toolbar, on the page, and through your system notifications. A due-and-overdue badge on the icon keeps the count in front of you across every site you track.

Added

• Due dates on tasks. Any to-do can now carry a due date and time. Set one from the task itself and Sukat shows how long you have — “Today, 4:30 PM”, “Tomorrow, 9:00 AM”, or “Overdue” once the time has passed — and tints the task amber when it slips. Works on both page tasks and site-wide tasks.
• A reminder 30 minutes before, and again when it’s due. Every dated task gets two reminders. Set a deadline less than half an hour away and Sukat skips the early one rather than firing it late.
• A count on the Sukat icon. The toolbar icon carries a badge showing how many tasks are overdue or due in the next 24 hours, across every site you track. It turns amber the moment something is overdue and clears itself when you’re caught up — nothing to enable, and no permission to grant.
• An on-page reminder. When a task comes due, Sukat shows a reminder in the corner of the page you’re reading, with the task, the page it belongs to, and a button that takes you there. It waits until you dismiss it rather than disappearing on a timer, and dismissing it once clears it from every tab. Because it lives in the page, it reaches you even when your system has notifications switched off.
• Jump from a task to its page. Each task tagged to a page gains a button that opens that page in a new tab with the audit panel already docked and showing that page’s report — so you go from “fix the H1 on /pricing” to the page and its findings in one click.

Request indexing without leaving the audit

Send a page straight to Google for indexing from the report itself. Sukat opens it in Search Console’s URL Inspection tool, on the exact page — no copying URLs between tabs.

Added

• Request indexing. Every audited page now opens directly into Search Console’s URL Inspection tool, where you can submit it to Google. Sukat takes you to the exact page — no copying URLs between tabs.

Ignore findings, keep notes, and a fuller Search Console

Mute any finding you’ve reviewed and it drops out of every count on that page. Keep as many notes per page as you need, and Search Console gains sitemaps and a per-page summary — with per-page numbers held to a steady 28-day window, so a page’s history moves only when the page does.

Added

• Ignore audit findings, per page. Any finding that needs review can now be muted on a specific page. Click the Ignore (eye-off) control on a finding — it greys out, keeps an undo toggle so you can restore it, and drops out of every review tally: the category badge, the report header’s “to review” count, the category nav icons, and the homepage list. Ignores are scoped to that URL and persist across re-audits.
• Across every category. Ignoring works in every section — Search Snippet, Canonical, Open Graph & Social, Headings, Content, Links, Images, and Page Speed.
• Down to a single schema property. In Structured Data, mute an individual missing property — a recommended aggregateRating, say — from the icon on the property itself. The block’s status recomputes from what’s left, so clearing its last gap flips it from Review to Valid, and the section badge and all counts follow.
• Multiple notes per page. Notes are now a list rather than a single box — add as many as you need, each timestamped, then edit any one in place or delete it. Site-wide notes work the same way, and an existing note carries over as your first entry.
• Search Console sitemaps. The Search Console panel now lists the sitemaps Google has on record for the site: submission and last-read dates, discovered-URL counts by type, and processing status. Multiple sitemaps collapse into compact rows you can expand.
• Per-page Search Console summary. The page report header now shows the page’s clicks, impressions, and average position for the last 28 days, alongside its index status.

Changed

• A steady 28-day window per page. Per-page Search Console data is now fixed to a 28-day window, independent of the date filter in the home Search Console view. A page’s numbers stay stable, so its history and saved versions change only when the page itself changes — not when you switch the site-wide date filter.
• Date filter scoped to the overview. The date filter in the home Search Console view now affects only the site-wide overview, and the “synced” label reflects the range you picked instead of a fixed one.
• Value-first average position. Average position now reads value-first — “7.1 pos” — to match clicks and impressions.

Fixed

• Header and card agree after a resync. The report header’s Search Console line and the per-page card no longer disagree after a resync; both refresh together.
• Index status without a date range. The index-status line no longer shows a date range — index status is a point-in-time state, not a 28-day metric.
• No deprecated sitemap count. Removed the sitemap “indexed” count, which Google deprecated in 2019 and always returned as zero. Only real discovered-URL counts are shown.
• Centered gauge number. The score gauge number is now horizontally centered within the ring.
• Structured Data nav icon clears. The Structured Data nav icon now turns green once its findings are cleared or ignored, instead of staying on the review color.
• Live homepage counts. The homepage review counts now update immediately when findings are ignored on a page.

A steadier Search Console panel

Connecting and disconnecting Search Console is cleaner, long pill rows scroll instead of overflow, and scanning stays reliable as you move from one site to the next.

Improved

• Reliable scanning across sites. Moving from one site to another and scanning right away no longer trips a “Can’t scan this page” error — the first scan on a freshly navigated tab now works every time.
• Scrollable Search Console pills. Long rows of pills now scroll sideways within the panel instead of spilling past its edge.

Fixed

• Connect and disconnect behave. The connect prompt and the data overview are now properly mutually exclusive — the overview stays in place after you connect, and the prompt returns cleanly after you disconnect, with no leftover panel lingering behind.
• One click, one toggle. Opening or closing the Search Console overview now responds to a single click; a double-firing handler that made it flicker is gone.

Google Search Console, built in

The audit gains a second dimension. Connect Google Search Console and every score sits beside the real clicks, impressions, and positions a page earns in Google — no separate dashboard, no tab-switching.

New

• Search Console overview. Connect your Google account and the home board gains a site-wide Search Console panel — clicks, impressions, click-through rate, and average position for the whole domain.
• Per-page Search Console card. Every page you audit shows its own Search Console numbers next to its on-page score, so you see how a page actually performs, not just how it’s built.
• Search-data recommendations. Five new recommendations draw on your real Search Console data and fold into the existing audit cards, surfacing the pages and fixes most likely to move rankings.
• One-click connect. Connect through Google’s standard OAuth consent, and disconnect whenever you like. Your Search Console data is requested straight from Google in your browser — the same privacy model as the rest of Sukat.

Improved

• Refresh in place. A single refresh pulls the latest Search Console figures — site-wide from the home board, or just the current page from its card.
• Cached across disconnects. Search Console data stays cached per site and date range, even after you disconnect, so reconnecting brings your numbers straight back.

Open & re-scan in one click

The open-in-tab arrow now does two jobs at once, so a saved page and its score never fall out of sync.

Improved

• Open & re-scan in one click. The open-in-tab arrow in the URL bar now opens the saved page in your current tab and immediately runs a fresh audit — so you never act on a stale score.

Notes for a whole site

Notes grow up: keep a running note for an entire domain, not just individual pages — with a tidier panel and an at-a-glance marker.

New

• Site-wide notes. Keep a running note for an entire site, not just individual pages. It lives beside the tech-stack button on the site overview, autosaves as you type, and stays attached to the domain.

Improved

• Cleaner notes panels. Tidier spacing, rounded corners, and consistent alignment for both page and site notes.
• Green note markers. Pages with a note are flagged with a green pencil in the audited list.

Structured-data value tree, notes & snapshots

The biggest release since v1.4.0 — see every schema block’s full contents, attach notes to a page, copy any point in its history as a shareable card, and get a fuller comparison view.

New

• Structured-data (JSON-LD) value tree. See the full contents of every schema block on a page, laid out as an indented key/value tree — one collapsible section per block (WebSite, Organization, FAQPage, and so on). Changes are tracked over time, and the comparison card diffs two versions value by value.
• Per-page notes. Attach a note to any audited page — a fix to remember, a client request — that autosaves per URL and travels with the page across re-scans.
• Per-version snapshot cards. Copy any point in a page’s history as a clean, shareable card showing that version’s values and the issues that were open at that scan.
• Text zoom. Scale the audit results up or down with A− / A+ and a reset. The setting sticks across sessions.
• Email & phone links. mailto: and tel: links are now surfaced with counts, and you can hover a link to locate it on the page.

Improved

• “Still open” issues in comparisons. The side-by-side comparison card now lists issues present in both scans — not just what changed — for a complete resolved / new / still-open picture.
• Richer link detail. Broken and external links now show their anchor text, and a same-origin redirect encountered on load is flagged.
• Wider platform detection. Joomla and BigCommerce are now recognized in the tech-stack readout.
• More history capacity. Audit history moved to a higher-capacity store, so long histories and large pages no longer run into limits.

Fixed

• Full-page screenshots. Sticky headers and fixed menus (including some page-builder navigation bars) are now hidden during full-page capture, so they no longer repeat down the image.

History, diffs & deeper checks

The audit now remembers. Every check keeps a dated history, you can diff any two scans, and the heavier signals — headings, canonical, structured data — get their own detail views.

What changed

• Audit history, per check. Every section keeps a dated version timeline, so you can see how a page’s title, canonical, schema, and the rest change scan over scan.
• Copy-diff. Compare two scans side by side, with old → new highlighted across title, description, canonical, Open Graph & social tags, and the search snippet.
• Heading outline history. The H1–H6 structure is now an indented outline tree, tracked over time.
• Canonical with indexability. Canonical history records the page’s index / follow state alongside the canonical URL on every scan.
• Live character counts. Title and description show their character count everywhere they appear — in history and in Copy-diff — so you stay inside search-result limits.
• Tidier report. Consistent title-casing across section labels, and long unbroken URLs now wrap cleanly instead of overflowing the panel.

Initial release

Scan any page and get a scored, section-by-section on-page SEO report — without leaving the tab you’re on.

What’s in it

• Eight on-page checks. Title, meta description, canonical, Open Graph & Twitter, H1–H6 structure, internal and external links, images, and JSON-LD schema — each scored and explained.
• Per-page scoring and a per-site board. A 0–100 score and grade for every page, with a home board that lists audited pages lowest-score-first so the worst offenders surface.
• Click to locate. Jump from any link, heading, or image in the report to the element on the live page, tagged with what it is — a heading, list item, or cell.
• Social card preview. Facebook, LinkedIn, and X previews, with og:image dimension and aspect-ratio checks.
• Live link checking. Internal, external, broken, and redirected links, checked from your browser.
• Sitemap reading. Lists the site’s sitemap URLs and marks the ones you’ve already audited.
• Schema validation. JSON-LD entities checked against Google rich-result requirements, with missing required and recommended properties named, and malformed blocks surfaced.
• Export. Download the full audit as CSV — every section’s data, not just the verdicts — or as JSON.
• Side panel mode. A sticky summary that follows the active tab and caches results locally, so revisiting a page is instant.