Release Notes — AuditSpark

August 2026

Aug 18 Improved Sharper findings: your report looks harder before it says something is wrong. A handful of checks were reporting the tool’s own blind spots as faults on your site. Forms are now read field by field, so a recommendation never asks you to change a setting that is already the way it suggests — and where every field is properly labelled, that counts as a pass instead of something to go and verify. Single-page sites are recognised as single-page sites, so their in-page section links are no longer counted as missing pages. Invisible framework markup is no longer flagged as leftover code to delete. Where a page runs longer than we read in one pass, the report says so plainly rather than leaving you wondering whether something on your site is cut off. And your AI Visibility score now reads the same on every page, with the cover stating clearly that it is scored separately from your composite — the composite measures the site itself.

Aug 17 New Agency now includes 100 audits a month, and our prices have changed. The Agency plan goes from 80 to 100 shared audits a month — a quarter more capacity across your whole team, at $349/month. Freelance moves to $99/month, which is the standard price the plan has always been advertised at; the $79 launch price has now ended. If you already subscribe, nothing changes for you. Every existing subscription stays at the price you signed up at, monthly and annual alike, for as long as you keep it — including everyone who joined at the $79 launch price, exactly as promised. Annual billing is still two months free: $990 a year for Freelance, $3,490 for Agency. Pro Report is unchanged at $39.

Aug 17 Improved Three more categories now come with the reasoning behind the score. Content quality, mobile experience, and trust & credibility were being scored on your report without a write-up of their own — you saw the number and a one-line summary, and had to look under a related category for the detail. Each now gets its own findings: what the copy actually does well or poorly, how the site behaves on a phone specifically rather than just how fast it loads, and what your page does to earn a first-time visitor's confidence. Your overall score is calculated exactly as before; there is simply more explanation behind it. Reports already generated are unchanged — this applies to new audits.

Aug 17 Improved The report has a clearer shape: one scorecard, then the detail behind it. Every category we score now sits on a single page with your overall score beside it and a one-line summary of which areas need attention — so you can read the health of the site in a few seconds. The section that follows explains each of those categories in the same order, so a score you want to understand is easy to find, and the prioritized plan at the back still tells you what to tackle first. Reports are also shorter and less repetitive than before.

Aug 12 Improved We changed the analytics tool behind our own website. AuditSpark now uses Umami instead of Plausible to understand how people find and use the site. Nothing about this changes what we collect or how we treat your data — it is still cookieless, still sets no advertising or tracking cookies, still needs no cookie-consent banner, and we still don't sell your data. We've updated our privacy policy to name the new provider, since it's listed there among the services we rely on.

Aug 11 New Change your plan yourself, any time. You can now move between Freelance and Agency, and between monthly and annual billing, from your billing portal — no email to us, no cancel-and-resubscribe. Open your billing portal, choose Change plan, and pick the one you want. The switch takes effect immediately and you're charged or credited only for the difference, so an upgrade mid-month costs the balance rather than a full second month. Your new monthly audit allowance applies right away, and audits you've already run this month still count toward it. If you move up to Agency, your team, client workspaces and white-label settings are set up for you on the spot; if you move down, everything you built is kept exactly as it is — scheduled monitoring and any embedded audit widgets simply pause, and switching back turns them on again with nothing lost.

Aug 10 Improved Older reports now show their overall score on the PDF cover. If you downloaded a report from earlier in the year, the cover's overall score dial was blank — a dash instead of a number — even though the score was right there in your dashboard and every category score in the report was correct. The cover was working the number out from the report text, and reports written before we moved that calculation into our own code phrased it differently enough that the cover couldn't read it. The cover now takes the score straight from your audit record, so it always matches what your dashboard shows. Re-download any affected report and the cover will be complete; nothing else in the report changes.

Aug 7 Improved Competitors no longer get credit for answering a question that named them. Some of the questions we put to AI engines mention rivals on purpose — head-to-head comparisons are part of reading your market. But an engine answering "how does X compare to Y" will always mention X and Y, and that says nothing about whether buyers hear about them unprompted. We were counting those mentions anyway. On one report a competitor sat at the top of the leaderboard with almost a quarter of the share on the strength of a single question that named it, having appeared nowhere else in the audit; it now shows its real share, a fraction of that. The same correction applies to where you place within an answer — you could be ranked behind rivals that only led because our own question named them, and that fed your visibility score. Expect a few competitor shares to fall and yours to rise a little at this measurement; earlier runs keep the figures they were scored with.

Aug 7 Improved A steadier competitor leaderboard, and repeat audits you can actually compare. Two audits of the same unchanged site could return almost entirely different sets of rivals, which made any move in your Share of Voice impossible to read. Two things caused it: we tracked whichever competitors came back first from the analysis rather than the ones engines named most often, so a firm mentioned twice could be dropped while one mentioned once was kept; and every repeat audit invented fresh buyer questions, so it often measured a slightly different market than the audit before. Now we track the competitors engines actually name most, a rival that keeps appearing stays on your leaderboard at 0% rather than quietly dropping off, and a repeat audit re-asks the previous questions for about three months before refreshing them. Expect your tracked competitor set to shift at this measurement — earlier runs keep the figures they were scored with, so a step this week may be the improvement rather than a change in your visibility.

Aug 7 Improved Your AI-readiness score no longer goes missing from the cover. A report whose readiness score landed on a fraction — 63.7 rather than 64 — showed "Not measured" on the cover and a blank row in the scorecard, on the very same report that went on to list the readiness findings in full. The report contradicted itself. Scores now print whatever their precision, and because this is fixed in how reports are drawn rather than how they're scored, it applies to reports you downloaded before today: download again and the score will be there.

Aug 7 Improved More accurate citations, and a Share of Voice that reflects them. Where Gemini was the source, cited links were being recorded as an opaque Google redirect rather than the site actually cited — so those citations couldn't be credited to a real source, and couldn't be recognised when the site cited was your own. We also missed links that appeared inside the body of an AI answer rather than in its source list. Both are fixed, which means your evidence is more complete and your Share of Voice is more honest: an engine finding your own website is discovery of your site, not proof that the market knows you, and it no longer counts as though it were. Expect some Share of Voice figures to come down at this measurement. Earlier runs keep the figures they were scored with, so a step in your trend this week may be the correction rather than a change in your visibility.

Aug 7 Improved Clearer citation reporting, and a steadier competitor list. The "where your citations come from" chart labelled its largest slice "Other", which read like a category of source when it actually means we couldn't identify the site. AI answers draw heavily on one-off blogs and vendor pages that belong to no category, so that slice is usually the majority — it's now labelled Unclassified, and the chart states how much of your citations we could identify so you can weigh the rest accordingly. Related findings no longer claim a source type is absent when most sources went unidentified; they say what was true of the ones we could read. Separately, the competitor list is now carried between runs of the same site even when the address was entered slightly differently — a trailing slash was enough to make a repeat audit rebuild the leaderboard from scratch and show a very different set of rivals.

Aug 7 Improved Two places the report argued with itself. A report could describe mobile performance as "likely acceptable pending measured Core Web Vitals" while printing the measured Core Web Vitals directly beneath it. And a section marked High priority told you to act on it "after the critical work above", even when it was the first section in the report and the urgent work came pages later. Performance verdicts are now written as settled wherever they appear, and priority advice points to your plan rather than to a position on the page.

Aug 6 Improved Two fixes to how AI Visibility is measured. Share of Voice could read far too low for a brand whose homepage title leads with a category phrase rather than its own name — the measurement mistook that phrase for the brand, and set aside most of the questions it should have scored. Separately, Google AI Overviews doesn't answer every question put to it, and a question it declined was being counted as an answer that left you out; those now drop out of the calculation instead of counting against you. Both corrections apply to measurements from today onward — earlier runs keep the figures they were scored with at the time, so a step in your trend this week may be the fix rather than a change on your site.

July 2026

Jul 31 Improved Downloading a report is now near-instant. The first time anyone downloaded a report, the PDF was built from scratch at that moment — several seconds of nothing after the click, with no sign the button had registered. Reports are now prepared the moment an audit finishes, so the download starts straight away. The button also tells you it's working, and can't be double-fired by an impatient second click.

Jul 30 Improved Your report's priorities now agree with its plan. Each section of the report ends with a verdict on where that area stands. Those verdicts were worked out purely from the score, which meant a section could tell you it was a strength worth holding — while the very first action in your 90-day plan was a fix in that same section. Verdicts now report where the work actually sits in your plan: "Fix first" if a P1 action lives there, "Scheduled" if it's P2, "Maintain" only when genuinely nothing in the list touches it. The AI-visibility findings also reach the plan itself now, not just the recommendations table, so the work you're told matters most has dates against it.

Jul 30 Improved The AI findings in your report now carry impact and effort, like every other recommendation. The checks we run on your site for AI visibility — schema, headings, crawler access, entity signals — landed in your prioritized list with blank impact and effort columns. They were the most certain findings in the report, and they looked like the least. Every one now shows what it's worth and what it takes, and each says which check produced it, so on-site fixes and things measured in live AI answers are told apart at a glance. This applies to reports you've already run, not just new ones.

Jul 30 Improved If your page already asks the right questions, the report now says so. Answer-engine readiness was pass-or-fail. A page with a dozen genuine questions on it but no FAQ markup was told it offered nothing for engines to lift, and advised to go and write question-style content it already had. There's now a middle state: when the questions are there and only the markup is missing, the report says exactly that and asks for the markup instead — usually an afternoon's work rather than a content project.

Jul 30 Improved Reports are shorter, and the verbatim AI answers explain themselves. Paragraphs used to stop short of the right margin while others ran full width, which read as a layout fault. Text now runs to one consistent edge, taking a typical report from 22 pages to 19. The section showing real AI answers also says why some of those questions never mention you: that's deliberate — we ask what your buyers ask, including questions naming other brands, then record whether engines bring you up anyway. A question you're absent from is a finding, not a mistake.

Jul 29 Improved Your PDF report has been rebuilt around what each finding actually does for you. The report used to be a list of categories — SEO, accessibility, conversion, and so on — scored one after another. It's now organised around the seven jobs your site has to do in order: be crawlable, get found, become citable, get chosen as the answer, be known across AI engines, convert visitors, and earn trust. Each section leads with the outcome it moves rather than the label, so "Gets you found" comes first and "SEO" sits underneath it. Every score is a dial you can read at a glance, each category says in one line why it scored what it did, and the AI-visibility work that used to live only in your dashboard — which competitors are being recommended ahead of you, where in the buying journey you show up, and the exact answers engines are giving about you — is now in the document you send to clients. Everyone gets this report; there's nothing to switch on.

Jul 29 Improved Annual plans are now simply two months free. Paying yearly used to save you 20%, which is a fine discount but an awkward one to hold in your head — and the yearly price came out at $63 or $199 a month, numbers that don't mean much on their own. Annual billing is now the straightforward version: you pay for ten months and get twelve. Freelance is $790 for the year and Agency is $2,490, saving you $158 and $498 respectively — exactly two months of your plan, which is the whole idea. Switch the pricing page to annual and you'll see what it works out to per month — $66 for Freelance, $208 for Agency — with the full yearly amount you'll actually be charged written underneath. If you're already on an annual plan, nothing changes: you keep the price you signed up at.

Jul 29 New Run out of audits mid-month? You can top up instead of waiting. Until now, using your last audit credit meant waiting until your quota reset — however urgent the work was. You can now add credits to your plan whenever you need them, in packs of 15 or 25, and they're available immediately. Credits are spent automatically once your monthly allowance runs out, so there's nothing to switch on or remember, and they don't expire while your subscription is active. If you hit the limit part-way through starting an audit, you can add credits right there and the audit picks up where you left off — no need to re-enter anything. Your monthly allowance is always used first, so you'll never spend a credit you've paid for while you still have included audits left.

Jul 28 Improved More reports now include real measured performance. Every audit tries to measure your site's actual loading speed — the Core Web Vitals Google grades you on — rather than estimating it from your code. When that measurement didn't come back in time, the report simply said it was unavailable, and that was happening to about one in five audits. Those failures turned out to be mostly bad timing rather than anything wrong with the site: the same site would measure perfectly on another run. Audits now wait a little longer, space out their measurement attempts, and retry once when a measurement fails for a transient reason. Sites that had never managed to return performance data are now returning it.

Jul 27 Improved Your AI-visibility fixes now appear in the prioritized recommendations. The AI & GEO readiness section has always listed what's stopping AI answer engines from finding and citing you — but those fixes sat on their own, with no sense of how they ranked against everything else in the report. They're now folded into the prioritized list alongside the rest: anything that blocks citation outright, like an answer engine's crawler being turned away, sits in the top tier, and the rest land one tier below. Each of those rows is marked so you can see it came from the readiness check rather than the written analysis. The readiness list itself is now colour-coded by severity too, so the urgent items read as urgent at a glance.

Jul 23 Improved More accurate competitor detection in your AI Visibility report. Share of Voice is only as trustworthy as the competitor list behind it — so AuditSpark now identifies your real competitors more reliably, filtering out generic placeholders and any name it can't confirm is a real business in your market. When it can't name real competitors, your report now says so plainly instead of showing an inflated score. Some AI Visibility scores will shift as a result — usually downward where the old number leaned on a weak list, and truer for it.

Jul 19 New A guided tour of what your audit measures, right on the homepage. Under the search box, the homepage now steps through the different kinds of visibility an audit covers — classic search, technical foundations, trust and authority, and how AI answer engines find and cite you — each paired with a sample of the report card you'd get. It's a quick way to see the full scope of what AuditSpark checks before you run your first scan.

Jul 18 New Name your agency. When you signed up we generated a placeholder agency name from your email address — so an account created with jane@studio.com showed up as "jane", including on the invitations you sent your teammates. You can now set a proper agency name under Settings → Branding, and it takes effect on team invites straight away. It's kept separate from the company name printed on your white-label report covers, so your internal name and the brand your clients see can differ.

Jul 17 New Your free report now grades answer-readiness — whether you're built to be chosen as the answer. Being found by AI engines isn't the same as being the answer they pick. The AI/GEO Readiness check now grades whether your page is structured to be lifted straight into a featured snippet, a People-Also-Ask box, or a voice answer: question-shaped headings with short, direct answers, the FAQ and HowTo schema answer engines actually pull from, and lists and tables they can quote cleanly. It also names which schema types you already have versus the ones that feed AI answers — so instead of a blanket "schema present," you see exactly which answer types you're missing. Like the rest of Readiness, it's on every report, including free scans, with specific do-this-next fixes.

Jul 15 New Google AI Overviews join your AI Visibility report. AuditSpark now measures how Google's AI Overviews — the AI-written answer that appears at the very top of Google Search — represent your brand: whether they mention, cite, or recommend you, and which sources they pull from, alongside ChatGPT, Claude, Gemini, Microsoft Copilot, and Perplexity. Because it's the AI answer millions of people see first, before any of the blue links, it's a direct read on how you show up at the top of search. It appears automatically as its own column in your per-engine scorecards and the Evidence Drawer.

Jul 15 Improved The free AI/GEO Readiness check now catches the crawlers that gate AI citations. AI engines use one set of crawlers to train their models and a different set to cite live web results — and blocking the citation crawler can make you invisible in an engine's answers even when everything else looks right. Readiness now checks both kinds across ChatGPT, Claude, Perplexity, Gemini, and Microsoft Copilot — including the Bing crawler that Copilot cites from — and tells you which is which: a blocked citation crawler is flagged as an urgent fix, while opting out of model training is noted as the legitimate privacy choice it is, not a failure. Because it now inspects more signals than before, some readiness scores will shift slightly; your report marks when a change comes from the improved check rather than from your site.

Jul 14 Improved A completely redesigned report PDF. Every downloadable report — the Pro Report, and every audit on Freelance and Agency plans — has been rebuilt into a polished, presentation-ready document. It opens with a cover that shows your overall score at a glance, then walks through visual score bars for each category, measured performance, AI-visibility charts, and a priority-ranked action plan. Every section now starts with a short, plain-language note on what it covers, and pages are numbered throughout. White-label reports get the same design in your own brand colour.

Jul 14 New Perplexity joins your AI Visibility report. AuditSpark now measures how Perplexity answers about your brand — whether it mentions, cites, or recommends you, and which sources it pulls from — alongside ChatGPT, Claude, Gemini, and Microsoft Copilot. As the AI answer engine built specifically for research and recommendations, it's a direct read on how you show up when people compare options in your category. It appears automatically as its own column in your per-engine scorecards and the Evidence Drawer.

Jul 13 New Microsoft Copilot joins your AI Visibility report. AuditSpark now measures how Microsoft Copilot answers about your brand — whether it mentions, cites, or recommends you, and which sources it pulls from — alongside ChatGPT, Claude, and Gemini. Because Copilot answers from live web results, it's a direct read on how one of the most widely used AI assistants represents you to people researching your category. It appears automatically as its own column in your per-engine scorecards and the Evidence Drawer.

Jul 11 New A look inside AuditSpark, right on the homepage. The homepage now features a rotating showcase of real product screens — the AI Visibility leaderboard, the per-engine scorecards, the verbatim Evidence Drawer, and a full Pro Report — presented on a set of gently turning phones under the search box. It's a quick way to see exactly what you get before you run your first audit.

Jul 10 Improved A refreshed, custom icon set across the whole site. AuditSpark now uses one consistent, custom-designed icon family everywhere — from the homepage and pricing to your dashboard — replacing the previous mix of icons and emoji. A small change on its own, but it makes the whole product feel more polished and cohesive, with a recurring teal "spark" as the through-line.

Jul 10 Improved Cleaner handling of paused agency accounts. If an Agency subscription lapses, its embedded lead-gen widgets now stop running new audits and the dashboard switches to a read-only view. Your clients, reports, monitoring history, and team stay fully visible — nothing is deleted — and everything resumes when you reactivate (widgets you re-enable yourself, so a paused site can't quietly keep generating audits in the background).

Jul 6 Improved Dedicated pages for every core feature. Scheduled monitoring, white-label reports, and the lead-gen widget each now have their own page walking through exactly how they work and what's included — so it's easier to see what AuditSpark does before you sign up, and easier to share a specific capability with a colleague or client.

Jul 6 New Automated monitoring history for your clients. Each monitored client now has its own run history — a timeline of every automated audit with its score trend and the "What's Changed" brief for each run. Open it from the client card to see how a site has moved over time, or jump straight to all audits for that client. The narrative brief that used to sit on the card now lives here, keeping cards tidy while giving the monitoring story a proper home.

Jul 4 New See what the AI engines actually said — the Evidence Drawer. Your AI Visibility tab now opens the verbatim answers behind your scores — exactly how ChatGPT, Claude, Gemini, and Perplexity described your brand, whether they mentioned, cited, or recommended you, and which sources they pulled from. Filter to the answers that named you, the ones that missed you, or any with factual errors.

Jul 3 New Answers to your most common questions. The homepage now has a Frequently Asked Questions section covering what's included free versus the Pro Report, how accurate the audit is, how your data is handled, and our money-back guarantee.

Jul 3 Improved A clearer homepage. Added a plain-language explanation of AI visibility — whether ChatGPT, Claude, Perplexity, and Gemini can find and cite your site — and made the 7-day money-back guarantee easy to spot right where you start an audit.

Jul 3 Improved Live progress while your audit runs. The loading screen now shows a real progress bar and step-by-step updates as each section of your report is written — so you can see exactly how far along it is instead of waiting on a spinner.

Jul 1 New AI Visibility dashboard. See how ChatGPT, Gemini, Perplexity, and Claude describe your brand versus competitors — with a share-of-voice breakdown and per-engine scores.

Jul 1 Improved Agency client workspaces are now restorable. Deleted a client by mistake? Removed clients can be brought back on request.

June 2026

Jun 30 New AI Visibility now checks OpenAI, Gemini, and Perplexity in addition to Claude — same audit credit, no extra charge.

Jun 30 Improved Bigger plan quotas. Freelance now includes 25 audits/month; Agency now includes 80/month, shared across your team.

Jun 30 Improved Redesigned the dashboard navigation for quicker access to Clients, Team, and AI Visibility.

Jun 29 Improved Refreshed the entire site with a cleaner look — new colors, typography, and layout.

Jun 23 New Agency Scheduled Monitoring. Put a client's site on a monthly or weekly check-in and get a "What's Changed" brief — with the full PDF — emailed automatically.

Jun 22 Improved More reliable performance scoring — we now average several test runs so one slow network blip can't throw off your Measured Performance numbers.

Jun 18 New Reports now include real Google PageSpeed / Core Web Vitals data — a Measured Performance section alongside the AI analysis.

Jun 11 Improved Your dashboard now shows every audit — web, API, and widget — in one sortable table.

Jun 10 New Launched the AuditSpark MCP server and free API keys, so you can run audits straight from Claude or your own tools.

Jun 9 New Shareable report pages. Every audit gets its own link you can send to a client or teammate.

Jun 8 New AI Visibility (GEO). A free AI/GEO Readiness score now appears on every report, plus an optional deeper add-on measuring whether AI engines actually recommend your brand.

Jun 8 New Added a Privacy Policy and Terms of Use, with one-click controls to make a shared report private or delete it.

Jun 5 Launch SiteIQ is now AuditSpark. Same product, new name, and a new home at auditspark.io.

Jun 5 Improved The Consultant plan is now called Freelance — $79/month, same features.

Jun 4 New Agency white-label PDF branding, clearer report filenames, and an emailed report link with every Pro Report purchase.

Jun 4 New Added a Refund Policy page.

Jun 3 Improved Switched to cookieless site analytics — no cookie banner needed, nothing personally identifying tracked.

Jun 3 New Agency Clients & Portfolio. Organize audits under named client workspaces and see rollup performance stats.

Jun 2 Improved Better handling of JavaScript-heavy sites that previously came back unreadable.

Jun 1 Improved Agency team seats raised from 3 to 10.

May 2026

May 30 New Agency Plan launched — team invites, client workspaces, white-label PDF, a portfolio dashboard, and an embeddable lead-gen widget.

May 28 Launch AuditSpark (then SiteIQ) launches — the first AI-powered website audit.

What's new in AuditSpark.

August 2026

July 2026

June 2026

May 2026

See where your website stands today.