Changelog
What we have shipped to SiteCurl, newest first. Every entry is live in the product today.
-
Improved
Clearer, easier-to-navigate reports
Reports open with a plain summary of what to fix first, and a category list lets you jump to or filter by one area. Failures you have accepted as a risk are counted separately and no longer read as open issues.
-
Improved
Portfolio health explanation now works on phones
Tap the help icon on the dashboard to see how the portfolio score is calculated, and tap anywhere else or press Escape to close it. The counts beside the score are also larger, easier tap targets.
-
Fixed
SSL expiry warns only when renewal has failed
Auto-renewing certificates no longer warn 30 days out. You now get a warning at 14 days left and a critical at 3 days, and a scan that cannot reach your certificate says it could not verify instead of staying silent.
-
Fixed
Sitemaps listed in robots.txt are now recognized
If your sitemap lives somewhere other than /sitemap.xml and robots.txt points to it, SiteCurl now finds it instead of reporting it missing.
-
Fixed
Sitemap and llms.txt checks no longer pass on a fake page
Sites that answer every URL with their homepage used to look like they had a sitemap and an llms.txt. SiteCurl now reads the file itself and only counts a real sitemap or a non-HTML llms.txt.
-
Fixed
AI crawler check now separates training from search
Blocking a training crawler like GPTBot or Google-Extended is now reported as info instead of a warning. Blocking a search crawler like OAI-SearchBot or PerplexityBot still warns. We also check ChatGPT-User, Claude-SearchBot, Claude-User and Perplexity-User.
-
Fixed
A missing robots.txt is no longer a warning
A 404 on robots.txt now reads as optional info, since search engines simply crawl everything. A server error (5xx) on robots.txt is now critical because Google pauses crawling your site, and a firewall bot challenge reads as could not verify.
-
Improved
Free audit page shows a sample report beside the form
The free website audit page now previews a sample report next to the URL field, and says how many checks the free scan runs.
-
Improved
Dashboard shows what to fix first, with clearer counts
Your priority inbox now leads the dashboard, with a smaller health gauge beside it. The counts say exactly what they mean (90+ with no critical issues, or not yet 90+ or with a critical issue), and the score change shows the date it started from.
-
Fixed
Speed score now says it is a lab score
A good speed result is now labelled a lab performance score instead of claiming your Core Web Vitals are healthy. Core Web Vitals are measured from real visitors, which a lab test cannot show. We also stopped listing INP, which a lab test cannot measure.
-
New
New API endpoint for lead-capture forms
On a paid Studio plan, POST a URL and an email address to the new /api/v1/lead_scans endpoint to scan it and send a whitelabeled top-5-issues report to that address, no site added to your dashboard. Built for a lead-capture form on your own site, and it works with any form tool that can send an authenticated request. If a scan can't be completed, you get an email with the visitor's address so you can follow up. Set the report's call-to-action link under Settings, Report branding.
-
Improved
Critical regression alerts now catch warning-level issues too
The critical regression email now fires when a new warning-severity issue appears, not just critical ones. Each issue in the email shows its real severity.
-
New
Send scheduled reports straight to your clients
On a paid Studio plan, add up to 3 client emails to a site and each scheduled scan sends them a summary email linking to the report, in your white-label branding if it is on. Clients can unsubscribe in one click, which stops reports for every site in your account.
-
Improved
Free scans show progress instead of a blank wait
The free site scan and the AI-readability check now open their result page right away with a scanning status, and the results appear on that page when they are ready.
-
Improved
Platform audit pages and Pricing easier to reach
The site footer now links the Shopify, WordPress, Webflow, Squarespace, and Wix audit pages. Pricing links in the menu and footer open the full pricing page instead of jumping down the homepage.
-
New
AI agents can find and learn the SiteCurl API
Coding agents that check a site for agent support can now discover the SiteCurl API, read how to get a key, and load a ready-made skill for running an audit and reporting the fixes.
-
Fixed
Pricing and About pages display cleanly when signed in
Signed in, the Pricing, About, and Contact pages drew the site menu over the app sidebar. They now show as the full page with a Dashboard link back into the app, and one main content area for screen readers.
-
Fixed
One skip link on shared scorecards
Shared scorecard pages offered keyboard and screen reader users two identical skip links. They now have one, pointing at a single main content area.
-
Fixed
Every page now has its own browser tab title
Signed-in pages used to all be titled SiteCurl. Each now names itself, and site pages name the site, so tabs are easy to tell apart and screen readers announce where you are.
-
Improved
Accepting a risk updates your report in place
Accepting, editing, or undoing an accepted risk no longer reloads the report. A confirmation with Undo appears right at the finding, the summary updates, and the sections you had open stay open.
-
Fixed
Checks we could not measure no longer show as fixed
When a CDN or firewall stops us measuring a check on one scan, it is no longer counted as fixed, and the next scan no longer reports a false regression. Reports, comparisons and alerts now compare against the last scan that did measure it, so real fixes and regressions still show.
-
Fixed
Pausing a site now stops its next scheduled scan
A scheduled scan could still run up to a few hours after you paused a site, switched it to manual, or scanned it yourself. The scan now checks again right before it starts and skips in each of those cases.
-
Fixed
Inactive account emails show the real deletion date
The warning sent to inactive free accounts now names one deletion date in its subject and body, and the account is scheduled for deletion on that date. The email now says the account was flagged for having no scans, not for missing logins. Cancelling a scheduled deletion now keeps the account instead of it being flagged again the next day.
-
Fixed
Pricing cards read correctly with a screen reader
Screen readers read the features a plan leaves out, like white-label reports on Pro, as if they were included. Screen readers now say "Not included" before each of those rows.
-
Fixed
Accurate Starter plan details in emails and help pages
Some emails, billing notices, and FAQs said scheduled scans stop on Starter, that scanning more than one page needs a paid plan, or that score alerts are paid-only. Starter keeps one site that can scan monthly, checks up to 5 pages per scan, and gets those alerts. The copy now says so.
-
Fixed
Sites added through MCP now get scheduled scans
Adding a site with the MCP sitecurl_add_site tool now schedules weekly scans on active Pro and Studio plans instead of leaving the site unscheduled. You can also pass any cadence your plan includes when you add it.
-
New
Get an email when your site can't be reached
If a scheduled scan can't reach your site (DNS, connection, timeout or SSL failure) and it still can't 15 minutes later, we email you once per outage. Turn it off under Site unreachable alerts in Settings. Webhooks get a new site_unreachable event.
-
Fixed
Webhook settings save while your receiver is down
Changing a webhook's events through the API failed if your receiving server's domain briefly could not be looked up. Only a change to the webhook address now checks that address.
-
Fixed
Clear reason when a scan cannot find your domain
A scan of a domain that could not be looked up, or that points to a private network, now says so and what to check, instead of a generic failure message.
-
Fixed
Finished scans are no longer marked as failed
A brief network hiccup right after a scan finished could mark that scan as failed and skip its report. The scan now keeps its result and the report is generated as usual.
-
Fixed
Site pages open again late in a free trial
Some trial sites showed an error page when their next scheduled scan fell after the trial ended. The page now opens and simply leaves out the next scan date.
-
Fixed
Pages no longer show as plain text in the browser
An AI tool asking for the text version of a page could make that page appear as raw text to other visitors for up to two hours. The text version is no longer cached for browsers.
-
Fixed
New API keys and sign-in errors show up again
Creating an API key now shows the new key so you can copy it, and a mistyped password now tells you so. Both messages were being cleared before the page appeared.
-
Improved
Free monthly scans are easier to find
Your free plan includes one automated scan a month. If a site is set to manual, its page now offers to turn monthly on. You pick the schedule, and you can dismiss the suggestion.
-
Fixed
Emails no longer repeat themselves
Report, alert, trial and billing emails arrived with their whole contents printed twice, one after the other, in a single message. Each email now shows its contents once.
-
New
Monthly automated scans on the free plan
Free plan sites were only ever scanned when you asked, so a score drop could go unnoticed until you checked. You can now set a site to scan automatically once a month, which is what makes score drop alerts arrive on their own. Choose Monthly on the site settings form. Weekly scans stay on Pro and Studio.
-
Fixed
Clearer answer when a scan's detail is unavailable
The API and MCP tools returned an empty finding list for older scans whose detail is no longer stored, which could read as a clean result, and such a site could drop out of the priority list entirely. Both now say the detail is unavailable and whether it aged out or was lost.
-
Fixed
One scheduled report per scan, not several
A scheduled scan could be queued more than once in the same cycle, so some sites were scanned several times and sent a report email for each one. Each cycle now runs a single scan and sends a single report.
-
Improved
Critical issue alerts on every plan
Critical issue alerts were limited to paid plans, so most sites never got told when a scan turned up a new critical problem. Every plan now receives them, and you can turn them on or off under Settings, Preferences. Scheduled scan reports stay on Pro and Studio.
-
Improved
Reports now lead with what changed
When you rescan a site, the report opens with what you fixed and what regressed since your last scan, instead of repeating the same summary every time. The PDF export carries the same line, so a report you forward matches the page you exported it from.
-
Fixed
New sites now get their scheduled rescan
A site created from a free audit or during signup was set to manual scanning, so it was scanned once and never again. Once you verify your email, those sites are scheduled automatically on a trial or paid plan, so you get the follow-up scan and can see what changed since the first one. You can still change the frequency on the site page.
-
Fixed
Checks we could not measure no longer look like passes
When we could not reach your server for a check, the report showed a green checkmark next to text saying the result could not be determined. Those checks now read as not measured in the report, the PDF, and the email, so you can tell a real pass from one we could not complete.
-
Fixed
Accurate HTTP/2 and HTTP/3 results on redirecting sites
If your site redirects between its plain domain and the www version, the HTTP/2 check tested one address but reported the other one's HTTP/3 support, so the result could be wrong in either direction. It now checks and reports the same address your visitors land on, and says when HTTP/3 is advertised rather than confirmed.
-
Fixed
Older reports no longer say a site is problem free
Some audits from before July lost their item by item detail, and their reports said no issues were found even when the original scan had failures. Those reports now show the real score and issue counts, and say plainly that the detail is unavailable.
-
Fixed
Password managers fill sign in forms again
Your password manager did not offer to fill the sign in, sign up, or password reset forms unless you reloaded the page first. Reaching these pages from anywhere in the site now works the first time, and the same fix covers the password fields in your account settings.
-
Improved
Better first fixes in your AI readiness report
The top three AI readiness fixes now lead with the cheapest changes that make the biggest difference, like heading structure and structured data. Publishing an llms.txt file used to come first for most sites, but it is an optional convention with no clear effect on how AI tools read you, so it now ranks last.
-
Fixed
Billing changes no longer report false success
Scheduling an account deletion, or cancelling one, could report success even when the billing change had not gone through. These now tell you when something went wrong and leave your account unchanged so you can retry.
-
Fixed
Annual plans stay annual when you downgrade
Downgrading to a lower plan while on annual billing moved you onto that plan's monthly price. Annual customers now keep annual pricing through a downgrade.
-
Fixed
Downgrade page no longer errors on some plans
Opening the downgrade confirmation page could show an error page instead of your billing details in rare cases. It now loads correctly or explains what went wrong.
-
Fixed
MCP docs now list all 12 tools
The MCP server has a portfolio-wide finding search, but the docs and tool tables still listed 11 tools and left it out. Every surface now lists all 12. The accessibility checker page also lists its 15th check.
-
Fixed
Page limits now shown per plan
Marketing pages, help answers and signup emails described the Studio page limit as if every plan had it. Each now names the plan it applies to, and your welcome and verification emails show the limit for your own account.
-
Improved
Faster answer when a site cannot be reached
A scan of a site that does not respond now gives up sooner instead of waiting the full timeout. Sites that are reachable but slow, including ones on unreliable networks, are still scanned in full.
-
Improved
Safer email preference links
Preference links now stop working once you change your email address, so a link sent to an old address cannot be used later. Every unsubscribe is recorded, which means we can tell you what happened if reports stop arriving unexpectedly.
-
Fixed
Check pages now match what the checks do
Fourteen check pages described coverage the checks do not have, such as scanning picture tags for broken images or requiring a minimum title length. Each description now reflects what the check actually does.
-
Fixed
Accurate descriptions of four check pages
The about page, privacy policy, terms and favicon check pages said we confirm the target returns a 200 status code. We do not: these checks read your HTML and never request the linked file. The descriptions now say exactly what each check does.
-
Improved
Unsubscribe from any recurring email
Report emails, score alerts, trial reminders and account warnings now carry a one-click unsubscribe and a link to your email preferences. Each person on an account gets their own copy, so unsubscribing stops your mail without affecting your teammates.
-
Improved
A way to reach us when something is wrong
Emails about account deletion, a data export, paused sites or a failed payment now offer a direct way to contact us, instead of only a small link in the footer. The welcome email also gets to the point faster.
-
Fixed
Report emails no longer contradict themselves
A report showing a score drop could say All checks passed directly above a list of regressions. It now only says that when nothing changed. Category tags in the What Changed section also count new issues, so a category no longer appears with no numbers beside it.
-
Improved
Clearer transactional emails
Buttons in our emails now render correctly in Outlook, onboarding emails name the next step instead of saying Go to Dashboard, and the password reset email tells you what to do if you did not request it.
-
Fixed
Emails now state retention accurately
Upgrade and trial emails said report retention was unlimited. Your score history is kept permanently, and full finding detail is kept for 365 days on Pro and Studio. The emails, your billing page, and the homepage now say exactly that.
-
Fixed
Expired certificates now show up as a finding
Scanning a site whose SSL certificate has expired used to return a generic error. You now get the certificate flagged as critical, with the steps to renew it.
-
Fixed
No score when we could not read your page
A site returning a 404 or 500 could score higher than a healthy one, because most checks never ran. We now show the failure and hold back the score.
-
Improved
Run a free scan from any audit page
The Shopify, WordPress and Webflow audit pages now take your URL directly, instead of sending you somewhere else to start.
-
Improved
See how we keep check results accurate
Reports and our FAQ now explain how to check a finding for yourself, how our checks are tested against real pages, and where we publish corrections when a check turns out to be wrong. Useful when you need to stand behind a report in front of a client.
-
Improved
Sign in without opening the menu on mobile
The sign in link on our marketing pages was hidden behind the menu button on phones, so returning customers had to go looking for it. It now sits in the top bar next to the trial button. Site names on your sites list also get more room on a phone.
-
Fixed
Buttons and links are easier to tap on phones
Small controls, including the edit button on your sites list, were too small to tap reliably on a phone. They now meet a 44px minimum, and neighbouring links no longer overlap, so a tap lands on the one you aimed at.
-
Fixed
Site cards no longer stretch off-screen on phones
On the dashboard, a site with a long name could stretch its card past the edge of a phone screen, and because the cards share a row, every card on that row stretched with it. Cards now stay within the screen and long names are shortened with an ellipsis.
-
New
AI agents can search your whole portfolio
The SiteCurl MCP server gained a tool that searches open findings across every site you manage in one call, grouped by site. An agent working through a client list no longer has to check each site one at a time.
-
Fixed
Your first scan now shows as running right away
Signing up used to land you on a page offering to run your first scan, directly under a message saying one was already starting. The scan now appears as running from the moment you arrive. A scan that never leaves the queue also explains itself accurately instead of reporting a worker crash.
-
New
Search every site you manage at once
Press Cmd+K and type a problem, like alt text or HSTS, to see which of your sites have it and how many pages are affected. Finding one issue across a whole client list used to mean opening every report.
-
Improved
Top fixes now show effort and a combined result
Each recommended fix is marked when it is a quick one, and the panel tells you the score you land on if you do all of them. You can see which fixes are cheap before you start, instead of finding out partway through.
-
New
See everything you have fixed on a site
Each site page now shows how many issues you have resolved since your first scan, and how the score moved. It is a record of the work you have done, not just what is still open, so you can show a client what changed on their site.
-
New
Your report opens with a plain summary
Every report now leads with a short paragraph saying what was checked, what is open, and where to start. It is written to be forwarded, so you can send the report to a client without writing a cover note first.
-
New
Every report shows how the score is calculated
How the score works used to appear only on our example report. It now sits on every report and every shared link, so anyone reviewing your results can see the method behind the number.
-
New
Scope and method on every report
Reports now close with the scan date, the site, how many checks ran across how many pages, and which categories were assessed. It prints with the report, so a copy handed to a client carries its own context.
-
Fixed
A clearer, more honest score scale on your report
The band bar under your score now shows each grade at its true size, so you can see that Critical covers half the scale and Excellent only the top tenth. The active band is labelled directly beneath itself, and the headline verdict no longer reads a reassuring word in a warning colour.
-
Fixed
Every passing check now appears in your report
The passed filter counted checks it never showed you. Categories that had something to fix listed only the problems and hid the checks that passed alongside them. Each category now has a short expander for its passing checks, so the number in the filter matches what you can actually read.
-
Fixed
Exports and the web report now say the same thing
A scan with an issue to fix could read one way on screen and another in the PDF or Markdown export you sent to a client. All three now lead with the same verdict, so the report you forward matches the report you reviewed.
-
Fixed
The sample report shows every fix
Our example report listed a fix at the top of the page and then asked you to sign up to see that same fix further down. All the remediation detail on the sample report is now open to read.
-
Fixed
Accurate docs on how your score is calculated
Our help pages said every check counted the same toward your score. It does not: failures are weighted by severity, so a critical moves your score about five times as much as a passing check. The docs now explain the real calculation. The plan comparison also showed Studio at 50 pages per scan instead of 100.
-
New
Short-lived access tokens for automated clients
Scripts and agents can now trade an API key for a one-hour access token at POST /api/v1/tokens. Keep your long-lived key in a secret manager and put only an expiring token on the wire. Revoking a key immediately invalidates the tokens it issued.
-
Fixed
Clearer pricing on what history you keep
The pricing page said unlimited history, which was not precise. Your score history is kept permanently, and full finding detail is kept for 365 days on Pro and Studio. The page now says exactly that.
-
Improved
Adding a site starts its first scan
Adding a site used to leave you on a page telling you to go run a scan. Now the scan starts as soon as the site is added, so your first results arrive without another click.
-
Improved
Trials now include a second automatic scan
Your trial re-scans your site after about three days, so you can see what changed, compare two scans, and get score alerts before the trial ends. Those need two scans to work, and a weekly schedule only fit one.
-
Fixed
Pricing page now shows the free plan
Starter was missing from the pricing page even though every account starts there. You can now see what the free plan includes, 1 site and 5 pages per scan, before you sign up.
-
Fixed
Pro plan correctly listed with API access
The pricing page showed the REST API, MCP server, and webhooks as Studio only. Pro has always included them, with Studio offering higher rate limits.
-
Improved
Starter reports explain the page limit
If a scan stopped at your plan's page limit, only Pro accounts were told why. Starter reports now say how many pages were covered and what a higher plan scans.
-
Fixed
Clearer wording on when we detect changes
Our site description implied we watch your site continuously. We compare each scan against the one before it, so the wording now says that. What the scans check has not changed.
-
Fixed
Issue counts now match between your site and report
Your site page counted every page a problem appeared on, while the report counted the problem once. The same scan could read 100 issues in one place and 1 in the other. Both now count the same way.
-
Fixed
Rechecked every competitor comparison claim
We reviewed all 74 feature claims across our comparison pages against each vendor's own documentation. Sixteen were out of date and are now corrected, and we removed price estimates we could not source.
-
Fixed
Corrected competitor comparison details
Our Screaming Frog and GTmetrix comparison pages had outdated pricing and understated what Screaming Frog covers. Both pages now reflect what those tools actually do today.
-
Improved
See how your score changed at a glance
Your site page now shows the score from your previous scan next to the current one, with the date it came from, so you can see what moved without opening a comparison.
-
Improved
Finished scans now hand off to your results
When a scan completes, the progress card finishes and clears instead of disappearing, so your score arrives as part of one smooth transition. Cancelled and failed scans skip it.
-
Improved
Your report score now animates as it loads
The health score ring fills to your score when a report opens, and the category breakdown follows just behind it. The number itself stays readable the whole time.
-
Fixed
PDF reports no longer fail on non-Latin text
Downloading a PDF report failed if your site contained Greek, Japanese, or other characters outside the Latin alphabet. Those reports now download, with a note where a character could not be shown in PDF.
-
Fixed
Trial emails show the right page limit
The trial expiry email said scans drop from 50 pages to 5. Trials actually include 100 pages per scan, so the email now shows the correct numbers.
-
Fixed
Pricing page now shows AI readiness on Pro
The pricing table listed the AI readiness report as a Studio feature, but Pro plans have always included it. The Pro column now shows it correctly.
-
Fixed
Account deletion completes reliably
Deleting an account could stall partway through if anyone on it had previously accepted a risk on a scan. Deletions now finish cleanly and the accepted risk history stays intact on the site.
-
Fixed
Contact messages deliver more reliably
Automated spam submissions with invalid email addresses could block the contact form from delivering. Messages from real visitors now reach us without interruption.
-
Fixed
Password managers work on signup and reset
Signing up, resetting a password, or sending a contact message could silently fail if your password manager filled a hidden spam-protection field. Those forms now work with 1Password and similar tools.
-
Improved
Easier tapping on phones and tablets
Category rows in your report, the site link in the report header, and the footer links were slightly too small to tap comfortably on a phone. They now meet the 44px minimum, so they are easier to hit on the first try.
-
Fixed
Broken external links now checked on every plan
The check for broken links to other sites could run out of time on the Starter plan and show a timeout notice instead of results. It now finishes within its time budget, so you get the real findings.
-
Fixed
Fairer content depth scoring for longer pages
The content depth check no longer penalizes longer, well-written pages for their length. It now judges how varied your writing is in a way that is fair whether a page is short or long, so a rich page is not mistaken for filler.
-
Fixed
Fewer false alarms in your scan results
Your scan no longer flags things that are actually fine: pages with a contact or privacy link at a less common address, location pages that share a similar opening line, retina and vector images sized correctly for the screen. Real issues are still caught.
-
Improved
More accurate security checks
Your security results are now more precise. We flag header values that look protective but are not (like a frame setting that still allows embedding), catch insecure images and forms in more places on the page, and no longer warn about protection that is actually in place on sites behind a CDN.
-
Improved
Share your AI-readability result with a link
Your result from the "Can AI read my site?" check now has its own link, so you can copy it and send it to a teammate or client. The link works for 24 hours.
-
Fixed
Password managers now fill the login form
A hidden anti-spam field on the sign-in page was named in a way that made password managers like 1Password fill it instead of your email and password. Autofill now lands in the right fields.
-
Improved
Fairer AI readiness grades
Your AI readiness grade now reflects how much of your content AI crawlers can actually read, so well-built pages get the credit they earn. Minimal pages are described plainly instead of flagged as a problem.
-
Improved
AI readiness now checks for llms.txt
Your AI readiness results now tell you whether your site has an llms.txt file, an emerging way to give AI tools a curated map of your key content, plus how to add one if you want it.
-
New
Free check: can AI read your website?
See in seconds whether AI crawlers like GPTBot and ClaudeBot can actually read your homepage, and get the specific fixes if they cannot. No signup.
-
Improved
Clearer, easier-to-spot links
Links across your dashboard, reports, and account pages now stand out clearly, so it is obvious what you can click at a glance.
-
Improved
Simplified the report page
We removed the embeddable score badge. It was not being used, and taking it out keeps the report page focused on the findings and fixes that matter.
-
Improved
Honest Core Web Vitals reporting
When a live speed measurement is not available for a scan, the performance section now says so plainly instead of quietly showing nothing, and it never penalizes your score for a measurement we could not take. Reports also now include INP (Interaction to Next Paint), part of Google's Core Web Vitals.
-
Fixed
LinkedIn links no longer flagged as broken
The external link check no longer reports LinkedIn (and similar) profile links as broken. Those sites return a non-standard status to automated checkers even though the page loads fine for visitors, so we now treat it as blocked rather than dead.
-
Improved
Clearer priority inbox on mobile
Each priority inbox row now leads with the site name and a compact status chip, so you can scan which site needs attention at a glance. On phones the action button drops onto its own line, so long site names stay readable.
-
Fixed
Priority inbox now matches your score
The dashboard priority inbox no longer flags a high-scoring site as having critical or quick-fix issues based on AI-readiness findings, which do not count toward your score. It also caps very large counts and links to the full list when there is more than fits.
-
Improved
Faster site pages
Individual site pages now load with far fewer database queries, so the scan history and trend chart appear noticeably quicker.
-
Fixed
Annual pricing toggle now shows which option is selected
On the plan chooser, switching between Monthly and Annual now clearly highlights the selected option, not just the prices.
-
Improved
A clear upgrade recap when you view your first report
After your first report during a trial, we recap what your scan covered and make it easy to keep Studio access before your trial ends.
-
New
Your trial countdown now follows you everywhere
A compact trial pill in the sidebar shows how many days remain and links straight to upgrade, on every page.
-
Fixed
Every new signup now gets a welcome email
Your welcome email and onboarding tips now send from the moment you sign up, even before you verify, so no one misses the getting-started sequence.
-
Improved
A welcome note when your trial starts
Sign up and you now see a clear confirmation that your 7-day Studio trial is live, so you know exactly where you stand from the first click.
-
New
A grade and a shareable card for your audit
Your audit scorecard now shows a letter grade and a branded card you can save and share, so passing your results along looks the part.
-
New
Share your audit scorecard
Run a free audit and share the scorecard with a colleague or client. They see the score and category breakdown, and can audit their own site in one click.
-
Improved
A warmer welcome after you upgrade
The page you land on after upgrading now names your new plan and what it unlocks, with a small celebration to mark the moment.
-
Fixed
Clearer messages when a scan cannot finish
Failed scans in your scan history now show a plain, readable explanation instead of raw technical error text, so it is easier to tell what happened.
-
Improved
Clear notice when a scan does not finish
If your most recent scan fails to complete, the site page now shows a friendly notice with your previous results still in view and a one-click retry, so you are never left guessing about a stalled scan.
-
New
Copy any fix as a prompt for your AI agent
Every finding in a report now has a "Copy fix for your AI agent" button. It copies a ready-to-paste prompt with the problem, the change to make, the affected pages, and a reference link, so you can hand the fix straight to Cursor, Claude Code, or Copilot.
-
Improved
AI readiness, grouped with a per-site badge
AI readiness checks are grouped into clear categories so you can see at a glance what an AI crawler can and cannot read on your site, and every site in your dashboard shows an AI readiness badge.
-
Improved
Side-by-side scan comparison
Compare any two scans of the same site on one page. It leads with what changed: the score movement, new and resolved issues, and category shifts, with the full check table below for detail.
-
Improved
Clearer fix guidance with copyable code
Each finding's fix steps now read as clear numbered instructions with copyable code snippets, so you can apply the fix without hunting for the syntax.
-
Fixed
More reliable report email delivery
Scheduled report emails now retry automatically when a mail server has a temporary network hiccup, so reports are far less likely to go missing.
-
Fixed
Dozens of scan accuracy fixes
A sweep across the audit checks removed many false positives and caught issues that were previously missed, so your report reflects your site more accurately.
-
Improved
Redesigned report and dashboard
A design overhaul brings a signature health gauge, a tighter category scorecard, and a more coherent look across your report and dashboard.
-
Improved
Reshaped report: verdict, category map, per-finding action
The report opens with a top-line verdict and a category map, and every finding now carries a clear next action.
-
Fixed
Scan field accepts bare domains
Enter a plain domain like example.com in the homepage scan field without needing to type the full address.
-
Improved
Studio scans up to 100 pages per site
The Studio plan page limit doubled from 50 to 100 pages per scan, with longer scan time budgets so large sites finish.
-
New
Content depth analysis in Studio reports
Studio reports now score your pages for content depth and flag thin or generic copy, with specific things to strengthen.
-
New
AI readiness now has a 0 to 100 score
AI readiness moved from a checklist to a real 0 to 100 score, with the impact of each fix and how the score has trended over time.
-
Improved
Redesigned homepage
A refreshed homepage that frames SiteCurl as an ongoing protection system for the sites you manage, not just a one-off scan.
-
New
Clean markdown for AI agents
SiteCurl pages serve clean markdown to AI agents that request it, so tools built on top of SiteCurl get structured content instead of HTML.
-
Improved
Score-drop alerts on every plan
Score-drop alerts are now available on all plans, so you hear about a regression no matter which tier you are on.
-
Fixed
Scans complete on more protected sites
Fallback fetch tiers and better bot-block detection mean scans now succeed on many sites that previously returned a 403 to the crawler.
-
New
Monitoring dashboard with trends and change detection
Your dashboard tracks score trends over time and detects what changed between scans, turning one-off audits into ongoing monitoring.
-
Improved
Findings as scannable table rows
Report findings render as compact table rows with progressive disclosure, so a long report is far less overwhelming to work through.
-
New
MCP server with 11 tools for AI agents
The Model Context Protocol server exposes 11 tools so AI agents can run scans and read reports directly.
-
New
Dismiss findings as accepted risks
Accept or dismiss a finding as a known risk so it stops counting against you, and track your setup with an onboarding checklist.
-
New
REST API, webhooks, and API docs
A full external API with webhooks, an MCP server, an llms.txt discovery file, and public API documentation for building on SiteCurl.
-
New
How-to-fix guides for every check
A library of public pages explaining each audit check and how to fix it, linked from your findings.
-
New
Dashboard sparklines and a priority inbox
Per-site score sparklines and a priority inbox surface the sites that need attention first.
-
Improved
JavaScript-rendered sites scan correctly
A browser-rendering fallback means sites that build their content with JavaScript now scan accurately.
-
Improved
Manage your subscription via Stripe
Update payment details, change plans, and view invoices through the Stripe billing portal.
-
New
Help center and scan-to-scan deltas
A searchable help center at /help, plus a delta view that shows exactly what changed between two scans of a site.
-
New
Self-service data export
Export your account data yourself through a secure, signed download link.
-
New
View your reports inside the app
Site owners can read full reports in-app with a navigation sidebar, not just through a shared link.
-
New
Cancel a running scan
Stop a scan in progress from your dashboard instead of waiting for it to finish.
-
New
White-label reports for Studio
Studio users can brand reports with their own logo and colors before sharing them with clients.
-
New
PDF and CSV export, plus one-click sharing
Export any report as a branded PDF or CSV, and share a live report with a single link.
-
New
SiteCurl launched
Scheduled website audits across SEO, speed, security, accessibility, technical health, uptime, and AI readiness, with scored, shareable reports and guided onboarding.
See what SiteCurl finds on your site.
Add your first site. The first report lands within minutes.