Every finding goes to AI auto-fix agents. They write the patch, cross-check it against your code, and deliver it as a pull request on your GitHub repo. Nothing is merged without your approval.
Works through the Claude Code CLI harness
SiteHealthScan runs 127 read-only checks across security, SEO, AI search, performance, domain, email, compliance and accessibility, then one click sends the findings to AI auto-fix agents that open a pull request on your repo.
Join The WaitList - Launching in Four Days
Free during launch week. One email at launch — no spam.
Use as an API-first engine
Site, code and backend. One scan.
Beyond the scan.
AEO scanner
ChatGPT, Claude and Perplexity are the new top of the funnel. SiteHealthScan audits whether they can parse your site, trust your claims and cite you, then ships the rewrites that flip the answer.
Live threat feed
Auth-bypass probes, credential-stuffing bursts, OWASP top-10 hits against your routes. We log the IP, the rule and the timestamp; you get a single Slack message.
Uptime
Sixty-second probes from 12 regions, with a one-minute minimum downtime floor so flapping doesn't blow up your inbox.
The full website health stack. One scan.
Paste the URL, get the 4 criticals, paste the fix prompt into your agent. Done before the coffee's cold.
One scan, one green badge. Investors don't have to take your word for it anymore.
Same check list, both branches, side-by-side score diff. Pick the winner with evidence.
Hooks into CI. A bad CSP header fails the build before the prod URL is even live.
Run scans across every site on your roster, export branded PDFs, schedule the monthly drop.
Every new client site goes through a baseline scan before kickoff. No surprises three weeks in.
40 client sites, one dashboard. Track regressions the moment a deploy lands.
Every finding is a fix-prompt ready to paste into Cursor. No more "where do we start?"
Public health score with a live timestamp. Customers see the same number you see.
The same scan output that powers the report also maps to your auditor's controls list.
Only fires when a real check regresses. No "your header is 1 char shorter than yesterday" spam.
Pipe every finding into Linear, Jira, Slack, or your own internal tool. No new tab required.
Pre-built ruleset for the headers your auditor actually checks, with a one-click paste fix.
Sub-minute probes against /cart and /checkout, with a fallback pinger if the main probe stalls.
Schema coverage report by template. Know which PDPs and PLPs the answer engines can cite.
Surfaces when third-party scripts are pulling more than they should off your storefront.
A clean public score on your portfolio is a quiet flex in the application queue.
Catches the broken og:image and missing alt text you missed when you shipped the case study.
DNSSEC, TLS chain, expiry runway - everything recruiters might check before they click through.
Set it once. The badge stays correct. The drift alerts only fire when something actually breaks.
"We replaced four separate tools with SiteHealthScan and the score-drift alerts caught a CSP regression our header-checker missed for two weeks. Already paid for itself."
"The fix prompts are the part I didn't expect to love. I paste them straight into Claude Code and half the work is already done. The other half is the wiring I would've done badly."
"The AEO pillar is the only place I've seen llms.txt coverage reported as a real check rather than a marketing bullet. Three of our top-cited pages were missing structured data."
Repo scanning
Daily monitoring
Exports
Domain health
Score over time
Your first scans are free (3 a day) with your score and critical count. Paid plans open every finding, one-click auto-fix and continuous monitoring.
See where you stand. No card, no expiry.
The full report, plus a fix for every line of it.
For agencies shipping other people's sites.
Every plan, one ladder
Annual billing saves 20%. Need more than 25 seats, a DPA or invoicing? Enterprise is a conversation, not a form.
Launching in four days. Reserve your spot and your first scan runs the moment we open.