Know when your site breaks, before a visitor tells you
Most sites find out they are down the way a shopkeeper finds out the front door is stuck: someone knocks. Site Watch stands outside the door from ten places at once, walks the aisles on a schedule, and sends one note when something changes and one more when it is fixed.
Free for one site. No card, no script to install for the checks that exist today.
Two views of one site
Outside view
The site as a visitor from 10 locations sees it
A small probe runs on three providers across North America, Europe, South America and Asia Pacific. Each one fetches your homepage every cycle, records how long DNS, the handshake and the first byte took, and reports in. A separate crawl walks the pages on a schedule and compares what it found with last time.
Live now.
Inside view
The site as your visitors experience it
A script under three kilobytes, optional, that reports script errors, images that failed to load, the dead pages people actually reached and the page speed they actually got. It sets no cookies and sends nothing that identifies a person.
Coming next. Not yet available.
What it checks
Availability
Homepage reachability, status and time to first byte from every location, every 5 minutes on Pro
DNS resolution and TLS handshake time as seen from each region
A verdict only by consensus: a single location failing is a regional notice, not a down alert
Response-time drift against your own 30-day median
Pages and links
Broken internal links, with the page that links to them
Broken external links, each target checked once a day at most
Redirect chains and loops, and pages that return 200 while rendering an error
Soft 404s, missing or mismatched canonicals, duplicate titles, missing alt text
Sitemap entries that 404, crawled pages the sitemap forgot, orphan pages
Pages whose word count fell by more than half since the last crawl
Certificate, domain, DNS
Certificate days left, chain validity and hostname match
Domain registration expiry from the registry record
DNS record drift since the last check
robots.txt newly blocking search or AI crawlers, and noindex tags that leak onto live pages
Coming next
Security headers and a fixed list of files that should never be public
Mail authentication for the domain (SPF, DKIM, DMARC)
Spelling, tuned for precision rather than noise
The visitor-side view: script errors, images that fail for real people, dead pages people actually hit
How alerts work
An email goes out when a state changes, never on a timer. Down and back up. A certificate crossing 21 days, then 7. New error-level problems after a crawl. Each one carries a line of context and a link to the dashboard, and the back-to-green email is not optional, because a green dot you cannot trust is worse than no dot.
Down is a verdict, not an opinion. It needs at least 3 healthy locations reporting in the same cycle, at least two of them failing, and failures in the majority. Every probe also checks two control sites each cycle; a location that cannot reach those is ignored for that cycle rather than allowed to vote. One location failing alone becomes a regional notice, which is often the more useful thing to know.
Worst case from outage to email: about 6 minutes on Pro, about 35 minutes on the free tier.
What it does not do
No backlinks. We do not keep a link index and will not build one. A borrowed one would be someone else’s coverage with our name on it.
No rank tracking, no keyword index. Site Watch answers whether the site is up, whole and findable. Where it ranks is a different product, sold by people who do that well.
No perfect uptime figure. The number you see is the share of cycles with a verdict that were up, with the cycles we could not observe counted beside it.
Privacy, by construction
The probe and the crawler identify themselves honestly, respect the limits a site sets, and are described on the bot page. Nothing about your visitors is collected today. When the visitor-side script arrives it will send no cookies, no identifiers, no page content and no form values, so it can sit on a public-sector site without touching the consent banner.
Yes, for one site. The free tier checks reachability every 30 minutes from 3 locations, crawls the site every 7 days, and emails you when the site goes down, comes back, or the certificate or domain is about to expire. Everything daily, historical or multi-site rides the existing Pro subscription.
Do I need to install anything?
Not for the checks that exist today. You add the site, prove it is yours with a meta tag or a DNS record, and the checks start within minutes. The visitor-side script is the next phase and will be optional.
How fast will I hear about an outage?
Pro sites are probed every 5 minutes from all 10 locations, so the worst case from outage to email is about 6 minutes. Free sites are probed every 30 minutes, so about 35 minutes. Both figures are stated rather than promised as instant, because they are what the system actually does.
Will it flood my inbox?
No. An alert fires once when something changes and once when it clears. Open issues roll into one digest, daily on Pro and weekly on free. There is no repeat reminder for a problem you already know about.
Does it track backlinks or rankings?
No. Site Watch does not keep a link index or a keyword index and does not pretend to by borrowing someone else's. It watches whether the site is up, whole and findable, which is the part no one else tells you about until a visitor does.