A day in the life of a monitored site
Your client gets a white-label report you did not write.
One site is checkable by hand, and a client list is not: open the page, run PageSpeed, catch the certificate before it lapses, write the update at the end of the month, on every one of them.
No card needed. Two weeks of full monitoring on your first site.
Nothing wakes you up until a second region agrees
The homepage gets an HTTP probe every 60 seconds from the EU, so a site going down is caught in minutes instead of at your next manual check. When the EU and the US disagree, a third probe in Asia-Pacific settles it.
Alert sent 03:42 by the uptime rule you set, one minute after the EU probe failed and the US confirmed.
no alert: only the EU probe saw it
Alert sent 03:42 by the uptime rule you set, one minute after the EU probe failed and the US confirmed.
19:22 · no alert: only the EU probe saw it
This is the homepage. Every other monitored page is reached by the daily scan below.
When a client asks how their site is doing, the answer is already scored
Every monitored page carries four Lighthouse scores, re-taken daily: Performance, Accessibility, Best Practices and SEO, audited from real Chrome in the same environment every time, so week-over-week deltas are comparable and a slow regression reads as a trend rather than a surprise.
Every site gets one fixed time of day, its slot, 12:30 for this one, and the scan walks every monitored page in a single pass. Which pages is not your job. Up to 15 are discovered from the sitemap and the homepage's own links, ranked by importance, and re-synced weekly so new pages get picked up.
Most visual changes are noise.
You only hear about the one that is not.
Every monitored page is also photographed in full and compared pixel by pixel against the last good capture, normally yesterday's. AI sorts each significant change into three verdicts, and only one of them interrupts you, which is what keeps a daily visual check from becoming a daily email. Every screenshot stays in your 2-year history, so you can check the call it made.
carousel on a different slide
new hero copy, the layout untouched
hero image gone, nothing in its place
Your client should not be the one who finds the dead link
Content Health requests every link and every image on every monitored page, once a day. Anything that answers with an error, or never answers at all, is listed on your dashboard and counted in the client report, and when the count of new breakages passes the threshold you set, it raises an alert.
One day's crawl
15 pages · every link, every imageone segment per monitored page
The certificate nobody is watching, counted down every day
Renewals watches both dates for you, the certificate and the domain registration, and you pick how much warning you want: 1, 3, 7, 14, 30, 60, or 90 days before either lapses. Miss the certificate date and everyone who visits meets a browser warning instead of the site. Both readings are taken on the site's first scan and refresh once a day on its slot after that.
SSL certificate
expires in 41 days
you pick the alert day (30 shown)
Domain registration
expires in 148 days
you pick the alert day (30 shown)
Not every registry publishes an expiry date. On .at, .au, .be, .bg, .ch, .de, .eu, .fi, .hu, .jp, .li, .nl, .no, .nz, .ro it may not, and a domain with no published date reads "Expiry date not published by registry" instead of counting down.
Monday is when your client sees the care plan working
Change nothing and the week rolls up on its own: on Monday, the report lands in your inbox, white-label in your agency's name, for you to read and forward. Verify a domain you own and it can go to the client directly instead, sent from your address. Forwarded or sent, it opens in your client's inbox with no SiteLetter login, ever.
Each schedule's cadence is chosen once at agency level: weekly, bi-weekly, monthly, or quarterly, and a new account already has two running. Each site's report can then be moved to the day that client expects it.
See the full report, section by sectionacme-client.com · 12 pages · 2026-09-21 - 2026-09-27
Checked 12 pages - 2 meaningful changes
Hero image area is blank - broken image placeholder visible.
Hero copy rewritten and a new testimonial card added.
What the whole day's work costs
The five monitors are switched on per site; the receipt is all of them, on one client site, for a month. There are no plan tiers and no subscription: you top up a prepaid balance when you want to, and every site draws its own days from it.
That is one site. 20 sites is $96.00 a month for the same day on every one of them. Set it next to one month of one client's care plan.
If the balance runs out, monitoring stops for the sites still switched on, and the history stays where it is.
Weighing it against another tool? Browse all alternatives.
- full-page capture of up to 15 pages
- compared with yesterday, AI triage
- Lighthouse scores on the same pages
- every link and image on those pages
- homepage probe, second region confirms
- certificate + domain watch
No alert is sent until you write the rule
Alerting is the one thing you set up. A new account has no rules at all, and you write them once at agency level. Results still land in the dashboard and in the report either way.
- From
- support@siteletter.com
- Subject
- brightharbordental.com is DOWN
The homepage failed its uptime check.
View incidentsWhere each check lands
| Check | Where the result goes |
|---|---|
| Uptime probe Uptime | Email or Slack, on the rule you set |
| Lighthouse audit Page Scores | Report, plus email or Slack under your score threshold |
| Visual change detection Visual Changes | Dashboard and report, plus email or Slack on anything possibly broken |
| Broken images and links Content Health | Report, plus email or Slack over your count threshold |
| SSL + domain expiry Renewals | Email or Slack, on the lead times you pick |
| Client report No separate charge | Your inbox, or your client's |
The small line under each check is the name it is billed under on the receipt.
Plain answers
Yes, identically. Monitoring runs from the outside, the way a visitor would, so WordPress, Webflow, Shopify, Squarespace, Framer, and custom builds all get the same coverage. There are no scripts, plugins, or DNS changes on any of them, and nothing that could slow the client's site down.
Visual changes only, and only ones big enough to matter. A change it cannot judge confidently is marked couldn't verify rather than guessed. Uptime, SSL, domain, and Lighthouse alerts use plain thresholds, no AI involved. Screenshots go to OpenAI in the US for visual change classification, under EU Standard Contractual Clauses.
Anything that isn't a public page. SiteLetter monitors public pages only, so sites behind Basic Auth or password protection and staging environments are not covered. There's no API-endpoint or checkout-flow monitoring either. Monitored pages are auto-selected, and choosing them yourself is not available today.
The report carries your agency's logo, colors, and sender name, replies go to the address you choose, and an org-level toggle removes the SiteLetter name from the report body. Verify a domain you own, with five DNS records, and the email is sent from that domain too, signed by it, with no 'via siteletter.com' line. Two honest caveats: the shared report link is still a siteletter.com URL, and alert emails still come from SiteLetter.
Email, to any of your team members' addresses, and Slack via Incoming Webhooks, in any combination. Alerts go to you and your team, never to clients. The branded report is the only thing a client ever receives.
2 years of Lighthouse scores, screenshots, asset checks and reports. Uptime is summarised over the last 90 days and raw uptime check rows are kept 280 days; downtime is kept as incident records. All uptime results are stored in the EU; probes running elsewhere keep only short-lived operational logs in their own region.
Up to 100 client sites in one organisation, and get in touch if you need more - you pay per site, nothing else changes. Your whole team comes along: up to 200 members per organization, with no per-seat pricing.
You set up care-plan reporting once.
After that it runs without you.
- ✓ Just add a domain. The first scan is booked for the next free quarter-hour, usually within minutes, not the site's next 12:30.
- ✓ The first report is already scheduled. A weekly summary and a monthly roundup are scheduled from the day you sign up.
- ✓ Nothing recurring. No subscription and no contract. You buy credits when you want them, each site spends its own, and you stop by not topping up.