WebChange Detector takes scheduled screenshots of your pages and compares each one to the previous version. When something changes, you get an alert. When nothing changes, you hear nothing. When the change is just a rotating banner or a cookie popup, our AI verification sorts it out so it does not spam your inbox.
This page is the reference: every Monitoring setting in the Web App, what it does, when you would change it and what it is set to if you leave it alone. If you just want Monitoring running for a website, start with the short how-to How to Set Up Monitoring and come back here when you want the details.
How Monitoring works
Every setting on this page changes one step of the same run, so it helps to know what a run does.
A run starts (schedule, page save or "Run monitoring now")
↓
WebChange Detector takes a screenshot of every selected page and device
↓
Layout-aware comparison with the last Monitoring screenshot
(content that only shifted and rendering noise are filtered out)
↓
Is there a visible difference left?
├─ No → OK. No alert sent.
└─ Yes → Our AI checks every changed area
↓
├─ OK (dynamic content like a slider, or matches one of your feedback rules) → No alert
├─ Unsure → Alert sent
└─ Alert (real change: text, layout, missing element) → Alert sent
A few terms you will meet in the settings and the results:
Monitoring settings per website: Every website you add has its own Monitoring settings and its own list of pages: how often to check, when to stay quiet, who gets the alert email, and how the screenshots are taken. These are separate from the website’s On-Demand Check settings.
Difference Percent: When WebChange Detector compares two screenshots, it calculates what percentage of the page changed. A value of 0.5% means roughly 1 in 200 pixels is different. Even a tiny value can be a real problem, like a missing menu item or a broken button.
AI Verification: Our AI looks at every changed area of a comparison and classifies it: is this a real content change, or something that changes on every page load, like a rotating ad? This is what keeps busy, dynamic pages from flooding your inbox.
Feedback Rules: When the AI flags something you know is fine, you tell it to ignore that change in the future. From then on, the same kind of change is marked as OK on that page or website.
Result Statuses: After processing, each comparison gets a status: New (change detected), OK (no change, or nothing that matters), To Fix (reviewed, needs action), or False Positive (confirmed not a problem).
Where the settings live
Every website has its own Monitoring settings and its own list of pages, separate from its On-Demand Checks. On the Monitoring tab of the Web App, open the ⋯ menu of a website row and choose Website Settings. Switching Monitoring on, choosing the pages and saving a first schedule is covered step by step in How to Set Up Monitoring; adding a website in Getting Started.
Most settings apply to every monitored page of the website. A few can also be set for a single page, in the page’s own settings in the page list:
| Setting | Website Level | Per Page |
|---|---|---|
| Schedule, quiet hours, alert emails | Yes | No |
| Desktop / Mobile | Default for newly synced pages | Yes |
| CSS Injection | Yes (applies to all pages) | Yes |
| JS Injection | Yes (applies to all pages) | Yes |
Schedule
The schedule controls when WebChange Detector takes screenshots, and with that how fast you hear about a change and how many checks a website uses. You find it in Website Settings under Run in intervals, with three schedule types, plus Never for WordPress websites that are only checked when a page is saved.
Every Day (Default)
The most common schedule: WebChange Detector runs at a fixed interval, every day.
Available intervals (Monitoring Interval):
- Every 15 minutes (higher plans)
- Every 30 minutes (higher plans)
- Every 1 hour
- Every 3 hours
- Every 6 hours
- Every 12 hours
- Every 24 hours (the default)
Start time: Next to the interval you choose at which hour the runs start (“at … o’clock”). The times are shown in your own time zone, and for shorter intervals the dropdown lists every run time of the day, so with a 6 hour interval you see for example 08:00, 14:00, 20:00 and 02:00.
When to use it: Ongoing monitoring of production pages, high-value pages that need frequent checks (homepage, checkout, pricing), or pages that update frequently (news, product listings).
Specific Weekdays
Run Monitoring only on the weekdays you select. On those days, the interval and start time apply as usual.
Day selection: Monday through Sunday, any combination. Example configurations: weekdays only (Mon to Fri), after weekend deployments (Monday only), or before releases (Tue, Wed, Thu).
When to use it: Pages that only need checks on business days, a check after updates on a predictable day, or lower-priority pages that do not need daily monitoring.
Specific Days in Month
Run Monitoring on chosen dates within the month.
Date selection: Any combination of the days 1 through 30, plus a Last day option that adapts to the length of the month.
When to use it: Compliance checks that only need monthly snapshots, end-of-month reporting baselines, or low-change pages like legal, privacy policy, or about pages.
Other Ways to Start a Monitoring Run
Check when a page is saved: For WordPress websites with the plugin, a page is checked a few minutes after someone saves it with changed content, on top of the schedule or instead of it (schedule Never). See How to Check a Page Automatically as Soon as It Is Saved.
Run monitoring now: The button on every website row starts a Monitoring run right away, for example after you pushed a staging site to production. See Run Monitoring Now After a Deploy.
Next Run and Checks per Month
Every website row shows Next check in with a countdown. At the top of the Monitoring tab, Next Monitoring Checks sums up your monitored websites and URLs and the estimated Checks / month 1 of all your schedules, quiet hours and selected devices included. Open it to see the upcoming runs per website under Upcoming 2 and the recent results. The monthly figure is an estimate, so use it to compare schedules before you save them.
Quiet hours
Quiet hours define a time window during which Monitoring is paused. No screenshots are taken during this window, regardless of what the schedule says, and because there is no run, there is no alert either.
When Quiet Hours Are Useful
Planned deployments: If you push updates every night between 22:00 and 02:00, set quiet hours for that window. Monitoring resumes after the deployment window, and the next comparison reflects the updated site rather than a half-finished state.
Scheduled maintenance: If a client runs automated backups or maintenance scripts that temporarily alter page content, quiet hours prevent alerts during that window.
Staging environments: Some staging sites show maintenance banners during off-hours. Quiet hours prevent those banners from generating alerts.
Save checks: Skip checks during the night or low-traffic hours and use them during the day instead.
Where to Set Quiet Hours
In Website Settings, under Quiet Hours (no checks during this time): the first dropdown is the hour when Monitoring pauses, the second the hour when it resumes.
The hours are shown in your own time zone. Set both dropdowns to None to switch quiet hours off.
Midnight-spanning windows work correctly. If you set the window from 22:00 to 06:00, WebChange Detector pauses overnight and resumes at 06:00. You do not need to split this into two separate configurations.
What decides whether you get an alert
A visual difference alone does not mean an alert. Three layers decide what reaches your inbox, and each one is a lever you can use when a page keeps producing alerts you do not care about.
1. The Comparison
WebChange Detector compares the new screenshot with the last Monitoring screenshot of the same page and device. The comparison understands the page layout: when a section only moved because something above it got taller, or a line of text renders a pixel lower, that shift is not reported as a change. Tiny rendering noise is cleaned up too, but every real change is kept, however small.
2. AI Verification
Our AI then looks at every changed area and labels it OK, Alert or Unsure, with a short description of what changed. It recognises content that changes on every page load, like sliders caught mid-transition, review counts in third-party widgets, relative times (“4 days ago”) or live tickers, and lets those pass as OK. Before it calls an area OK, it checks that every element is still there and compares the text word by word, so a missing button or changed wording is always flagged.
The AI is conservative: if it cannot verify a change, the change is still shown to you rather than hidden. Only areas labelled Alert or Unsure end up in the alert. AI verification is included with every paid plan.
3. Your Feedback Rules
Some changes are expected on your site but look important to anyone else: a footer date, a rotating testimonial, a known console warning. When the AI flags one of them, click Ignore in future on that area in the comparison. This creates a feedback rule, and from the next check on, the same kind of change is marked Auto-ignored by your rule instead of raising an alert.
Rules are matched by meaning, not by exact wording, so “the date in the footer” still matches when the date is different. You can scope a rule to one page or to the whole website, and you find all rules on the AI Rules page. For the step-by-step, see Teach the AI What to Ignore.
Elements You Never Want in the Screenshot
A live chat bubble or a countdown timer does not need to be in your screenshots at all. Hide it with a line of CSS, or close a popup with a bit of JavaScript, and the comparison never sees it (see Advanced screenshot settings).
Website Settings also still shows a Difference threshold field; as the hint below it says, keep it at 0, because the AI classification already filters out the noise.
Alerts
Email Alerts
WebChange Detector sends an alert email when a Monitoring run finds changes that are still flagged after the AI verification. The same email also lists pages that could not be captured, so a page that silently fails never goes unnoticed.
Multiple recipients: Each website has its own list of alert email addresses for Monitoring, separate from the list for its On-Demand Checks. Add as many recipients as you need, separated by commas. This is useful for agency setups where both you and the client want to be notified. Leave the field empty and no alert emails are sent for Monitoring; the results still show up in the dashboard.
What the alert email contains:
- An AI-generated summary of what changed in the run
- A table of the changed pages with device, status, difference percentage and a short AI summary per page
- A link to each before/after comparison that opens without a login, so you can forward it to a client
- The pages that could not be captured, with the reason
When alerts are triggered:
- When our AI labels a changed area as Alert or Unsure
- When new browser console errors appear on a page (plans with browser console tracking)
- When pages of the run could not be captured
To configure alert emails, open Website Settings and fill in Email addresses for alerts. Enter one or more email addresses, separated by commas, and save.
Webhook Alerts
For teams that have their own alerting pipelines, WebChange Detector supports webhooks. Next to the email, WebChange Detector sends a POST request to your URL whenever a subscribed event occurs.
Available webhook events:
| Event | When It Fires |
|---|---|
| New Change Detected | A comparison was flagged as a change |
| Batch Completed | A complete run has finished processing |
| Screenshot Completed | A single screenshot was taken successfully |
| Screenshot Failed | A single screenshot failed |
Common integrations: Slack (route alerts directly into a channel via an incoming webhook URL), Zapier (connect to thousands of apps: create tickets, send messages, update spreadsheets), or your own system that processes the webhook payload.
To configure webhooks, open the Automations page in your dashboard and click Create Webhook. For a step-by-step guide, see Connect Slack, Zapier and Your Deploy Pipeline.
Devices
For each page, you can independently enable desktop and mobile screenshots:
- Desktop only: Full-width browser screenshot. Best for sites with no mobile-specific layout differences.
- Mobile only: Narrow viewport screenshot simulating a mobile device. Best for mobile-first sites.
- Both: WebChange Detector takes two independent screenshots and runs two separate comparisons, so it counts as two checks. Layout regressions on mobile are detected independently from desktop issues.
Why monitoring both matters: Responsive CSS and media queries mean the desktop and mobile views can break independently. A change to a flexbox layout might only affect mobile, while a navigation dropdown issue might only appear on desktop.
You set this per page with the Desktop 1 and Mobile 2 switches in the page list of a website. In Website Settings, Activate newly synced URLs by default decides whether pages that are synced later start with Desktop and/or Mobile switched on.
Advanced screenshot settings
These settings sit in Website Settings under the Advanced Settings accordion, except the screenshot region, which sits right above it.
Screenshot Delay
The Delay between screenshots (seconds) is the pause between two screenshot starts on the same website. It spreads the page loads of a run over time, so your server is not hit by many requests at once. It is not a waiting time before each screenshot.
Default: 10 seconds. Leave the field empty to use it.
When to increase the delay: Your hosting is small or shared, the server has strict rate limiting, or the site struggles while a run is going on.
The field sits in Website Settings under Advanced Settings 1.
Screenshot Region
Where the screenshots are taken from: Auto-detect (the closest region for the site), United States (US) or Europe (EU). EU renders the site from an EU IP address, which gives European sites the right geo-location and lower latency.
CSS Injection
CSS injection lets you add custom CSS to the page before the screenshot is taken. The CSS is applied and rendered, then the screenshot is captured. It does not affect the live site in any way.
Primary use case: hiding elements that should never be compared. Some elements change on every page load and carry no information for you, like live chat widgets or countdown timers. Instead of reviewing them again and again, hide them with CSS before the screenshot. For changes you still want to see in the screenshot, a feedback rule is the better tool.
Examples:
/* Hide a live chat widget */
.intercom-lightweight-app { display: none !important; }
/* Hide a countdown timer */
.countdown-timer { display: none !important; }
Website-level CSS applies to every page of the website: the right choice when the same element appears on every page (e.g., a sitewide chat widget). Page-level CSS applies only to one page. Use it for a page that has a unique element not present elsewhere.
JavaScript Injection
JS injection runs custom JavaScript on the page before the screenshot is taken. This is more powerful than CSS injection and enables interactions that CSS alone cannot handle.
Common use cases: Accepting a cookie consent programmatically, closing a modal that opens on page load, scrolling to a specific section before the screenshot, setting a localStorage value that affects page behavior, or dismissing a notification banner.
// Close a welcome modal by removing it from the DOM
const modal = document.querySelector('.welcome-modal');
if (modal) modal.remove();
Like CSS injection, JS can be set for the whole website (runs on every page) or for a single page.
Basic Authentication
If a page is behind HTTP basic authentication (username and password required at the server level, typically used for staging environments), WebChange Detector needs the credentials to access it.
To configure: open Website Settings, expand Advanced Settings, enter the username and password under Basic Authentication, and save. The password is stored encrypted and applies to all pages of the website.
Proxy Settings
By default, screenshots are taken from changing IP addresses in our cloud infrastructure, and when a site blocks them, WebChange Detector automatically tries other ways to reach it and remembers what worked. For sites with strict firewall rules that only let known IP addresses in, there is the static IP proxy.
Switch on Enable Static IP Proxy and all screenshots of the website come from one fixed IP address, shown right next to the switch with a copy button. Whitelist this IP in your firewall. Best for sites behind WAFs like Cloudflare, Sucuri, Wordfence, or managed hosting firewalls.
For a full guide on resolving firewall blocking issues, see our post: How to Stop Firewalls From Blocking Your Screenshots.
Where the results show up
Monitoring results appear in the Checks tab. Each Monitoring run shows up as one entry with all comparisons of that run. Click any comparison to open the full view with the before/after slider, the highlighted differences and the AI analysis. Each comparison also has a public share link you can send to clients without requiring them to log in.
For a complete walkthrough of the comparison view and the review workflow, see: How to Review Checks.
Configuration Quick Reference
| Setting | Where to Configure | Notes |
|---|---|---|
| Monitoring on / off | Switch on the website row | Per website |
| Schedule type (every day / specific weekdays / specific days in month / never) | Website Settings > Run in intervals | “Never” only for WordPress websites with the plugin |
| Monitoring interval | Website Settings > Monitoring Interval | 15 minutes to 24 hours |
| Start time | Website Settings > at … o’clock | Shown in your time zone |
| Quiet hours | Website Settings > Quiet Hours | Supports midnight-spanning windows |
| Check when a page is saved | Website Settings | WordPress websites with the plugin |
| Alert emails | Website Settings > Email addresses for alerts | Comma-separated list, empty means no emails |
| Feedback rules | “Ignore in future” in a comparison; AI Rules page | Per page or per website |
| Webhooks | Automations page | Event-based, for the whole account |
| Desktop / Mobile | Page list of the website | Per page; desktop plus mobile counts as two checks |
| Screenshot region | Website Settings > Screenshot Region | Auto-detect, US or EU |
| Screenshot delay | Website Settings > Advanced Settings | Pause between screenshot starts, default 10 seconds |
| CSS / JS injection (website) | Website Settings > Advanced Settings | Applies to every page |
| CSS / JS injection (page) | Settings of the page in the page list | Only this page |
| Basic authentication | Website Settings > Advanced Settings | Password stored encrypted |
| Static IP proxy | Website Settings > Advanced Settings | Whitelist the IP shown next to the switch |
Related Guides
- How to Set Up Monitoring: The short how-to to get Monitoring running for a website
- How to Review Checks: The complete guide to the comparison view, AI analysis, feedback rules, and review workflow
- Check a Page as Soon as It Is Saved: Monitoring that starts when someone edits a page
- On-Demand Checks: Before/after comparisons for planned updates
- Auto Update Checks: Automatically check your site after every WordPress auto-update
- Manage Multiple Client Sites: Bulk Changes and the agency workflow
- Firewall Troubleshooting: IP whitelisting guides for Cloudflare, Sucuri, Wordfence, and managed hosting