Markup churn looks important
Hydration identifiers, timestamps, reordered attributes and injected widgets can change the document while leaving the page's meaning untouched.
Rendered DOM and visible-text evidence
See the HTML change behind the page update.
Compare the page response or its browser-rendered DOM, isolate the region that matters, and review text, structure, section and metadata evidence without reading a wall of source code.
- Comparison
- Text + structure
- Targeting
- CSS or XPath
- Rendered capture
- Pro and above
Rendered HTML comparison
acme.com/pricing · capture 0048
<main id="pricing"> <h1>Simple pricing</h1> <strong>$12 / month</strong> <strong>$15 / month</strong> <p>50 monitors included</p></main>- Text
- 6.7%
- Structure
- Stable
- Class
- Pricing
Price changed from $12 to $15. The surrounding component structure remained stable.
A changed hash does not explain what moved.
Modern pages regenerate markup constantly. Useful HTML monitoring has to separate a meaningful content edit from template churn, head-tag drift and an application that merely rendered the same message differently.
Visible copy can hide inside structure
A liability clause, stock label or plan limit may move within a large document. The reviewer needs the changed section, not thousands of surrounding lines.
Search metadata lives outside the body
Titles, canonicals, robots directives, hreflang and JSON-LD can drift without creating an obvious visual change for the person opening the page.
One change, separated into useful evidence.
OnChange classifies the surfaces independently so text, structure and search-facing changes can carry different review priority.
Visible-text comparison
Script and style content is removed from the reading layer, making customer-visible wording changes easier to distinguish from document plumbing.
Structural HTML signal
Tag and document movement remains available when a layout, component wrapper or interaction changes without rewriting the visible copy.
Section-anchored redlines
Long documentation, terms and policy pages are divided by headings so the changed section can be reviewed with its own before and after context.
Field-level SEO evidence
Head metadata is extracted into named fields, which turns a canonical or robots change into a precise value comparison instead of an opaque source diff.
Selector and normalization controls
Target one stable element and remove known volatile patterns before comparison, reducing noise without discarding the source snapshot.
Optional browser rendering
Pro plans can load JavaScript pages in Chromium and wait for a known selector or bounded delay before the captured content is compared.
From response to reviewable HTML diff.
The monitor keeps capture, comparison and notification as separate decisions, which prevents a noisy document from becoming a noisy inbox.
- 01
Capture the chosen surface
Fetch the document or render it in Chromium, then apply the configured selector and normalization rules.
- 02
Compare against the baseline
Calculate visible-text, structural, section and metadata differences against the last successful state.
- 03
Classify consequence
Attach change class, confidence, severity and optional AI context without replacing the underlying diff.
- 04
Route or retain
Send a meaningful alert or keep a low-priority change in the 90-day history for later investigation.
Know what the HTML evidence can prove.
A source-level record is strong operational evidence, but it is not interchangeable with a screenshot, browser session recording or permanent legal archive.
- Pair HTML monitoring with a visual monitor when CSS, images, responsive layout or overlays affect the customer-visible outcome.
- Use JavaScript rendering only when the important content is absent from the initial response, since a real browser is slower and more expensive to operate.
- Keep selectors stable and narrow. A selector tied to generated classes can fail even though the business content is still present elsewhere on the page.
- OnChange retains operational change history for 90 days. Export or use a dedicated records system when a longer legal retention policy applies.
Questions about html change detection
Before you add the first monitor
- Does HTML change detection compare source code or rendered content?
- A standard text monitor compares the fetched page response. When JavaScript rendering is enabled on Pro or above, OnChange loads the page in Chromium and compares the rendered result after the configured wait. The change record also separates visible-text movement from broader structural HTML movement.
- Can I monitor only one HTML element?
- Yes. Use a CSS or XPath selector to isolate a price, policy section, release note, table, headline or other stable region. Targeting the smallest meaningful region reduces navigation, timestamp and recommendation noise before diffing begins.
- How is an HTML diff different from a visual diff?
- An HTML diff explains changed text, tags, attributes, sections and metadata. A visual diff compares rendered screenshots and can catch appearance changes caused by CSS, images or layout. Use both when the markup and the customer-visible result are equally important.
- Will every HTML change create an alert?
- No. OnChange records the evidence first, then applies thresholds, normalization patterns, confidence signals and optional semantic-only rules. A cosmetic re-render or wording-only edit can remain in history without creating the same urgency as a changed price, clause or canonical URL.
Start with one page whose HTML matters.
The Free plan includes five text monitors. Upgrade to Pro when the page requires browser rendering or visual evidence.