Tool guideWeb & SEO Utilities

Website Screenshot & Browser Inspector – Complete Guide to Page Captures, Load Timings and Resources

How the NetTools.tech Website Screenshot & Browser Inspector captures a public page in a sandboxed Chromium at several screen sizes, and how to read its screenshots, load timeline, resource waterfall, redirects and headers.

By NetTools.tech · · 11 min read

A redesign went live and someone reports that the mobile menu covers the content, a marketing page looks fine on your laptop but a colleague abroad sees a consent wall, or a page suddenly takes seconds to show anything and you need to know which requests are responsible. Looking at a page from outside your own browser, cache and network is the quickest way to separate "it is my machine" from "it is the site".

The Website Screenshot & Browser Inspector opens a public page in a sandboxed Chromium on the NetTools.tech server, captures it at up to three screen sizes, and reports what the browser saw on the way: the redirect chain, the response headers and TLS certificate, the load timeline and every network request the page made.

What is a website screenshot and browser inspection?

In simple terms

A website screenshot is an image of a page as a real browser renders it, at a given screen size. A browser inspection records the details behind that image: which address the page finally ended up at, how long it took until something was painted, and which files (scripts, stylesheets, images, fonts, API calls) were loaded from which servers. Together they answer two questions: what does the page look like, and why does it look or behave that way.

Technical background

Modern browsers expose their own measurements to the page through standard APIs. The Navigation Timing entry records when the first byte of the HTML arrived, when the DOM became interactive and when the DOMContentLoaded and load events finished. The Paint Timing API records first paint and first contentful paint, the Largest Contentful Paint API reports when the largest image or text block was rendered, and layout-shift entries are combined into Cumulative Layout Shift (CLS) using session windows of at most five seconds.

Network details come from the browser's DevTools protocol: every request with its resource type, status code, MIME type, transferred bytes, protocol (for example h2) and start and end times. Screen size is emulated with a viewport width and height plus a device pixel ratio; a 390×844 mobile viewport at 3× density produces an image 1,170 pixels wide.

Why use the Website Screenshot & Browser Inspector?

  • Responsive checks: see desktop, tablet and phone layouts of the same page in one run, side by side.
  • Visual evidence: keep a full-page image of a page at a point in time, for a bug report, a change review or a hand-over.
  • First-visit view: see consent banners, interstitials and geo-dependent content that your own browser hides because it already has cookies.
  • Performance triage: find the slow or heavy requests in the waterfall, and see how much comes from other sites such as tag managers and CDNs.
  • Redirect and header checks: confirm where a URL really ends up, which status code it returns and which security headers are present.
  • Before and after: capture the same URL again after a change and compare the two images with a slider.

Who should use it?

Web developers and designers checking layouts at several sizes, SEO specialists verifying titles, canonicals and redirects as rendered by a real browser, operations teams confirming that a deployment or CDN change looks right from outside, and support staff who need a neutral screenshot of what a public page shows.

How the Website Screenshot & Browser Inspector works

When you start a capture, the NetTools.tech API validates the address (http or https, a public host name or address, ports 80, 443, 8080 or 8443, no embedded credentials), resolves the host and refuses it immediately if it points to a private or reserved address. The job then joins a small queue; at most two captures run at the same time and the page shows your position.

A separate browser service runs a headless Chromium with its sandbox enabled. For every screen size it creates a new, empty browser context: no cookies, storage or cache from earlier captures, service workers blocked, downloads refused, pop-up windows closed and dialogs dismissed. The locale is en-US and the time zone UTC. The browser navigates to the page, waits until the condition you chose is reached (DOM ready, the load event, or half a second without network activity, for at most 15 seconds after the document arrived), waits the extra delay you set, measures the page and captures the image.

All of the browser's traffic, including requests to localhost, goes through a filtering proxy. For every connection the proxy resolves the host name once, refuses it if any returned address is private, loopback, link-local (such as the cloud metadata address 169.254.169.254), carrier-grade NAT, multicast, reserved or one of NetTools' own servers, and otherwise connects to exactly the address it checked. Because there is no second lookup, a DNS answer that changes between check and connection (DNS rebinding) cannot redirect the browser into an internal network. Only ports 80, 443, 8080 and 8443 are allowed. The browser service itself has no network access except to this proxy, which is enforced by the operating system as a second layer.

Limits keep the service fair and safe: three sizes per capture, one active capture per visitor, 12 captures per hour per visitor, 6 per site per 10 minutes across all visitors, 30 seconds to receive the page, a 16,000-pixel image height and 40 megapixels per image. Results and images are stored under a random job identifier and deleted after 24 hours.

How to use the NetTools.tech Website Screenshot & Browser Inspector

  1. Open the Website Screenshot & Browser Inspector and enter the page address, for example https://example.com/pricing. A missing scheme is treated as https.
  2. Under Screen sizes, select up to three presets (Desktop 1366×768, Desktop HD 1920×1080, Laptop 1280×800, Tablet 768×1024, Mobile 390×844, Small mobile 360×780), or type a custom size such as 1440x900 and press Add.
  3. Choose Capture: Visible area (exactly one screen) or Full page.
  4. Choose the Format: PNG (lossless, largest), JPEG or WebP (smaller, with a quality setting under Advanced).
  5. Optionally open Advanced: when the page counts as ready (DOM ready, Load event, Network idle), an extra wait of 0–10 seconds for animations and late content, the pixel density (Auto uses each device's typical density), a dark colour scheme for sites that support prefers-color-scheme, and whether the page's JavaScript runs.
  6. Press Capture. The progress panel shows the queue position and which size is being captured; Cancel stops the job.
  7. Use the tabs Screenshot, Overview, Resources, Network and Compare. With several sizes, the size switcher above the tabs selects which one the tabs show.

Understanding the results

ResultMeaningWhat to look for
ScreenshotThe captured image with Fit, 100%, zoom, Download, Copy image and Open"Full page" images are cut at the height limit and marked truncated
HTTP statusStatus code of the final page response, plus the number of redirects4xx/5xx means the screenshot shows an error page
Page readyTime until the chosen ready condition was reached"not reached" means the page never settled; it was captured at the time limit
Largest paint (LCP)When the largest text block or image was renderedLate values point to slow images, fonts or render-blocking scripts
Layout shift (CLS)Unexpected movement of content while loadingValues above 0.1 are noticeable; above 0.25 is poor
Requests / TransferredNumber of network requests and bytes receivedCompare sizes: mobile pages sometimes load more, not less
FailedRequests with an error or an HTTP status of 400 or higher, including blocked onesMissing assets (404) and refused connections
Other sitesRequests to a different registrable domain than the pageThird-party scripts, fonts, analytics, CDNs
Load timelineFirst byte, first paint, first contentful paint, LCP, DOM interactive, DOMContentLoaded, loadn/a means the browser did not report the value
ResourcesEvery request with type, status, size, duration and a waterfall barLong bars early in the waterfall delay rendering
NetworkFinal URL, protocol, server address, TLS version and certificate, redirect chain, response headersMissing security headers, unexpected redirects, certificate expiry
CompareAll sizes side by side, and a slider against an earlier capture of the same URLLayout changes between sizes or between two points in time

Cookie values in response headers are removed before anything is shown or stored; only cookie names remain. Messages under Browser messages are errors reported by the page's own scripts, plus pop-ups and downloads the capture browser refused.

Example: comparing desktop and mobile

An illustrative capture of a shop page, https://shop.example.com/, at Desktop and Mobile, full page, PNG, ready at the load event plus one second, could look like this:

Desktop 1366×768Mobile 390×844
Image1366×2210 px, 1×1170×5490 px, 3×
HTTP status200200
Largest contentful paint420 ms610 ms
Layout shift0.0120.180
Requests3437
Transferred1.21 MB1.48 MB

The mobile layout shift stands out. In the Resources tab of the Mobile capture, filtering by image shows a hero image loaded at 980 ms without width and height attributes, which pushes the text down when it arrives. The Compare tab shows both images side by side, and after the fix a second capture of the same URL can be compared with the first using the slider. These numbers illustrate the layout of a result; your values depend on the page and the moment of capture.

Common use cases

  • Checking a responsive layout at three widths before or after a release.
  • Producing a dated full-page image of a public page for documentation or a ticket.
  • Seeing what a first-time visitor sees, including consent banners and redirects to a regional site.
  • Finding which third-party domains a page loads and how much they transfer.
  • Confirming that HTTP redirects to HTTPS and that the final URL is the canonical one.
  • Checking that a dark-mode stylesheet works by capturing with the Dark colour scheme.

Troubleshooting

"Blocked by the safety filter"

The page address, a redirect or the host's DNS points to an address the capture browser may not contact (private network, loopback, link-local, reserved or a NetTools server). Check where the URL redirects with the Redirect Checker and what the name resolves to with a DNS lookup. Internal or staging sites cannot be captured.

"Site not reachable" or "Timed out"

The DNS lookup or connection failed, or the server did not send the page within 30 seconds. Check whether the site is up with Is It Down and the HTTP Status Checker. Some sites block data-centre addresses or automated browsers and never answer.

The screenshot shows a challenge page, captcha or empty page

Bot protection often treats headless browsers differently. NetTools does not try to bypass these checks. Turning off JavaScript sometimes shows the server-rendered page; otherwise the page cannot be captured from here.

"did not reach the load event" in Browser messages

A request never finished (a long-polling connection, a stalled third-party script), so the chosen condition was not reached within 15 seconds. The image was captured at that moment. Choose DOM ready or add an extra wait instead of Network idle for pages with continuous traffic.

The full-page image is cut off

The page is taller than the image limit (16,000 pixels, or less at higher densities because the limit counts device pixels). Capture at 1× density, or capture the visible area of the sections you need.

Lazy-loaded images are missing

Many pages load images only when they scroll into view. The browser does not scroll, so images below the first screen may be placeholders. An extra wait does not help with scroll-triggered loading.

Common mistakes

  • Comparing timings from two different tools or locations as if they were the same measurement.
  • Expecting logged-in content: no cookies or credentials are ever sent.
  • Choosing Network idle for pages with analytics beacons or chat widgets that never go quiet.
  • Reading "Other sites" as "trackers": it counts every request to another domain, including your own CDN on a different domain.
  • Choosing PNG at 3× for very long pages and hitting the size limit; WebP or JPEG is much smaller.

Limitations

  • One run per size from one location; results are a lab snapshot, not field data from real visitors.
  • DNS and connection times in the browser reflect the filtering proxy, not the visitor's path.
  • Only public pages, ports 80, 443, 8080 and 8443; no IPv6-only sites (the NetTools server has no IPv6 connectivity).
  • The browser does not scroll, click, log in, accept consent dialogs or run user-supplied scripts.
  • "Other sites" uses an approximation of the registrable domain that recognises common two-part suffixes such as co.uk, not the full public suffix list.
  • At most 400 resources are listed per size (all are counted in the summary); the image height limit is 16,000 pixels.
  • Sites may serve different content to data-centre addresses, to headless browsers or by location.

Security & privacy

The page address and options are sent to the NetTools.tech server, which loads the page with a disposable browser profile. Nothing you enter is executed: there is no way to supply scripts, headers, cookies, browser flags or extensions. The captured images and the inspection result are stored under a random identifier, are not linked to your IP address or an account, and are deleted after 24 hours or when you press Delete. Anyone who has the result link can view it during that time, so share it deliberately. The history of recent captures is kept in your browser's local storage only. Do not capture pages whose address contains secrets such as one-time tokens or private document links: the address is stored with the result and appears in the server's request log.

Summary

The Website Screenshot & Browser Inspector shows how a public page looks to a first-time visitor at up to three screen sizes and records what the browser did to get there: redirects, headers, TLS, load timings and every request. It runs in an isolated browser that can only reach public addresses, keeps results for 24 hours, and reports values it could not measure as unavailable rather than estimating them.

Frequently asked questions

How do I take a full-page screenshot of a website?

Enter the address in the Website Screenshot & Browser Inspector, choose Full page under Capture and start the capture. The browser measures the page height after loading and captures the whole page in one image, up to 16,000 image pixels tall; longer pages are cut and marked as truncated.

Can I capture desktop and mobile versions at the same time?

Yes. Select up to three screen sizes, for example Desktop and Mobile, or add custom sizes such as 1440x900. Each size is opened in its own fresh browser profile, so the page is loaded separately for every size and the results can be compared side by side.

Why does the screenshot show a cookie banner or a different language?

Every capture is a first visit with an empty profile: no cookies, no stored consent and the en-US language setting. Sites therefore show what a new visitor sees, which often includes consent banners, and they may adapt content to the location of the NetTools server.

Can it capture pages that require a login?

No. You cannot send cookies, credentials, custom headers or scripts, so only publicly accessible pages can be captured. Pages behind a login show whatever the site serves to an anonymous visitor, usually the login page.

Why are the load timings different from my own browser?

They are a single lab measurement from the NetTools server with an empty cache, and all traffic passes a filtering proxy. Network distance, caching and the proxy all change the numbers, so use them to compare sizes or runs rather than as your visitors' real experience.

Why are some requests marked as blocked?

The capture browser may only connect to public internet addresses. Requests to private networks, loopback, link-local addresses such as cloud metadata services, reserved ranges or NetTools' own servers are refused by the egress filter and shown as blocked with the reason.

How long are screenshots stored?

Images and results are kept on the NetTools server for 24 hours and then deleted automatically. You can delete a capture earlier with the Delete button; the history list is stored only in your own browser.