PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteUse Lighthouse to diagnose how one page performs, WebPageTest to compare browser loads under chosen conditions, and k6 when you need scripted browser interactions or concurrent traffic. These methods answer different questions: a browser audit can reveal page-level bottlenecks, but it does not prove how many simultaneous users a service can handle.
Choose the test that answers your question
| Your question | Start with | Record |
|---|---|---|
| Where can one page’s performance be improved? | Lighthouse in Chrome DevTools | Page, audit configuration, metrics, opportunities, and diagnostics observed. Chrome’s Lighthouse guide describes the report contents and workflow. |
| How does a browser load vary by location, browser, or connection? | WebPageTest | Browser, test location, connection profile, number of runs, and whether the result is first view or repeat view. See WebPageTest and its documentation. |
| How does a scripted journey behave in a browser? | k6 browser | The journey, browser environment, and collected metrics. See k6 browser documentation. |
| Can the server handle concurrent traffic? | Protocol-level load test, such as k6 protocol testing; add browser tests if user-visible rendering also matters | Workload shape and server-side measures. Grafana explains the distinction in its website load testing guide. |
“How fast is it?” is a useful starting question, but the answer depends on what you measured and under which conditions. A browser test represents its selected browser, location, connection, and visit state—not every user or production region.
As an Amazon Associate I earn from qualifying purchases.
Establish a page-level baseline with Lighthouse
- Open the page in Chrome and open DevTools.
- Open the Lighthouse panel, choose the audit settings relevant to your question, and generate a report.
- Keep the page and audit conditions consistent when comparing a change with the baseline. Chrome recommends auditing after each change so its isolated effect can be assessed.
- Use the report’s metrics, screenshots, opportunities, diagnostics, and passed audits to investigate what happened. Do not treat a single score as a complete explanation.
For repeatable automation or more configuration, Lighthouse also offers a command-line interface and a Node module. Check the Lighthouse project README for current installation and runtime requirements; those can change.
Represent a real audience with WebPageTest
- Enter the page URL and choose a test location near the users whose experience you want to approximate.
- Select a browser available at that location and a connection profile suited to the scenario you are investigating.
- Choose the number of runs, then decide whether to examine a first view, a repeat view, or both.
- When results vary, inspect the request waterfall and visual filmstrip to understand the timeline rather than reporting one favorable run as a stable result.
Keep first view and repeat view separate
A first view represents a fresh-visit scenario. A repeat view is intended to represent another visit with browser state or cache retained according to the test setup. They answer different questions; label which one you are reporting. Multiple runs help expose variability, while repeat view represents another visit state. WebPageTest documents its run and view controls in its documentation.
#1 Best Overall
Make comparisons interpretable
Change one relevant factor at a time when diagnosing a page. Compare like-for-like runs, and include the browser, location, connection profile, run count, and view type alongside any reported result. A test from one location or connection should not be generalized to all visitors.
Use k6 for scripted journeys and concurrent traffic
k6 supports browser scripts through a Chromium-based browser as well as protocol-based tests. A browser script can exercise interactions and observe browser metrics; a protocol script can generate server requests more efficiently. They are complementary, not interchangeable. Combine them if the objective includes both backend load and the experience of a real browser journey.
Rank #2
Define the workload before running a capacity test. Specify the number of virtual users, arrival pattern, duration, test geography, browser mix, and user journey. There is no universal value for these settings: tailor them to expected traffic and the environment being tested, and avoid testing production systems without authorization and an appropriate plan.
What to include in a useful result
- The page or user journey tested and the date of the run.
- Tool, browser and relevant browser configuration.
- Test location and connection profile for remote browser tests.
- First-view or repeat-view state and number of runs.
- For concurrency tests, virtual-user count, arrival pattern, duration, geography, browser mix, and server-side measures.
- The traces, waterfalls, filmstrip, or browser metrics used to support the interpretation.
Results can vary with conditions. State the setup and use repeated comparable measurements before claiming that a change improved performance. No single Lighthouse score or remote browser run establishes service capacity.
Troubleshoot misleading or unusable results
- One score does not explain the slowdown: inspect Lighthouse diagnostics and the underlying page evidence, then change one factor and audit again.
- Remote runs disagree: compare the same browser, location, connection, and view type across multiple runs; use the waterfall and filmstrip to locate where timelines diverge.
- The result seems unlike your audience: check that the selected test location is near the intended users and that the browser and connection profile reflect the question. The result is scoped to those choices.
- A repeat visit looks much faster or different: confirm whether the run is a repeat view and report it separately from a fresh first view.
- A browser audit is being used to claim capacity: use a designed concurrent protocol or browser load test with an explicit workload. Lighthouse is a diagnostic audit, not a capacity test.
- Automated Lighthouse setup fails after installation: verify the current runtime and installation requirements in the project README rather than relying on an old setup recipe.
Or skip the browser setup
For a screenshot of a page rather than a performance or capacity test, ScreenshotNeo provides a website screenshot API and MCP server. One GET request can return a PNG, JPEG, WebP, or PDF. It is not a load-testing substitute: it captures a page, while Lighthouse, WebPageTest, and k6 address performance and workload questions.
Example cURL request (see the ScreenshotNeo API documentation):
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpCookie banners, popups, and chat widgets are removed before the shot; bot checks, blank pages, and failed loads are never billed. An MCP server lets AI agents take screenshots. The free plan includes 1,000 screenshots a month with no card, and paid plans start at $5 for 3,000. Sign up for 1,000 free screenshots a month, with no card required.
Frequently Asked Questions
Does Lighthouse load-test a website with multiple users?
No. Lighthouse audits an individual page; it does not establish concurrent-user capacity.
Should I use browser-based or protocol-based testing?
Use browser-based testing when interactions, rendering, or browser metrics matter; use protocol testing for efficient server traffic generation. Use both when the objective requires both views.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
