CDP throttling · 4× CPU
Loaded like a mid-range phone on a weak signal.
Every scan reloads your page on Slow 3G and Fast 3G with a 4× slower CPU, then shows how far LCP drifts from the baseline.
Real Chromium, throttled to Slow 3G and a 4× slower CPU. LCP, INP and CLS graded on Google's thresholds, then traced to the file behind them.
1440 × 900 · no throttling
Baseline passes. Every lab vital is inside Google's good range.
What gets measured
Every check answers two questions: what is slow, and which file made it slow?
CDP throttling · 4× CPU
Every scan reloads your page on Slow 3G and Fast 3G with a 4× slower CPU, then shows how far LCP drifts from the baseline.
Field data first
When Google has CrUX data for your page, that's the number you get. Every vital is labelled with its source.
JS & CSS coverage
Precise V8 and CSS rule coverage flags files over 40% unused, and names third-party scripts.
Long tasks · TBT
Every task over 50 ms is counted toward Total Blocking Time, and the longest is traced to its script.
WebP · AVIF
We download your largest PNGs and JPEGs and re-encode them, so the savings are real bytes, not estimates.
LCP element
Fold position, lazy-loading, fetchpriority and served-vs-shown size all go into the AI fix.
Weight budgets
Scripts, images, CSS and fonts are each checked against a budget, so you know which one to cut first.
Grading
Score penalties ramp between good and poor, so a 2.6 s reading can't sink your grade the way a 6 s one does.
Largest Contentful Paint
Interaction to Next Paint
Cumulative Layout Shift
First Contentful Paint
Time to First Byte
Finding → fix
AI writes each fix for your actual markup and files. It ships in the PDF, right next to the finding.
Detected
hero.jpg · above the fold
Why it hurts
The browser waits for layout before fetching it, and downloads 4× the pixels it paints.
Hurts LCPFix · in your PDF
html<link rel="preload" as="image" href="/hero-1200.avif"
fetchpriority="high">
<img src="/hero-1200.avif" width="1200" height="630"
fetchpriority="high" decoding="async" alt="…">Fetched first, sized right
Where it lands
Every vital labelled with its source, plus AI fixes.
Sort crawled pages by LCP in the crawl explorer.
Weekly or monthly runs catch regressions.
Export the raw numbers for your own tooling.
FAQ
First Input Delay (FID) only measured the delay before the browser started processing the very first interaction on a page. INP evaluates all user interactions (clicks, taps, keypresses) throughout the entire session lifecycle and reports the worst-case interaction latency. Google officially replaced FID with INP in March 2024.
When Google has real-user (CrUX) data for your page or site, BugViso reports that 28-day p75 value, the same field data PageSpeed Insights shows. Without it, BugViso falls back to a PageSpeed mobile lab run, then to its own Chromium measurement, and labels every metric with its source. The Slow 3G and Fast 3G runs are deliberate stress tests, so expect them to be slower than PageSpeed's default profile.
INP needs real taps, clicks and keypresses, so the most reliable value is CrUX field data, which BugViso uses whenever Google has it. In the lab, BugViso measures what drives INP: main-thread tasks longer than 50 ms, Total Blocking Time, and the script each long task is attributed to.
The primary cause of CLS is dynamically loaded images and embeds without explicit dimensions. Always specify width and height attributes (or CSS aspect-ratio) on all images and video containers so the browser can reserve the necessary layout space before the asset finishes downloading.
Yes. TTFB is the starting line for every paint metric: if your server takes 1.5 seconds to send the first byte of HTML, LCP can't be faster than 1.5 seconds. BugViso reports TTFB separately from asset download time, as the median of several loads.
One URL. Results in about five minutes.