LIVE|TBT|v.1

Diagnostic Web Vitals

Pass TBT in every preview build

Always-on lab monitoring of Total Blocking Time. TBT is the Lighthouse proxy for INP - the only interactivity signal a preview deploy can produce before real users touch it. Sentry catches main-thread regressions in CI before they ship to field CrUX. Cortex handles the fix.

sentry.tbt.live● 2 min ago
03:30:00GET https://capconvert.com/
03:30:00PSI: mobile + desktop · Lighthouse lab
03:30:01Parsing 7 TBT rules...
03:30:02 PASS mobile_lab_tbt (148ms ≤ 200ms)
03:30:03 PASS main_thread_work (2.1s ≤ 3s)
03:30:04 PASS long_tasks_under_5 (2 long tasks)
03:30:05 PASS js_execution_time (1.4s ≤ 2s)
03:30:06 WARN third_party_main_thread (612ms > 500ms)
03:30:07 PASS dom_size_acceptable (842 ≤ 1500)
03:30:08 PASS unused_javascript_low (18% ≤ 30%)
03:30:09 Score 6/7 · Grade A · optimized
Page Speed

Continuous Main-Thread Monitoring

Continuous lab audits of mobile + desktop TBT against Lighthouse's 200ms threshold and the documented diagnostics that explain main-thread cost. TBT is the metric Lighthouse pays attention to when scoring Performance because INP can't be measured in lab. Sentry catches what fails. Cortex fixes it.

RULE · 1

mobile_lab_tbt

Mobile lab TBT ≤ 200ms

Lighthouse Total Blocking Time on mobile under 200ms. Sums every main-thread task longer than 50ms between FCP and TTI - the chunks that block user input from being processed. The Lighthouse-recommended lab proxy for INP.

RULE · 2

main_thread_work

Main-thread work ≤ 3s

Lighthouse mainthread-work-breakdown total under 3 seconds. Every millisecond of main-thread work past 50ms is a long task that becomes TBT - and a candidate INP-blocking event for real users.

RULE · 3

long_tasks_under_5

Fewer than 5 long tasks

Lighthouse long-tasks audit reports fewer than 5 tasks > 50ms on the throttled mobile profile. Each long task is one potential interaction stall. Five is the practical cap Lighthouse uses when it tags a page as TBT-heavy.

RULE · 4

js_execution_time

JavaScript execution ≤ 2s

Lighthouse bootup-time audit shows total JS execution under 2 seconds on the throttled mobile profile. Long bootup time is the single largest TBT contributor on JavaScript-heavy SPAs.

RULE · 5

third_party_main_thread

Third-party main-thread ≤ 500ms

Third-party scripts (ad tags, analytics, chat widgets, A/B tools) consume less than 500ms of main-thread time. Third parties are the #2 TBT contributor after first-party JS - and unlike first-party, you can't tree-shake them.

RULE · 6

dom_size_acceptable

DOM size ≤ 1500 elements

Total DOM tree under 1500 elements. Style recalculation and layout work after JavaScript scale with DOM size - large DOMs inflate TBT even with optimized scripts.

RULE · 7

unused_javascript_low

Unused JavaScript ≤ 30%

Lighthouse unused-javascript audit shows less than 30% of total JS bytes go unused on page load. Unused JS is parse + compile + execute cost paid for nothing - and a documented TBT regression source.

Page Speed

Free TBT Checker

Paste your homepage URL. Sentry runs PSI for mobile + desktop, pulls Lighthouse lab TBT, main-thread breakdown, bootup time, long-task count, third-party summary, DOM size, and unused-JS audits, then ships a per-rule report. No signup, instant results, always free.

Comprehensive auditInstant resultsCompletely free
Instant

Audit in under a minute

Sentry fetches your site, runs every TBT rule, and renders the full result page before your next sip of coffee.

Actionable

Every failure gets a fix

Each failed rule ships with a prescription paragraph. Hand it to engineering and the gap is closed before lunch.

Ongoing

Locked in for the long haul

Add your site to the daily Sentry sweep with one click. New regressions get caught the next morning.

7 rules in the TBT Sentry. Daily 3:30 AM ET sweep.

Optimize Main-Thread Work

Stop Guessing. Start Seeing. Get Cortex.

One brain. Thirty-six pairs of eyes. Sentries monitor every visibility signal that decides whether search engines, AI engines, and ad platforms show you. Cortex reads what they see, weighs it against a unified corpus of platform documentation, and acts. Every move follows a defined decision protocol: action stated, reason given, impact named.

250
Ranking signals
30
Sentries
60
Platforms
Daily
Always-on
llms.txtai-citationsai-crawlersbrand-pulsetitle-tagsmeta-descstructuredsitemapcore-vitalspage-speedaccessibilitydomain-agehttpsaboutauthorsbacklinksmentionsreviewsinternalgbpnapyelptrusthelpfultopicalfirst-handcronavboostfreshnesshreflangimage-seopage-qualitycanniballlm-outputtrackingssl-tlsCORTEXdecision engine