Googlebot renders your JavaScript.
GPTBot, ClaudeBot and PerplexityBot don't.
So a PWA storefront can rank fine in Google and simply not exist in ChatGPT, Claude or Perplexity.
No rank tracker shows that gap.
https://t.co/uZ2khJ64qo
A bilingual Laravel site usually keeps four lists of the same URLs: routes, sitemap, hreflang tags, language switcher. Nothing makes them agree, and when they drift nothing breaks. Search Console tells you weeks later. We built our site so those four lists cannot exist.
One national news site. 36,476,875 requests in 7.4 days.
/llms.txt got 45 of them.
/.well-known/mcp got 0.
/robots.txt got 19,704.
Same window: 1.5% of requests claiming to be Googlebot failed verification. 4% of Bingbot.
A user agent is a claim, not an identity.
The tool that measured it — Python, standard library only, no dependencies. Reverse-DNS confirmation plus the operators' own published IP ranges.
https://t.co/x6Uq0dqo2Y
Full write-up: https://t.co/8IvWvLCiGe
Part 1 of 3 on how NessFlow is built. Parts 2 (Blade, Inertia and Filament) and 3 (millions of log lines in PHP) are already live.
Our public site ships 4,323 bytes of gzipped JavaScript. No CDN, no front-end framework, no headless CMS. It runs inside the same Laravel app as the product it sells. And one of its perfect Lighthouse scores turned out to be wrong. We wrote up the decisions, measured.
Hello world.
Right now, someone is asking an AI for the best provider in your market. The answer cites someone.
We built one engine to measure whether it is you, and everything that decides it.
Crawl. Logs. Rankings. AI answers. Accessibility.