Free scan · No signup · ~40 seconds
One URL in, four scores out: speed, technical SEO, AI search readiness and security. Plus the fixes that matter, ordered by what buys you the most for the least work.
Public sites only · 5 scans per hour · Results cached for 24 hours
Real Core Web Vitals from Google's field data where your site has enough traffic to have any, and a Lighthouse mobile run either way. Plus the things that cause slow pages — render-blocking resources, missing compression, unsized images, legacy image formats.
The unglamorous layer: titles and descriptions, one h1, canonicals, robots.txt, sitemaps, Open Graph, image alt text, accidental noindex tags. Mostly it catches the thing someone switched on during a redesign and never switched off.
Whether ChatGPT, Claude, Perplexity and Google's AI can actually reach your pages, and whether what they find says clearly who you are. Structured data, entity schema, answer-shaped headings, and the emerging conventions — llms.txt and friends — marked honestly as emerging rather than dressed up as failures.
HTTPS enforcement, security headers, and the WordPress-specific exposures that turn up constantly: readable version numbers, an open xmlrpc.php, user enumeration through the REST API, directory listings left on.
Four things. Performance, using real Core Web Vitals field data from Google where it exists and a Lighthouse mobile run where it does not. Technical SEO — titles, descriptions, canonicals, robots.txt, sitemaps, indexability. AI readiness — whether AI assistants can reach and understand the site. And security posture, including the WordPress-specific exposures that account for most of what I find in practice.
Because they do different jobs and conflating them produces false alarms. Blocking GPTBot, CCBot, Applebot-Extended or Bytespider keeps your content out of AI training corpora — a legitimate editorial choice with no cost to your visibility. Blocking ClaudeBot, PerplexityBot, OAI-SearchBot, ChatGPT-User, Claude-User or Google-Extended is different: those are the crawlers that fetch and cite pages when someone asks a question. Blocking them removes you from AI answers entirely. Only the second group is scored as a problem here.
No. Scores start at 100 and issues subtract from it gently — a site in reasonable shape scores in the 70s or 80s and gets told so. Free audit tools that fail everyone to manufacture urgency are obvious to anyone technical, and they cost more trust than they win. If your site is fine, the report will say your site is fine.
Usually 30 to 60 seconds. Most of that is Google PageSpeed Insights, which runs a real Lighthouse test against your site rather than guessing. Results are cached for 24 hours, so a repeat scan of the same domain returns instantly.
There isn't one. The whole report is free — every score, the full crawler table, and the complete prioritised plan, with no address and no account. I'd rather you read the thing, find it useful, and remember who wrote it. If it turns up work you'd rather hand to someone, there's a message box at the bottom. If it doesn't, take the list and fix them yourself — that's a perfectly good outcome.
That's the other half of what I do. There's a message box at the end of the report — tell me what you'd like sorted and I'll come back to you. I read and reply to everything myself, usually within a working day.