01What does this audit check?
Four families of checks on the page: security headers (CSP, HSTS, anti-clickjacking…), server-side performance (redirects, time to first byte, compression, weight), SEO tags (title, meta description, canonical, H1, viewport, alt, structured data) and discoverability — robots.txt, sitemap, Open Graph, plus a still-rare criterion: is the page accessible to AI crawlers (GPTBot, ClaudeBot…) and does the site offer an llms.txt?
Each category is graded and detailed line by line, with a link to the specialized tool to dig deeper: security headers, redirects, TLS certificates, DNS. The global score out of 100 weighs: SEO tags 30, security 25, performance 25, discoverability 20.
02Understanding the criteria
- Security
- Reuses our header analyzer's A+–F grade: CSP, HSTS (≥ 6 months), nosniff, X-Frame-Options/frame-ancestors, Referrer-Policy, Permissions-Policy.
- TTFB
- Time between the request and the first byte of the response, measured from Cloudflare's network. < 300 ms: excellent; > 800 ms: visitors are waiting.
- Title / description
- The title (10–70 characters) and meta description (50–160) are what Google shows in its results: too short they say nothing, too long they get truncated.
- Canonical
- The page's reference URL: prevents variants (parameters, http/https, www) from competing against each other in rankings.
- AI crawlers
- GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot… Blocking them in robots.txt removes the page from AI assistants' answers — a choice to make consciously.
- llms.txt
- An emerging convention: a markdown file at /llms.txt describing the site to LLMs. Still rare, hence its low weight (2 points).
03From the command line
The same URL as this page answers in plain text, and the JSON API returns the full four-category breakdown (open CORS) — pluggable into a CI to watch every deployment:
$ curl "its-my-ip.com/audit?url=https://exemple.fr" score: 78/100 (B) security: A (23/25) seo: B (24/30) $ curl "its-my-ip.com/api/audit?url=https://exemple.fr" { "score": 78, "categories": [ { "key": "security", … } ] }
04Frequently asked questions
How is this different from Lighthouse or PageSpeed Insights?
PageSpeed runs the page in a real browser and measures Core Web Vitals (LCP, CLS, INP): the visitor's visual experience. This audit works server-side, without executing JavaScript: headers, redirects, TTFB, tags, discoverability files. The two are complementary — and our “no JavaScript” view is precisely what most crawlers, AI included, actually see.
My site uses React/Vue and the SEO score is poor. Why?
If the page is rendered entirely client-side, its initial HTML is nearly empty: no meaningful title, no content, no tags. That is exactly what crawlers that don't execute JavaScript see — most AI bots in particular. The fix: server-side rendering (SSR), static generation (SSG), or pre-rendering your important pages.
What does the “AI crawlers allowed” criterion mean?
We read your robots.txt and check whether GPTBot, ClaudeBot, Google-Extended, PerplexityBot or CCBot are blocked. Blocked, your content cannot be cited by ChatGPT, Claude or Perplexity — a growing share of site discovery. It is a legitimate choice (protecting your content) but it must be a conscious one: many sites block them unknowingly, through a firewall or CDN setting.
Why isn't my page found in the sitemap?
We look for the exact URL in the first declared sitemap file (robots.txt, otherwise /sitemap.xml). If your sitemap is an index pointing to sub-sitemaps, the page may be listed without us seeing it: the criterion then becomes a light warning, not an error. What matters most is that the sitemap exists and is declared in robots.txt.
What do you store when I audit a page?
Only what concerns the audited page: its URL, the date, the scores and the per-criterion detail — never your IP address nor any information about you. This data is used to show the evolution between two runs on the same URL; only the last 20 audits of each URL are kept. It is the single exception to our “nothing is stored” rule, mentioned in the footer.
Does a score of 100 guarantee good rankings?
No — it guarantees that technology isn't holding your rankings back. Ranking depends first on content (relevance, freshness, depth) and reputation (inbound links). A high score means the technical side is settled: search engines and AIs can discover, read and present the page properly. The opposite — good content sabotaged by technique — is the most frustrating case, and the easiest to fix.