AI Readiness Audit. Can AI read your site?
Enter your domain and get a 0-100 readiness score for the whole site: 21 structural checks across up to 12 pages, from crawler access and robots.txt to schema, metadata and content hidden behind JavaScript. Every deduction is named and every fix is spelled out.
Fetched the way a crawler fetches, checked the way an engineer checks.
- Access. We load your homepage twice: once as a normal browser, once with GPTBot's published user-agent string. If the browser gets a page and the bot gets a 403, no AI engine will ever read you. We also evaluate your robots.txt against 15 AI crawler tokens, check that http redirects to https, and that www and non-www do not both serve the site.
- Discovery. robots.txt, the XML sitemap (from robots.txt first, then the standard locations), llms.txt. Then the public WordPress REST API for recent pages, posts and products, which also tells us whether permalinks are readable and hands us the content WordPress actually stores.
- Pages. Up to 12 pages, fetched one at a time with a pause between each. For every page: noindex, title length, meta description, H1 count, canonical link, JSON-LD types, and whether the stored content appears in the raw HTML or only exists after JavaScript runs.
- The score. Severity-weighted deductions from 100, capped per check, with the letter never reading better than the worst finding. The same weights run inside Viz.
Fair questions.
What does the score mean?
Start at 100 and deduct for every finding: 25 for a critical, 10 for high, 4 for medium, 1 for low, with each check capped at three pages so ten missing meta descriptions cannot outweigh one blocked crawler. A means AI crawlers can reach, read and identify the site. F means they are turned away or told to stay out. The letter can never read better than the worst finding on the list.
What exactly gets checked?
21 structural checks in three groups. Access: can an AI crawler fetch your homepage at all, does robots.txt block the crawlers behind ChatGPT, Claude, Perplexity and Gemini, is the homepage indexable, does http redirect to https, do www and non-www agree. Discovery: sitemap, sitemap in robots.txt, llms.txt, readable permalinks. Pages: every audited page loads, is not noindex, has its content in the raw HTML rather than behind JavaScript, and carries a sensible title, a meta description, one H1 and a canonical link. Plus schema coverage: Organization on the homepage, Article on posts, Product on products, FAQ anywhere.
Which pages do you look at?
The homepage, always. Then we ask your site's public WordPress REST API for recent pages, posts and products, which also gives us the content WordPress stores so we can check whether it actually appears in the HTML a crawler receives. If the REST API is closed we fall back to your XML sitemap for URLs. Free audits cover up to 12 pages, fetched one at a time with a pause between each, so your host never sees a burst.
Does it work on sites that are not WordPress?
Mostly. Access, discovery, metadata and schema checks apply to any site, and the score is honest for any site. The fixes are written for WordPress, and the JavaScript-visibility check needs the content WordPress stores to compare against, so it is skipped elsewhere. The report says so when that happens.
Is this exactly what the real crawlers get?
Almost. For the access check we request your homepage with GPTBot's published user-agent string, from our servers rather than OpenAI's IP ranges. Hosts that verify crawlers by IP address may challenge us and let the real bot through. When we can tell that happened we report it as a note, not as a block, and say what to confirm in your CDN.
How is this different from the audit inside Viz?
Same checks, same weights, less information. The Viz plugin runs the audit from inside WordPress, so it knows your reading and permalink settings, which bots your AI Control rules block, your SEO plugin, and the exact content you stored for up to 30 pages. Free Viz plans run that structural audit once per site; Lite re-runs it on demand; Pro adds the full audit with semantic and content-gap passes, up to four runs a month.
Is it really free?
Yes. Three audits a day, no account. Results are cached for ten minutes so re-running the same site does not spend an audit.
Viz runs it with the facts only your site knows.
This tool sees your site from the outside. The Viz plugin runs the same checks from inside WordPress, where it knows your reading and permalink settings, which AI bots your own rules block, your SEO plugin, and the exact content you stored for up to 30 pages. Findings live in your dashboard next to the AI traffic they explain, so a fix shows up as more crawls, more citations and more AI-referred visitors, not as a number on a page you audited once.
Free plans run the structural audit once per site. Lite re-runs it on demand. Pro adds the full audit: is your content answer-shaped, is your brand clear, is it fresh, and which questions people ask that no page even tries to answer, up to four runs a month. See pricing.
The audit says what to fix. Viz shows what it earned you.
AI already reads your site. Viz sees which bots, wins the answers, publishes what is missing, credits the customers AI sends, and hands agents the tools to act.