Site audit
The site audit reads your website the way AI engines do: can their crawlers get in, is there structured data they understand, and is the content easy to quote? It gives an AI readiness score and a list of what to fix, most important first.
The audit itself is free. Only the optional AI analysis of individual pages uses credits.

Run an audit
Section titled “Run an audit”Editors, admins and owners can run an audit. One audit runs per brand at a time.
- Open Site audit and choose Find pages (or New audit for a later one).
- LaCleo reads your robots.txt, sitemaps and internal links, politely, at about one request per second. Most sites take a few minutes; very large sites stop at the limits (5,000 addresses) and say so.
- Choose pages to audit. The pages worth auditing (products, pricing, guides, comparisons…) are ticked for you, up to 200. Pages that rarely matter to AI engines are left out with the reason: sign-in and cart pages, search results, tag archives, files, legal pages, pages your robots.txt blocks, and duplicates. You can tick most of them back in.
- Choose Audit N pages (up to 500). Pages are read one per second, so 500 pages take about 10 minutes. The page updates by itself.
The results
Section titled “The results”- AI readiness score (out of 100): 70% from your pages’ average score and 30% from site-wide checks. 80 and above is good; under 50 needs work.
- Critical and high: fix these first. Medium and low: improvements.
- AI crawlers allowed: how many of the 10 main AI crawlers your robots.txt lets in.
- What to fix: every issue, most important first, with the pages it affects and a Fix: line saying what to do.
AI crawler access
Section titled “AI crawler access”The audit checks your robots.txt for each AI crawler:
| Crawler | Company | Blocking it… |
|---|---|---|
| OAI-SearchBot, ChatGPT-User | OpenAI | Critical: keeps you out of ChatGPT answers |
| PerplexityBot | Perplexity | Critical |
| Claude-SearchBot | Anthropic | Critical |
| Googlebot | Critical: also affects AI Overviews | |
| Bingbot | Microsoft | Critical: Bing and Copilot |
| GPTBot, ClaudeBot, Google-Extended, CCBot | Training crawlers | Medium: a choice about AI training |
Site-wide checks
Section titled “Site-wide checks”- HTTPS: every page should be served over HTTPS.
- llms.txt: a short guide to your key pages, written for AI assistants. Missing is a medium issue.
- XML sitemap and robots.txt.
Page checks
Section titled “Page checks”Each audited page is checked for what AI engines need: a title and meta description, one H1, headings in order, enough content, valid structured data of the right type, no noindex, a canonical address, Open Graph tags, a language and image alt text. Pages built only with JavaScript (almost no text in the page itself) are flagged, because many AI crawlers do not run scripts.
The pages table
Section titled “The pages table”
For each page:
- Score: 100 minus points for each issue (25 per critical, 10 per high, 4 per medium, 1 per low).
- Citability: how quotable the page’s structure is, from 12 checks such as answers up front, question headings, short paragraphs, facts and figures, cited sources, a named author and a date.
- Schema: whether the page has valid structured data of the right type, with the required properties.
- AI: the AI analysis score, once you have run one.
Open a page to see its checks and run the AI analysis. See Page analysis.