Google readability
Crawlability, indexation, speed, internal linking, duplicate content, thin pages. The classic technical picture.
Can AI actually read your hotel website?
Most hotel sites look perfect in a browser and arrive empty to a crawler. Tharro checks every page the way an AI assistant does β then tells you which fixes move the score.
Your site was tested by humans, in browsers
A modern hotel website loads a shell, then fills it with JavaScript. That is standard practice and it produces fast, beautiful sites. It also means that anything fetching the page without running a browser gets a near-empty document.
Googlebot renders JavaScript and mostly copes, which is why the problem stays invisible β your search rankings look fine. AI crawlers are far more conservative, and live retrieval, the fetch an assistant makes mid-answer, is the most time-constrained of all. If your content only exists after JavaScript runs, you are betting your AI visibility on the most expensive path being taken every time.
One site, two different audiences
AI readability
Content present in raw HTML, text volume per page, structured data, language coverage, whether key claims survive without rendering.
Checked on every page, not sampled from the homepage
- Rendering Is the content in the raw response, or only after JavaScript.
- Text volume How many characters of real prose a crawler receives per page.
- Structured data Schema.org coverage for hotel, rooms, offers, reviews, FAQ.
- Language coverage Whether translated pages serve translated content.
- Crawl health Status codes, redirect chains, canonicals, orphan pages.
- Content depth Pages with nothing for a model to work with.
- llms.txt and robots What you are telling crawlers, and whether it is current.
Two briefs, not a list of errors
Dev brief
The technical fixes, in priority order, with the score impact of each. Written to be handed to whoever maintains the site.
Content brief
The pages that need words rather than code: what is thin, what is missing, what a model cannot currently learn about you.
An unreadable page cannot be cited
When an assistant recommends a hotel it names a source. Across AI hotel recommendations, a hotel's own website is the cited source in under 10% of cases β OTAs and review platforms take the rest. An unreadable site guarantees that outcome: a page with no text in the response has nothing to cite. The model still recommends you. It just credits someone else for knowing about you.
See how that plays out in AI answers: AI Visibility
Frequently asked questions
Does Google penalise JavaScript-rendered websites?+β
No. Googlebot renders JavaScript and usually indexes the content. The exposure is with AI crawlers, several of which do not render, and with live retrieval, which is time-constrained.
How is AI readability different from a normal SEO audit?+β
An SEO audit asks whether Google can index you. This asks whether a model that does not run a browser receives anything worth reading β and whether there is enough substance on the page to be quoted.
Do I need to rebuild my website?+β
Usually not. Server-side rendering or pre-rendering is often a configuration change in Next.js, Nuxt and similar frameworks. The audit tells you which pages actually need it.
What is a good AI readability score?+β
Above 80 means a crawler receives your content on nearly every page. Below 50 usually means a rendering problem rather than a content one.
How often is the site re-checked?+β
Weekly, so a regression from a deploy shows up in days rather than at the next agency review.
Does structured data compensate for unrendered content?+β
Partially. Schema.org markup in the initial HTML is readable and does help. It summarises your page; it does not replace it.
Still have questions?Contact us
See what a crawler sees
A full readability audit of your site, scored twice, with the fix list in priority order.