Agentic BrowsingReadiness Report
Analyze how well any URL is optimized for AI agents, crawlers, and real users. Checks page speed, structured data, accessibility, and semantic HTML.
What this report actually checks
An AI agent does not skim a rendered page the way a person does. It fetches HTML, follows the structure, and decides what the page is and whether it is worth trusting. This report runs four groups of checks against a live fetch of your URL, so every score is based on what a crawler would actually receive.
Access and crawlability
Before anything else, an agent has to be able to get the page. This group verifies the URL is reachable, served over HTTPS, indexable rather than noindex, declared with a canonical URL, and discoverable through a working robots.txt and XML sitemap. A page that fails here is invisible regardless of how good the content is.
Structured data and content
This group checks whether the page describes itself: a real title, a meta description, exactly one H1, JSON-LD structured data, and an llms.txt file at the root that explains your site structure in plain markdown. It also measures visible word count, because a thin page gives an agent nothing to summarise.
Interaction and accessibility
Agents and screen readers share a problem: controls with no name. This group verifies that buttons and links carry accessible names, that form fields have real label elements, that link text describes its destination, and that the page uses semantic landmarks instead of unlabelled divs.
Performance and stability
Core Web Vitals decide whether an agent waits for your page. This group measures CLS, LCP and interaction responsiveness through the PageSpeed API, and separately verifies that your content actually exists in the served HTML rather than being injected only after hydration.
The four failures we see most often
- Content injected by JavaScript. The page looks complete in a browser but arrives nearly empty over HTTP. Anything important needs to be in the server-rendered HTML.
- Missing or duplicated H1. The H1 is how an agent decides what the page is about. Zero or several of them makes the topic ambiguous.
- Unlabelled interactive elements. Icon-only buttons and generic link text like "click here" give an agent nothing to act on.
- Layout shift on load. Images and embeds without explicit dimensions push content around, which raises CLS and can abort a crawl mid-read.
How to use the score
Treat the score as a priority order, not a grade. Fix everything in Access and Crawlability first, because those failures block the rest from mattering. Then clear Structured Data and Content, which is usually the largest block of missing points on a typical site. Interaction and Performance matter most once the page is actually reachable.
The most common high-impact fix is also the cheapest: put the main content in the HTML, give the page one H1, and add a meta description. Those three changes resolve a large share of failed checks without touching layout or code.