Agent-Ready Website Check
AI agents now research, compare and buy on behalf of people. If they can't read your site or work out what they're allowed to do there, they move on. Get a free score out of 100 in about 15 seconds.
Free check
Enter your website and work email. We scan your public files, show your score here, and email you the full report with your top three fixes.
What we check
The score adds up five categories, 100 points in all. Your email report breaks down each one.
Scored with the open-source Agents First rubric by Joshua Baer.
Can an agent find the rules for using your site? We look for an AGENTS.md file, an llms.txt file, and a robots.txt that names the AI crawlers.
Can an agent read your pages and data without scraping? We check whether your site serves markdown when asked, publishes a sitemap, and exposes an OpenAPI description.
Does your robots.txt say clearly what AI systems may do with your content, through a Content-Signal line or rules for each AI crawler, rather than a blanket block?
Can an agent act on your product, not just read about it? We look for an MCP server card, a CLI or SDK, and OAuth discovery so agents can sign in properly.
Does your homepage tell people how to connect their agents, the same way it tells them how to sign up?
About the check
An outside scan of public files, scored with fixed rules. No AI model reads your site.
AI agents now research vendors, compare options and complete tasks on behalf of people. An agent that can't read your site, or can't tell what it's allowed to do there, moves on to a competitor that is easier to work with. Agent readiness is the set of public files and signals that make your site legible to them.
With fixed rules, not an AI model. Clustr's scanner fetches about 17 public files from your site (robots.txt, llms.txt, AGENTS.md, sitemap, OpenAPI, MCP server card, OAuth discovery and your homepage) and scores them out of 100 with the open-source Agents First rubric by Joshua Baer, rubric version 0.7. The same site gets the same score.
It's an outside scan of public files only. Some sites block automated scanners, which hides files that do exist. Most marketing sites also score low on agent capabilities, which is normal unless you offer a product agents can use directly.
We send your full report to it: the five category scores with notes, what's missing, and your top three fixes. The Clustr team also receives your result so we can follow up if it's useful. We ask for a work email because the report is about your company's site.
Book a call to walk through your report and decide what to fix first.
We use cookies
We use cookies, Google Analytics, and business visitor identification (Apollo) to understand how visitors and companies use our site. Privacy Policy