Is your site ready for AI agents?
See what ChatGPT, Claude and other AI assistants can and cannot read on your website. Every finding is something we measured, shown to you, never a guess.
No email needed to see your result.
What it checks
Three questions, ten measurements.
Can AI crawlers get in?
What your robots.txt tells each AI crawler, and what a request identifying as one actually receives. A firewall can turn them away where robots.txt cannot show it.
Can they read what is there?
Whether your content is in the page itself or only appears after JavaScript runs, plus structured data, titles and your sitemap.
Is it built for agents?
llms.txt, markdown versions of pages, an OpenAPI description and an MCP endpoint. Reported as facts; they do not change your score.
The 10 AI crawlers we read your robots.txt for
- GPTBot · ChatGPT (training)
- OAI-SearchBot · ChatGPT search
- ChatGPT-User · ChatGPT browsing for a user
- ClaudeBot · Claude (training)
- Claude-User · Claude browsing for a user
- PerplexityBot · Perplexity
- Google-Extended · Gemini
- Applebot-Extended · Apple Intelligence
- CCBot · Common Crawl
- Bytespider · ByteDance
How it works
Measured, and it shows its working.
If we could not check something, the report says so. “Not checked” and “not present” are different answers, and only real findings move your score.
What does the score mean?
It is access multiplied by readability: how much of the AI world can get into your site, times how readable your content is once there. Great content behind a blocked door does not score well, and neither does an open door onto an empty page.
Is any of this an AI opinion?
No. Nothing in this check asks a language model anything. Every finding is a measurement, and the report shows it: the status code, the line of robots.txt, the number of words in your HTML.
How do you test what a crawler receives?
We request your page three times: as an ordinary browser, and identifying as GPTBot and as ClaudeBot. If the browser gets the page and the crawler is refused, something in front of your site is turning AI away. A site that verifies crawlers by IP address may treat the real ones differently, and the report says so.
What if you could not check something?
The report says "Not checked" and why. That is different from "not present", and it never counts for or against your score.
Is blocking AI crawlers wrong?
No, it is a choice, and the report words it as one. But if you want assistants to recommend you, they have to be able to read you.