AI Answer Readiness
Score any page on whether AI answer engines can crawl, parse and cite it.
The URL you enter is fetched by our server to run the checks. Nothing you paste is stored.
About AI Answer Readiness
Answer engines don't rank pages, they quote them. Before ChatGPT, Perplexity or Google's AI Overviews can quote you, three things have to be true: their crawler has to be allowed in by your robots.txt, your actual words have to exist in the HTML that crawler receives, and the page has to say clearly what it is. This audit checks all three against one URL and gives you a score out of 100 with the reasoning attached.
It fetches your page, your robots.txt and your llms.txt, then runs four scored checks — crawler access, server-rendered content, structured data and core metadata — plus a fifth reported but deliberately unscored section covering llms.txt and freshness. Every one of the thirteen AI user-agents we track gets an individual verdict with the robots.txt line that produced it quoted back at you, so a 'blocked' result is something you can act on rather than argue with.
Nothing here involves a language model. Every check is a parser: robots.txt matched with Google's longest-rule-wins semantics, JSON-LD parsed and inspected for a usable entity type, visible text extracted from the raw HTML before any JavaScript runs. That determinism is why the same page scores the same twice, and why the tool is cheap enough to be free.
Reach for it after a framework migration, after a robots.txt edit, or the first time someone asks why a competitor keeps showing up in AI answers and you don't. The report ends with a link to the specific tool that fixes each failing section, and a copyable link that reproduces the whole report.
Frequently asked
›Which AI crawlers does it check?
Thirteen: GPTBot, OAI-SearchBot and ChatGPT-User from OpenAI; ClaudeBot, Claude-User and anthropic-ai from Anthropic; PerplexityBot and Perplexity-User; Google-Extended; CCBot, which feeds a long list of open models; Bytespider; Applebot-Extended; and Meta-ExternalAgent. Five of those are weighted as majors in the crawler-access score.
›Why does my page score badly when it looks fine in a browser?
Almost always because the content arrives via JavaScript. Most AI crawlers fetch the HTML and never execute scripts, so a client-rendered app hands them an empty shell. The 'Server-rendered content' section measures exactly this — how many characters of readable text exist before any script runs.
›Is llms.txt part of the score?
No, deliberately. Google has documented llms.txt as having no effect on its systems and no major vendor has committed to reading one. We still report whether you have it, because it costs an hour and does no harm — but scoring a site down for the absence of a file nobody reads would be dishonest.
›Does the audit send my page to an AI model?
No. Your HTML is fetched by our server, parsed in memory, scored, and discarded when the request ends. It is never sent to a language model and never written to storage.
›Can I audit a staging or password-protected site?
Not yet — the fetcher only follows public http(s) URLs, and it refuses private and link-local addresses outright. That guard is what stops the endpoint being used as a proxy into someone else's network.