AXRAY: See what ChatGPT actually reads on your website

Hey Product Hunt 👋 I built AXRAY because I could not answer a simple question about my own site: when someone asks ChatGPT about it, what does ChatGPT actually get? Search engines run a full browser. AI agents mostly do not. They fetch the HTML and read what is there — no JavaScript, no rendering, no second chance. So a site that looks perfect to Google can arrive as an empty shell. To find out how common that is, I measured 292 popular sites against the same rubric. The median scored 68 out of 100. Only 8 scored above 85. 45 blocked AI crawlers outright, and another 72 half-blocked them without, I suspect, meaning to. That study is public and the raw numbers are on the site. What AXRAY does: - Paste a URL, get a score out of 100 across 70 published checks, in about two seconds. No account, no card. - An ordered fix list — what to change, why it matters, and the exact lines to paste. - A command-line version (npx axray-cli) so it can run in CI. - An MCP server, so you can add it to Claude or ChatGPT and just ask "is my pricing page readable by an agent?" wherever you already work. Building a tool that measures agent experience and leaving it un-callable by an agent would have been embarrassing. Two honest things: - The rubric is published in full, weights included. If you disagree with a check you can see exactly what it cost you and argue with it. - It scores its own site with the same rubric and shows the result. It is not 100. I would genuinely like to know where the rubric is wrong. Scan something and tell me what it got wrong about your page — that is the feedback I want most.