Ready
Press start. Two requests: one as a browser, one asking for markdown.
- content-type (html)
- —
- content-type (markdown)
- —
- Vary
- —
- alternate link
- —
- llms.txt
- —
Paste a URL and press Start. The checker asks for the page twice, once as a browser and once as an agent that wants markdown, and compares what came back.
Why this matters: Making Next.js agent-readable
Two GET requests to the URL, following redirects, 8 seconds each:
accept: text/html, like a browser.accept: text/markdown, text/html;q=0.9, */*;q=0.8, like an agent that prefers markdown.Then a HEAD for /llms.txt on the same origin, and a look for <link rel="alternate" type="text/markdown"> in the HTML. Bytes are counted after decompression, the way curl | wc -c counts them.
| Readout | When |
|---|---|
| Agent ready | markdown, at most half the HTML |
| Negotiates, barely | markdown, but over half the HTML |
| Partial | HTML, with llms.txt or an alternate link |
| Ships the navbar | HTML, nothing else |
| No signal | refused, failed or timed out |
The checker refuses private and internal addresses and allows ten checks a minute per visitor.