🤖 Agent Readiness Checker
Enter a public website URL (http or https). The checker fetches only public pages — homepage, /llms.txt, /robots.txt, /sitemap.xml — through a fetch proxy and inspects what they actually contain. Nothing is submitted to any AI agent and no login is ever requested.
Paste your website URL
Try it on a site you own first. The proxy refuses private/internal addresses (localhost, 169.254.169.254, RFC 1918 ranges) — see FAQ.
How it scores — and what it cannot prove
Each check is a statically verifiable fact about a public page. Weights add up to 100:
| Check | Points | Pass condition (basis) | Limitation |
|---|---|---|---|
| HTTPS | 15 | Target URL uses https:// | Says nothing about certificate validity beyond the proxy fetch succeeding. |
| llms.txt present | 15 | GET /llms.txt returns 200 with non-empty, non-HTML body | Presence only — content quality is not judged. |
| robots.txt present | 10 | GET /robots.txt returns 200 with non-empty body | Does not validate whether agent crawlers are allowed or blocked. |
| sitemap.xml present | 10 | GET /sitemap.xml returns 200 with XML-looking body | Does not validate every listed URL. |
| Meta description | 15 | Homepage has a non-empty meta description (30–300 chars ideal; shorter still passes with a note) | Only the homepage is inspected. |
| Open Graph tags | 10 | og:title and og:description present (og:image noted, not required) | Only the homepage is inspected. |
| Homepage has substance | 15 | Visible text (scripts/styles stripped) ≥ 200 characters | Heuristic only: a JS shell with server-rendered fallback text, or a thin page, can fool it either way. |
| Mobile viewport | 10 | Homepage has meta viewport | Tag presence only — not a real device rendering test. |
Deliberately not claimed: agent payment success rates, agent traffic, checkout conversion, or any behaviour that would need a real agent transaction, a login, or private analytics. If a tool claims those from a public URL alone, it is guessing.
The proxy fetches at most 2 MB per page, with a 10-second timeout, and refuses private/loopback/link-local/reserved addresses after resolving the hostname through DNS-over-HTTPS (1.1.1.1). It forwards no Authorization or Cookie headers. Per-IP rate limiting is per-Worker-isolate (see Worker notes) and is a basic abuse brake, not a billing-grade limiter.
FAQ
Why does the check go through a proxy?
Browsers block cross-origin fetches to arbitrary sites (CORS). A server-side proxy is the only way a static page can inspect another site's public HTML. The proxy only fetches the four public paths listed above.
Why was my localhost / internal URL refused?
Deliberately. The proxy resolves the hostname via DNS-over-HTTPS and rejects loopback, private (RFC 1918), link-local (including 169.254.169.254, the cloud metadata address), carrier-grade NAT, documentation and reserved ranges, for both literal IPs and resolved hostnames, including IPv4-mapped IPv6. This is SSRF protection; a public checker must never be usable to probe internal networks.
What does Pro unlock?
Pro ($19 one-time, no subscription) unlocks the copy-paste fix pack for the site you just checked — llms.txt template, robots.txt additions, meta snippets — plus the in-panel Re-check button. The score and the main issue list are always free, and you can always re-run a fresh free check from the form above.
Is my data private?
The target URL you type is sent to the fetch proxy so it can fetch that public site; nothing else about you is collected, and no login or payment details are handled by this page except the standard PayPal Pro unlock (Transaction ID redeem, verified server-side by the existing paywall worker).
Does a high score guarantee agents will use or buy from my site?
No. It means the machine-readable basics are in place. Real agent compatibility also depends on structured product data, checkout APIs and terms that this static check cannot verify — that is stated on the results, not hidden.
Feedback or a check we should add?
Tell us which statically verifiable check would help you most.
General information only — not legal, SEO or business advice. Scores are heuristic snapshots of public pages at check time; sites change, and a re-check after you deploy fixes is the only confirmation.