Site Visibility Check API

• Can a crawler find the pages: soft 404 (a made-up address that answers 200), robots.txt rules for search and AI bots, the sitemap, llms.txt.
• One address per page: www and the bare domain, http and https, the canonical link, redirect chains.
• What a machine reads: visible text without JavaScript, title and description, JSON-LD, Open Graph, the language of the page.
• Who gets turned away: AI bots that get a 403 or a challenge at the edge while a browser does not.
• Two findings are marked private: true (a secret leaked in the page, a session cookie without protection). They are in your answer, never on a public page and never in our logs.

POST the address of the site, usually its home page:

POST https://api.aiskills402.com/v1/run/site-visibility-check
content-type: application/json

{"url": "https://your-site.com/"}

The first answer is 402 with the price in the PAYMENT-REQUIRED header. Sign it with any x402 client and send the same body again with PAYMENT-SIGNATURE. The answer is 200 with the report: score, findings[] (each with fix_skill), fixes[] (each of our skills with the findings it fixes or helps with), and report_url, the full report with evidence as a web page.

The free example, with the input schema and a real recorded report, is at https://api.aiskills402.com/v1/run/site-visibility-check/example.

Every refusal comes before the payment is settled, so nothing is taken:

• 422 url_refused or site_refused: only https on port 443 and public DNS names are checked; IP addresses, private names and *.workers.dev are refused.
• 422 redirect_other_site: the address redirects to a different site. Submit that address instead.
• 429 site_rate_limited: a site gets at most 3 checks per hour and 10 per day across all buyers, to protect the site being checked. The answer says when to retry.
• 503 check_unavailable: the checking service did not answer in time. Retry shortly.

The check is run by 402registry, from Cloudflare's network, under its own name, which it never changes or hides:

402registry-check/1.0

It sends at most 21 requests per check (the home page and its redirects, robots.txt, the sitemap, llms.txt, one made-up address, the www and http forms, and a few requests with the names of AI bots), runs no JavaScript, never logs in and never submits a form. What it sends and how to block it is on https://402registry.com/bot.

We keep one log line per check with the host name, the time it took and whether the report was stored; no findings, no evidence. 402registry stores the report so that report_url works; that page is not indexed and is not shown on the public profile of the site. If it could not be stored, report_url is null and the report is only in your answer.

Paying for calls in general: pay-per-call APIs.