AI Search + Agent Readiness Audit

Can AI assistants find, read and cite your site, and can AI agents discover what it offers? This Actor runs one audit per site and returns one score (0-100), a score per section and a ranked fix list. Paste one site or a bulk list. No API keys, no login.

What does it check?

Five sections, each with a handful of yes/no/partial checks on public files and the page itself:

• AI crawler access: whether common AI search, user-triggered and training crawlers are allowed by robots.txt.
• Readable content and page basics: main text present without JavaScript, title, meta description, canonical, H1, language, indexability.
• Structured data: JSON-LD present and valid, organization or site markup, page-type markup, required fields.
• Discovery files: llms.txt (optional), XML sitemap, sitemap declared in robots.txt.
• Agent readiness: public discovery files for agents, each shown as pass (found and well formed), partial (found but malformed or incomplete), fail (not found) or unknown (could not be read: blocked by robots.txt, access denied or server error). Covered: API catalog (RFC 9727), MCP server card (draft), A2A agent card, OpenAPI description, OAuth authorization server metadata (RFC 8414), OAuth protected resource metadata (RFC 9728) and discovery links in the home page Link header (RFC 8288). All of these only matter if you offer an API, an agent, OAuth or an MCP server, so they are marked optional: missing ones never lower your score, found ones count.

Input

Field	Description	Default
urls	1-100 sites (home page or domain)	required
urlsText	Optional bulk text: many sites separated by new lines, spaces or commas	empty
maxExtraPages	More pages read from the sitemap besides the home page (0-5)	2
language	en or tr (titles and fixes)	en

{ "urls": ["https://example.com"], "maxExtraPages": 2, "language": "en" }

Output

One row per site: score, grade, sections[] (each with score and checks[]: id, status = pass / partial / fail / unknown, optional, evidence, title, fix), fixes[] (ranked, up to 10), aiCrawlers, aiCrawlersBlocked, pagesChecked. A final summary row (also saved as SUMMARY) with the average score and the most common fixes.

unknown means the check could not be measured (for example robots.txt was unreachable). It is left out of the score. The score is a readiness indicator for public signals. It is not a ranking, citation or traffic prediction.

Pricing (pay per event)

See the Pricing tab: one event per site that could be audited. Unreachable sites and sites whose robots.txt blocks the crawler are not charged. You can set a maximum cost per run and the Actor stops when it is reached.

Safety and politeness

At most 22 requests per site, one every 0.5 seconds, own User-Agent, robots.txt respected (a site that disallows the home page is skipped and not charged). Every request and every redirect is checked: private, loopback, link-local and metadata addresses are refused. No cookies, no login, no personal data. Only public files are read.

Full report: https://ruesandora.com