Agent Readiness Test

Scan any website to see whether AI agents can find, read, and act on it, with a scored report and copy-paste fixes.

How it works+

What this tool does

The Agent Readiness Test scans a website for the standards and signals that let AI agents find, read, and act on it, from robots.txt rules for GPTBot, ClaudeBot, and PerplexityBot to llms.txt, structured data, markdown negotiation, MCP server cards, OAuth discovery, and agentic commerce protocols. It returns a scored report with specific, copy-paste fixes for every gap.

How it works

  1. Scan: The API fetches your homepage, robots.txt, and a dozen well-known paths in parallel, plus a markdown content-negotiation probe.
  2. Three tiers of checks: Access & discovery (can agents find you and are they allowed in), content readability (can they actually read your pages without JavaScript), and agent capabilities (MCP, API catalogs, OAuth discovery).
  3. Honest scoring: The score is scoped to your site type, so a portfolio isn't penalized for lacking an MCP server or agentic checkout. Emerging standards (llms.txt, Content Signals, Web Bot Auth) award bonus points when present but never count against you, and mature standards that are still rare for your site type (like OAuth discovery on a storefront) score as bonus too.
  4. Fixes: Every failing check includes a specific remediation: a robots.txt snippet, a JSON-LD template, or a ready-to-paste prompt for your coding agent.

Why site type matters

Composite agent-readiness scores are misleading when they grade a portfolio on commerce protocols it will never need. Pick the type that matches your site and the audit adjusts what it expects:

  • Simple Site (portfolio, blog, content): discovery, access, and readability. No login or checkout expectations.
  • Editorial (publisher, news, gated content): everything above, plus Article structured data expectations, paywall markup (isAccessibleForFree) so gating isn't mistaken for cloaking, and OAuth discovery for subscriber access as bonus signals.
  • E-Commerce (storefront, checkout): Product structured data expectations, plus the agentic commerce protocols (x402, UCP, ACP), MCP, and OAuth for account access as bonus signals.
  • SaaS (product, app, API): the full capability tier scored, including API catalog and OAuth discovery, with MCP server cards, Agent Skills, and WebMCP as bonus signals.
Note: This tool uses backend services that may take up to a minute to start up on first use. If your request doesn't process immediately, please wait a moment and try again.