Which agent readiness checker should you use?

Use two kinds of checker, because they answer different questions. A standards scanner such as Cloudflare’s isitagentready.com checks which agent-facing files and protocols your site publishes. A behavioural scan such as the Axidort Score sends an agent to attempt a real journey and reports whether it could finish. Scores from different tools are not comparable.

By the Axidort team · Updated

The two families

Standards scannersBehavioural scans
MethodRequest well-known files and headers and check they are valid.Run an AI agent through a task in a real browser, plus a no-browser read.
Typical checksrobots.txt and AI bot rules, sitemap, Markdown responses, MCP server card, API catalog, OAuth discovery.Find a product, read price and stock, add to cart, reach checkout, note any challenge.
TimeSecondsA few minutes
Best forEngineering teams adopting emerging standards.Commerce and product teams who need to know if an agent can buy.
Blind spotA site can pass and still block agents at a bot challenge.Does not certify support for each protocol.

Cloudflare: isitagentready.com

Cloudflare introduced its Agent Readiness score in April 2026. It requests your site and checks standards in four scored groups: discoverability (robots.txt, sitemap, Link headers), content (Markdown for agents), bot access control (Content Signals, AI bot rules, Web Bot Auth) and capabilities (Agent Skills, API catalog, OAuth discovery, MCP server card, WebMCP). Agentic commerce standards (x402, UCP, ACP) are checked but do not count towards the score. For each failed check it gives a prompt you can hand to a coding agent. Source: Cloudflare’s announcement.

It is the reference point for standards coverage. Read the result with one caveat: several of the checked standards are drafts, and Cloudflare’s own data shows fewer than 15 of the 200,000 most visited domains published an MCP server card or API catalog at the time. Failing those checks does not mean agents cannot use your site today.

Other standards and AI-visibility checkers

Several SEO and AI vendors now offer free checkers that score a single URL on access, discovery and parseability, and some add checks for citations in AI answers. They are useful for a quick read, and their weights differ, so the same site gets different numbers. Record the tool and the date next to any score you report.

Axidort: the Axidort Score

Axidort measures what an agent could do. A browser agent that identifies itself as AxidortAgentCheck searches the site, opens a product, reads price and stock and tries to reach checkout, while a second agent reads the same page without a browser. The result is a score from 0 to 100, seven component scores, a grade and a step-by-step record with screenshots. Ninety per cent of the score comes from behaviour and ten per cent from static signals. The methodology lists the weights and limits.

Its limits: it is built around commerce journeys, it does not scan pages behind a login, and because the agent is a language model a score can move slightly between runs.

Which to run, and in what order

  1. Behavioural scan first. If an agent is challenged or cannot reach checkout, that is the problem to fix, and no file will fix it.
  2. Standards scan second. Close the cheap gaps: AI bot rules, sitemap, structured data, llms.txt.
  3. Protocols last. Add an MCP server, WebMCP or a commerce protocol when you have a use for it.
Disclosure: Axidort publishes this guide and the Axidort Score. Facts about other tools come from their own published material, linked above. If something here is out of date, tell us through the contact details on the agent page.

Frequently asked questions

01Why does my site get different agent readiness scores from different tools?

Each tool checks different things and weights them differently. A standards scanner counts files and protocols. A behavioural scan counts what an agent managed to do. Compare a score only with earlier scores from the same tool.

02Is isitagentready.com the same as the Axidort Score?

No. Cloudflare’s scanner checks which agent-facing standards a site supports. The Axidort Score measures whether an AI agent could find a product, read price and stock and reach checkout on the site.

03Do I need to pass every check?

No. Several checked standards are still drafts with almost no adoption. Fix what blocks agents today first, then add emerging standards where they serve your customers.

See how an AI agent experiences your site. A free scan returns your Axidort Score, seven component scores and the step where the agent stopped.
Want your site made agent ready?

Related guides