AgentReadyWebsite vs Cloudflare’s scanner

At a glance

What each product is

isitagentready.com is Cloudflare’s free Agent Readiness scanner, launched April 17, 2026 during Agents Week (launch blog). It requests the site, scores emerging standards across Discoverability, Content Accessibility, Bot Access Control, and Protocol Discovery / API-Auth-MCP, shows Commerce checks informatively, and offers coding-agent prompts for failing checks. The tool also exposes an MCP discovery document at /.well-known/mcp.json and an Agent Skills index. The same checks appear in Cloudflare URL Scanner’s Agent Readiness tab.

AgentReadyWebsite sells a one-time package. The free Pre-Check calculates a website’s Agent Readiness Score. After checkout, a site analysis generates a custom package (13 files) for you to install. For 12 months, Domain Watch verifies those files stay hosted and emails you monthly (see how it works and the files).

Agent readiness score models are not the same

Do not treat the two scoring models as interchangeable.

Where the check surfaces overlap

Cloudflare’s scanner mixes checks for files and signals on the site with checks that need server-side or runtime capabilities. Our package aims to cover the file-and-signal set. The second set is outside what a static-file delivery can provide.

Often satisfiable by a static package (when files are correctly shaped and served)

• robots.txt AI bot rules and Content Signals
• sitemap.xml
• .well-known/api-catalog
• .well-known/mcp/server-card.json
• .well-known/agent-skills/index.json
• A2A .well-known/agent-card.json (our package includes this path)
• .well-known/ucp (placeholder or populated profile)
• Static webmcp-registration.js presence (their live WebMCP runtime check is a separate bar)

Outside a static-file package (their score can still fail after install)

• Markdown content negotiation (Accept: text/markdown)
• Link response headers (RFC 8288)
• DNS for AI Discovery (DNS-AID)
• Web Bot Auth
• OAuth discovery / OAuth Protected Resource / Auth.md as live endpoints
• Live MCP tools / WebMCP runtime registration
• ARD manifest served as a live discovery endpoint
• Agentic commerce protocols that require payment or checkout infrastructure (x402, MPP, ACP)

On July 7, 2026 we scanned our own domain on their tool: checks for files our package covers passed; failures were in the server-side / runtime group above. That result is dated evidence for that scan, not a guarantee for any site.

One update worth stating plainly, because it cuts against us. As of September 27, 2026 this website does serve Markdown content negotiation: curl -H "Accept: text/markdown" https://agentreadywebsite.io/ returns Markdown with Vary: Accept, and you can verify that yourself in one command. We did it with edge middleware on our own hosting, which is server-side work. Your installed package does not include it, because a zip of files cannot change how your server answers a request header. We are telling you this so you do not read our own result as something the package delivers.

llms.txt: a real standards disagreement

Cloudflare’s default Content Accessibility check is Markdown content negotiation. As of September 27, 2026, their Customize checklist still has no llms.txt toggle (the April 2026 launch blog still says you can customize the scan to include it, that claim is outdated compared with what their tool shows today). Their homepage meta tag description still mentions llms.txt. This gap has now held across three checks of their live tool, on July 11, July 20, and September 27, 2026.

AgentReadyWebsite ships llms.txt as a foundational package file. Google Chrome Lighthouse Agentic Browsing audits for llms.txt. That is a genuine difference of opinion between Cloudflare’s preferred content answer and Lighthouse’s audit, not a trick question, and not something either vendor resolves for the other.

What we ship that their primary content check does not score

Their Content Accessibility dimension scores Markdown negotiation. Several files in our package are scored in our model and documented on the files hub, but are not their default content check:

• llms.txt, foundational discovery index (Howard / llmstxt.org format)
• llm-info.html, agent-facing business briefing
• agenticweb.md, agent-capability narrative
• schema.jsonld, Schema.org structured data (augment-if-present)
• .well-known/agent-permissions.json, action and resource policy

Sources and last verified

isitagentready.com facts on this page were re-verified September 27, 2026 against their live homepage and Customize panel, and previously in July 2026, alongside the Cloudflare launch blog and /.well-known/mcp.json. AgentReadyWebsite facts come from our published product details (tiers, score dimensions, package composition).

What changed since July 20, 2026. Their Customize panel now names three site-type presets, All Checks, Content Site, and API / Application, and you pick one in their interface rather than declaring it on your own page. Their Protocol Discovery group now also lists OAuth Protected Resource (RFC 9728), Auth.md, and an ARD manifest (agenticresourcediscovery.org), and their Commerce group lists MPP alongside x402, UCP, and ACP. What did not change: there is still no llms.txt check anywhere in their tool.

• isitagentready.com
• Introducing the Agent Readiness score (Cloudflare Blog, April 17, 2026)
• Cloudflare Radar AI Insights (adoption charts; refresh before quoting percentages)
• FAQs, free-scanner entry (category-level) and score definition