agenticplug.ai - Agentic Commerce Optimization company

Free · No sign-in · MCP connector

Agent Ready Report

Ask your AI assistant how ready your website is for AI agents.

Add the Agent Ready Report to Claude, ChatGPT or any MCP-compatible assistant, then ask about any public website. It runs 35 checks across agent readiness, AI search foundations and agentic commerce, and explains every result in plain English and technically, with the evidence it found.

Add it to your assistant

What you get

A diagnostic report your assistant can walk you through. It tells you what's in place and what's missing. It doesn't sell you anything.

  • Where you stand, as a tally

    Three counts: agent readiness, AI search foundations and, if you sell online, agentic commerce. For example, "9 of 11 checks passing". Never a grade or a level.

  • Every check explained twice

    What the check is in plain English, why it matters for AI agents and AI search, and exactly how it's tested, with links to the official standard.

  • Evidence, not opinions

    Each finding shows what the scan actually saw: the response code, the robots.txt rule, the header, the structured data. You can verify every line.

  • Blocked isn't failed

    If a site's bot protection turns the scanner away, the check is marked "couldn't check", never a fail, and the report explains what that means for AI agents.

What it checks

Every business is scored against the same agent-readiness checks, because this is where AI search is heading. Checks built on standards that are still drafts are labeled emerging.

Agent readiness

16 scored checks + 1 for information

Whether AI agents can find, read and work with the website. Scored for every business, because this is where AI search is heading.

  • robots.txt

    The file that tells crawlers, including AI crawlers, which parts of the site they may visit.

  • XML sitemap

    A machine-readable list of the site's pages.

  • Link headers for agents

    Signposts on the homepage pointing machines to the site's APIs and documentation.

  • DNS-AIDemerging

    DNS records that announce the business's agent services, signed with DNSSEC.

  • Markdown for agents

    A clean text version of pages for AI tools that ask for one.

  • AI crawler rules

    robots.txt rules that apply to AI crawlers such as GPTBot or ClaudeBot, by name or through rules for all crawlers.

  • Content Signalsemerging

    A robots.txt declaration of whether AI may use the content for search, answers or training.

  • API catalog

    One standard file listing the ways software can connect to the business.

  • OAuth discovery

    Where and how software signs in to the site's services.

  • OAuth protected resource metadata

    Which sign-in service protects the site's tools and data.

  • auth.mdemerging

    A guide for AI agents on registering with the service.

  • MCP server cardemerging

    A description of the business's MCP server, so AI assistants can find its tools.

  • A2A agent card

    A profile of the business's own AI agent, for agent-to-agent conversations.

  • Agent skills indexemerging

    Step-by-step skills the business publishes to teach agents its service.

  • WebMCPemerging

    Actions a page offers to AI assistants built into the browser, checked in a real browser.

  • ARD capability manifestemerging

    One catalogue of everything the business offers to AI agents.

  • Web Bot Authemergingnot scored

    Signing keys for the business's own bots. Only relevant if it runs bots.

AI search foundations

11 scored checks + 2 for information

The technical basics that decide whether AI search engines can read, understand and cite the website today.

  • HTTPS

    A secure connection, with plain HTTP redirected to it.

  • AI search crawler access

    Whether ChatGPT, Claude and Perplexity's search crawlers can actually reach the site.

  • Indexable homepage

    No 'noindex' instruction keeping the page out of search indexes.

  • Content readable without JavaScript

    Whether the text is in the page itself, for crawlers that don't run code.

  • Heading structure (H1–H6)

    A clear outline: a main heading, section headings and no skipped levels.

  • Page title and description

    The headline and summary machines read first.

  • Canonical URL

    The page's single official address.

  • Structured data

    Machine-readable facts about the business and its content.

  • Business identity markup

    Who the business is, linked to its official profiles elsewhere.

  • Page language

    The language the page declares it's written in.

  • Response time

    How quickly the site starts responding.

  • Content-type markupnot scored

    FAQ, product, service or event markup, where it fits.

  • llms.txtnot scored

    A proposed AI summary file. No major AI platform has said it reads one.

Agentic commerce

5 scored checks

Whether AI shopping agents can buy from the business. Scored only for sites that sell online.

  • x402emerging

    Instant payments by AI agents inside the web request, typically in stablecoins.

  • MPP (Machine Payments Protocol)emerging

    Machine-to-machine payments by Stripe and Tempo, as one-time charges or pay-as-you-go sessions.

  • UCP (Universal Commerce Protocol)

    A store profile that lets shopping agents search, build a cart and check out.

  • ACP (Agentic Commerce Protocol)emerging

    OpenAI and Stripe's standard for agents buying on a customer's behalf.

  • AP2 (Agent Payments Protocol)emerging

    A protocol, started by Google and now run by the FIDO Alliance, giving businesses proof of what a customer authorized an agent to buy.

Add it to your assistant

One URL, no sign-in, no API key.

MCP server URL (Streamable HTTP)

https://scan.agenticplug.ai/mcp

Tools: scan_websiteexplain_check

Technical documentation: tools, report format, scoring and limits

  1. Claude

    Available

    • It's listed in Claude's Connectors Directory. Open Customize, then Connectors, select Discover and search for Agent Ready Report, or open the listing below.
    • Select Connect to Claude. It connects straight away, with no sign-in or API key.
    • On Team and Enterprise plans, an Owner first adds it for the organization (Connect for your team), then members connect.

    Agent Ready Report in Claude's Connectors Directory

    Anthropic: get started with connectors

  2. ChatGPT

    Available

    • Developer mode is where your plan offers it (OpenAI documents it for Business, Enterprise and Edu). On Business an admin turns it on in Workspace settings; on Enterprise and Edu, under Settings, then Apps, then Advanced settings.
    • Go to Apps, then Create, paste the server URL above, choose no authentication and scan the tools.

    OpenAI: developer mode and MCP apps in ChatGPT

  3. Claude Code and other MCP clients

    Available

    • Add the server URL as a remote MCP server using Streamable HTTP. For Claude Code:
    claude mcp add --transport http agent-ready-report https://scan.agenticplug.ai/mcp

  4. Meta Muse

    Planned

    • We've submitted it to Meta's Muse connector directory. Muse lists connectors after Meta's review.

Then just ask

  • “How agent-ready is example.com?”
  • “Can ChatGPT and Perplexity read my website? It's myplumbing.com”
  • “My store sells online. Can AI shopping agents buy from mystore.com?”
  • “What is DNS-AID and why does my report say it's missing?”

Questions

What is the Agent Ready Report?
A free connector for AI assistants that scans a public website and reports how ready it is for AI agents and AI search. It gives a pass or a miss for each check, explains what each check is in plain English and technically, and shows the evidence it found.
Is it free? Do I need an account?
It's free and needs no sign-in. Add it to your assistant and ask about any public website.
Does it fix anything on my website?
No. The report is a diagnostic: it tells you where you stand and what each finding means. If you want help acting on it, agenticplug.ai does that work as a service.
What data does it keep?
It reads the public pages and files of the website you ask about. It doesn't ask for or store your name, email or IP address. Each report is cached for 10 minutes so repeat questions don't re-scan the site, your IP address is used for one minute to limit how often scans run, and we keep anonymous usage statistics (the site scanned, which assistant asked, and the results) for three months. The full details are in our privacy policy.
Why can my report differ from my row in the Agentic Web Readiness Index?
The index table comes from a separate weekly scan of each listed site, so a site can change between that scan and your report. Where the two differ, the report shows its evidence for every check.
Why do I get 'couldn't check' on some items?
Some websites or hosting providers turn automated visitors away. The report marks those checks as 'couldn't check', never as a fail, and notes the bot protection, because AI agents that browse from data centers are likely turned away the same way.
Why is my small business scored on things like MCP and OAuth?
Because that's where AI search and agentic commerce are heading. The report scores every business against the same agent-readiness checks so you can see where you stand now. Emerging standards are labeled, and commerce protocols are only scored for sites that sell online.

Want help acting on your report?

The report is free and stays a diagnostic. If you'd like a team to take your site from the report to agent-ready, that's what agenticplug.ai does.

Talk to us