Introduction
I have been building WordPress sites with Elementor for years, and for most of that time the job was the same: make a page that looks good to a human. Somewhere in the last few months, the job quietly changed. The visitor arriving at your site might not be a person at all. It might be an AI agent — Claude in a browser tab, an assistant inside ChatGPT, a shopping agent comparing your services — and it does not see your hero section or your animations — just HTML, JavaScript, and navigation chrome it has to guess its way through.
Elementor noticed. On November 5, 2026, Miriam Schwab — Elementor's Head of WordPress — published a blog post titled "From reading to acting: WebMCP and the agent-ready web." The feature behind it is a new menu in Elementor 4.4 called Agent Ready. The beta has run since October 26, Beta 2 landed November 2, and the release notes put the stable date at November 9 — an estimate, so check your dashboard.
Agent Ready is free, off by default, and genuinely useful — but only if you understand what each of its six pieces does.
What "Agent-Ready" Actually Means
When you or I visit a website, we read the layout. An AI agent reads the code. A modern Elementor page is a lot of code: containers, widgets, inline styles, scripts. Buried inside all of that is the actual content — what the business does, what it charges, how to contact it. Agents have been getting by with scraping and guessing, and the guessing is the problem. They hallucinate your prices, miss your services, and quote pages that no longer exist.
The agent-ready idea flips that around. Instead of making the agent guess, your site hands it a clean, structured summary of itself. It is the same direction the wider web is moving — WordPress is planning native AI agent support for version 7.3, and Google, Microsoft, and OpenAI recently agreed to honour a shared robots.txt signal for AI training opt-outs.
The Six Pieces Inside the Agent Ready Menu
The menu bundles six components, all free and off by default — nothing changes until you turn something on.
1. Readiness Scanner
Start here. The scanner rates your site in three areas — Find, Read, and Act — and suggests fixes. Find is whether an agent can discover your content. Read is whether it can understand it. Act is whether it can do anything useful there, like submit a form. Run it, fix the gaps, rescan. It turns "make my site agent-friendly" into a checklist.
2. llms.txt
This is the piece most people will have heard of. The
llms.txt file is a Markdown index of your site's content, served at /llms.txt. Elementor generates it for you, and you choose which content types go into it. There is also an optional llms-full.txt with the full content rather than just the index. Think of it as a sitemap written for a language model: instead of "these pages exist," it says "these pages exist, and here is what they say." One warning: if you edit the generated llms.txt by hand, the plugin stops updating it automatically when your content changes.3. Markdown Pages
Elementor can serve clean Markdown versions of your pages — the content without the layout, navigation, and scripts. A classic Elementor page is heavy. The Markdown version is light, fast, and unambiguous.
4. Agent Discovery
This one is the most technical, and the easiest to skip over — do not. Agent Discovery publishes the files agents use to find out what your site offers:
agent.json (or agent-skills.md) and an API catalog, served under /.well-known/. That directory is the web's standard place for "ask here before you do anything" files. In practice, this is your site introducing itself. Publishing discovery files costs you nothing and positions your site for the agents that are coming, not just the ones that exist today.5. Bot Access
Read this piece most carefully before touching anything. Bot Access lets you control your
robots.txt separately by three categories — Search, Answer, and Train — globally, per bot, or for custom bots. Those three are not the same permission. Search means "you may crawl me so people can find me." Answer means "you may read my content to answer a user's question." Train means "you may use my content to train your model." Now you can say yes to search, yes to answering, and no to training — or any other combination that fits your business. Allowing a bot to answer questions about your services is good marketing. Allowing it to train on your copyrighted course material is a different decision entirely.6. WebMCP
WebMCP points at the future, and has the most honest limitations. It exposes four read-only tools that a browser-based agent can call directly in the visitor's tab: business information, content search, page structure, and page content. Instead of scraping your page, the agent asks your site for the information through a defined tool. Read-only is the key word. Today, an agent can look things up. It cannot do things yet. Actions such as submitting a form are planned for later, starting with Atomic Forms, and Elementor has not given a date. There is also a real browser limitation. Chrome and Edge are testing WebMCP in an Origin Trial until March 30, 2027. WebKit rejects the standard, and every browser on the iPhone uses WebKit, so iPhone visitors probably cannot reach these tools at all for now. It is also still unclear when agents in ordinary browsers will look for tools before reading a page, and prompt injection — manipulated page content tricking an agent into unwanted calls — is still unsolved, in Schwab's words. Treat WebMCP as an experiment worth running, not a feature to bet a client project on.
How It Fits With the Elementor MCP From 4.3
If you have been following Elementor's AI moves, Agent Ready might sound familiar — and it should, because it is the second half of a story that started in September. Elementor announced its official MCP integration on September 15, 2026, and shipped it in Elementor 4.3 on September 22. MCP — the Model Context Protocol — lets tools like Claude Code, Claude Desktop, OpenAI Codex, and Cursor connect directly to your WordPress site and build real, editable Elementor pages instead of raw HTML pasted into a widget. Access is restricted to site administrators, the connection is per site, and everything the MCP generates is saved as a draft, never published on its own. It works on pages built with the new Atomic elements; classic widget (V3) pages are outside it for now.
Here is the distinction. The Elementor MCP from 4.3 is about you using an AI agent to build your site. Agent Ready in 4.4 is about other people's agents finding, reading, and using your site once it is built. One is a building tool. The other is a welcome mat. To try the beta, enable the beta tester feature, install the latest beta via Dashboard → Updates, activate the Atomic Editor, and turn on the features you want under Elementor → Agent Ready.
What It Costs, and What You Actually Need
Agent Ready is free, and you do not need Elementor Pro for it. Pro is only required for Pro beta features such as the carousel, not for Agent Ready. There are no country restrictions, and Elementor does not overwrite your existing files.
The longer version is about risk, and it is small but real. This is beta software. The release date of November 9 is an estimate, and the beta has already gone through two releases since October 26. Do not test it on a live client site first. Clone the site to a staging copy, run the Readiness Scanner there, turn features on one at a time, and rescan after each fix. The other cost is attention. Agent Ready being off by default means nothing changes until you decide — but your competitors who do turn it on will be the ones agents can read clearly.
What I Would Do First
Here is the order I would work through it in. First, run the Readiness Scanner on a staging copy and fix what it flags in Find and Read — the cheap wins. Second, turn on
llms.txt and the Markdown pages — the lowest-risk pieces, free, and they immediately give agents a clean version of your content. Third, set your Bot Access rules deliberately — Search yes, Answer yes, Train only if you mean it — and write down why, so nobody undoes it by accident. Fourth, publish the Agent Discovery files. Fifth, treat WebMCP as a test: turn it on, watch what happens, and do not build anything that depends on it until the browser picture clears up. And keep notes: when you turn a feature on, note the date and the version.Conclusion
Elementor 4.4's Agent Ready menu is not a flashy redesign or a new widget. It is a recognition that your website's audience now includes machines that read, summarize, and act on what they find. The six pieces — a scanner that grades you, an
llms.txt file that indexes you, Markdown pages that simplify you, discovery files that introduce you, bot controls that let you choose, and WebMCP tools that let agents ask instead of guess — add up to a site that can speak for itself in a language agents understand.It is free. It is off by default. It is still a beta, with real limits — no write actions yet, no iPhone support for WebMCP, and open questions about when browsers will even look for these tools. None of that is a reason to ignore it. Test it on staging this week, turn on the safe pieces, and decide your bot rules on purpose instead of by accident. The web is being rebuilt for agents, one plugin at a time. WordPress is planning its own agent support for 7.3, and Elementor just shipped its second major piece of the puzzle in two months. Start with the scanner. It will tell you exactly where you stand.
Frequently Asked Questions
What is Elementor 4.4 Agent Ready?
Agent Ready is a new menu in Elementor 4.4 that bundles six free features — a Readiness Scanner, llms.txt generation, Markdown pages, Agent Discovery files, Bot Access controls, and WebMCP tools — designed to help AI agents find, read, and use your website.
Is Elementor Agent Ready free?
Yes. All six Agent Ready features are free and you do not need Elementor Pro. Pro is only required for Pro features of the beta, such as the carousel. Everything is switched off by default.
What is llms.txt and why does my site need one?
An
llms.txt file is a Markdown index of your site's content at /llms.txt. It gives AI agents a clean summary of what your site contains, so they do not have to scrape your HTML and guess. Elementor 4.4 generates it automatically, with an optional llms-full.txt holding the complete content.What is WebMCP and does it work on iPhone?
WebMCP lets browser-based AI agents call read-only tools on your site — business information, content search, page structure, and page content — instead of scraping the page. It is in an Origin Trial in Chrome and Edge until March 30, 2027, and WebKit rejects the standard, so iPhone browsers probably cannot use it for now.
How is Agent Ready different from Elementor MCP?
Elementor MCP, shipped in Elementor 4.3, lets you use AI tools like Claude Code and Cursor to build and edit pages on your own site. Agent Ready, in Elementor 4.4, is about other people's AI agents finding, reading, and using your finished site. One helps you build; the other helps your site be found and understood.
Should I enable Agent Ready on a live site?
Test it on a staging copy first — Elementor recommends running the Readiness Scanner there, not on the live site. Then turn features on one at a time and rescan after each change.
---
Sources: Elementor Blog — "From reading to acting: WebMCP and the agent-ready web" (Miriam Schwab, November 5, 2026); Elementor Editor v4.4 Beta release notes (GitHub Discussion #37846); Elementor MCP product documentation (Elementor 4.3, September 2026).
