--- title: "AI Ready: llms.txt, Markdown pages and content signals for WordPress" url: "https://aeoultimate.com/docs/wordpress/ai-ready/" description: "Make your WordPress site readable to AI: Markdown versions of every page, an llms.txt that leads with your main pages, and content signals. Video + guide." author: "David Chamberlain" published: "2026-10-15T01:22:21+00:00" modified: "2026-10-15T16:03:30+00:00" language: "en-US" site: "AEO Ultimate" --- Source: https://aeoultimate.com/docs/wordpress/ai-ready/ AEO Ultimate for WordPress # AI Ready: make your site readable to AI AI agents and crawlers read your site differently from people. AI Ready gives them clean Markdown versions of your pages, a guide to your site at llms.txt, and a clear statement of how AI may use your content. Watch the walkthrough, or read the guide below. A tour of AI Ready on aeoultimate.com: Markdown pages, llms.txt and content signals, live. Read the transcript **0:00 What AI Ready does** AI Ready makes your site readable to AI agents and crawlers: clean Markdown versions of your pages, a guide to your site at llms. txt, and a clear statement of how AI may use your content. The safe features are on from the start. Anything that lets an agent interact with your site stays off until you switch it on. **0:31 Markdown versions of your pages** When an AI agent asks for Markdown, it gets a clean copy of the page instead of the HTML: fewer tokens, less noise. Add dot M D to any page address to see it. It opens with the title, address and dates, and names the HTML page as the original, so search engines never treat it as a duplicate. **0:56 llms.txt: a guide to your site for AI** llms. txt is a short guide to your site for AI models. It opens with your main pages, your home page and the pages in your menu, so the pages that say what you do are always first. Then come your most recent posts and pages. **1:15 Content signals: how AI may use your content** Content signals tell AI companies what they may do with your content: train their models on it, index it for search, and use it as input, which means reading your page to answer someone's question and citing you. That's the whole point of answer engine optimization, so allow it. Training is your call; we suggest no. The signals go out with every page and in robots dot text. **1:43 Breadcrumbs and discovery headers** Every page also carries breadcrumbs, so agents see where it sits on your site, and Link headers that point them to your sitemap, llms. txt and the other discovery files, even if they only ever load one page. **1:58 Advanced agent features stay off** The advanced features let agents search your site through an API, or use its tools through MCP. They're off by default. Turn one on only when you have a use for it, and set up the IP whitelist first. **2:12 Per-bot controls** Per-bot controls let you block a single crawler outright, or give it a lower rate limit, without touching the rest. Blocking writes a disallow rule to robots dot text and turns its requests away. **2:25 Check your site** After you change these settings, save your permalinks and clear your page cache, then test your site at is it agent ready dot com. Thanks for watching. ## Where is AI Ready in AEO Ultimate? Open **TWT AEO → AI Ready**. Each feature has its own switch and a note on what it does, and the header shows how many are on. The **zero-footprint features** are on from the start: they serve AI-friendly versions of content you already publish and grant no interactive access. Anything that lets an agent query your site stays off until you switch it on. [Image: AEO Ultimate AI Ready screen in WordPress showing 6 of 10 features active, the zero-footprint features section and Markdown Content Negotiation switched on with the /page.md and ?aeo_format=markdown addresses ticked] AI Ready: what AI agents and crawlers get from your site, one switch per feature. ## What does an AI agent see when it asks for Markdown? When an agent asks for a page with an `Accept: text/markdown` header, it gets a clean Markdown version instead of the HTML: fewer tokens, less noise. Agents that do not send that header, or sit behind a CDN that strips it, can add `.md` to any page address (or `?aeo_format=markdown`). Each Markdown version opens with a header giving the title, address, description and dates, and names the HTML page as canonical, so search engines never treat it as duplicate content. Markdown is never stored in a page cache, so every AI crawler fetch is logged in AI Crawler Watch. [Image: The E-E-A-T guide served as Markdown at its .md address: a header with title, url, description, published and modified dates, language and site, then the source address and the page text] A page as an AI agent receives it: a header it can parse, then clean text. ## What is llms.txt, and which pages does it list? `/llms.txt` is a short guide to your site for AI models. It opens with **Main pages**: your home page, then the pages in your header menu, in menu order, whether your theme uses classic menus or a block Navigation. Then come your **Recent pages**, the 20 newest posts and pages not already listed. Pages marked noindex in your SEO plugin, password-protected pages and WooCommerce cart and checkout pages are left out. [Image: aeoultimate.com/llms.txt: the site name and tagline, a link to the knowledge bundle, then a Main pages section listing the home page, Features, the plugin comparison, FAQ, Getting Started and About, each with a short excerpt] llms.txt opens with the home page and the main menu, then the most recent pages. ## Should I let AI use my content as input? Yes, if you want to be cited. **Content signals** state three separate permissions. **AI training** is an AI company using your content to build its models. **AI search indexing** is an AI search engine indexing your pages. **AI input** is an assistant reading your page to answer someone’s question and quoting or citing you, which is what answer engine optimization is for. AEO Ultimate recommends allowing search indexing and AI input; training is your call, and the default is do not train. The signals are sent as a `Content-Signal` header with every page and written into `robots.txt`. You can also have AEO Ultimate write per-bot Allow and Disallow rules into `robots.txt`; leave that off if Yoast SEO, Rank Math or another SEO plugin already manages AI bot access. [Image: Content-Signal Headers and robots.txt settings: AI training set to do not train, AI search indexing allowed, AI input allowed, giving the directive Content-Signal: ai-train=no, search=yes, ai-input=yes] Three separate permissions: training, search indexing, and use as input to an answer. ## What do breadcrumbs and Link headers do for AI agents? **Semantic breadcrumbs** add BreadcrumbList schema to every page except the home page, so an agent knows where the page sits on your site. **Link headers** on every response point agents to your sitemap, llms.txt and other discovery files, even if they only ever load one page. **Vary: Accept** tells CDNs to cache the HTML and Markdown versions separately, so a person never gets the Markdown page and an agent never gets the HTML one by mistake. [Image: Semantic Breadcrumbs switched on, Link Headers (RFC 8288) always on and currently sending links to the sitemap, llms.txt, the API catalog and the agent skills index, and Vary: Accept switched on] Breadcrumbs and Link headers point agents to your discovery files from any page. ## Should I turn on the Agent Search API or MCP? Only when you have a use for them. The **Agent Search API** lets agents search your content through a read-only REST endpoint, and **MCP Integration** publishes a tool manifest so agents can use your site’s tools. Both are off by default. Before you switch on any interactive feature, set up the **IP whitelist** under Security & Rate Limiting, and never give an AI agent admin or write access. ## How do I block one AI crawler? Use **Per-Bot Controls** under Security & Rate Limiting. Set a crawler to **Block** and AEO Ultimate writes a Disallow rule for it into `robots.txt` and answers its requests with 403. To slow one down instead, give it its own rate limit; it overrides the global limit for that crawler only. [Image: Per-Bot Controls table listing GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, Claude-User, Google-Extended, Googlebot and other crawlers with their company, an Allow or Block status and a per-bot rate limit, under a global limit of 60 requests per IP per hour] Block one crawler, or give it its own rate limit, without touching the rest. ## How do I check my site is agent ready? After you change these settings or update the plugin, go to **Settings → Permalinks** and click **Save Changes**, then clear any page cache. Then test your site at [isitagentready.com](https://isitagentready.com/). Say yes to being cited Saying no to AI input asks assistants to leave your content out of their answers. If your goal is to be quoted and credited, allow AI input and AI search indexing, and decide on training separately. Something not adding up? See [Troubleshooting](https://aeoultimate.com/docs/wordpress/troubleshooting/) or ask in the [support forum](https://wordpress.org/support/plugin/twt-aeo-ultimate/).