Serfe AI Discovery
Overview shows the current state of public discovery features and optional agent-facing resources after Discovery Checks run.
AI Discovery helps your website communicate with AI-powered search engines and agents through clear crawling signals, structured discovery information, and AI-friendly content formats.
As AI becomes another way people find information online, you can make your public content easier for compatible systems to discover and interpret. Serfe AI Discovery gives you practical, opt-in controls for robots.txt, llms.txt, Markdown versions of eligible content, and HTTP Link headers for verified discovery resources.
This can help compatible services and agents find the public information you choose to expose. It remains one part of your broader site SEO strategy, alongside useful content, technical accessibility, and your established SEO practices.
AI Discovery is a practical technical foundation for Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO). These broader disciplines also depend on useful, authoritative content and each platform’s own systems. This plugin helps with a key part: making the public content and discovery signals that AI agents can access clearer and easier to validate.
What you can configure
You do not have to enable every capability. Start with what suits your website, review the public result, and add more when useful.
robots.txt— Review crawler instructions and sitemap references. They are preferences, not access control.llms.txt— Publish a concise, human-reviewed guide to public content and key URLs. It is an emerging convention, not a universal standard.- Markdown — Offer eligible public pages, posts, archives, products, and supported custom content in a cleaner representation; HTML remains canonical for visitors.
- Link headers — Advertise verified public
llms.txtand XML sitemap resources with RFC 8288 Link headers. - API Catalog — Optionally describe the public WordPress REST API without creating endpoints or permissions.
- Discovery Checks — Review discovery files, Markdown, Link headers, the API Catalog, and optional agent-facing resources.
Advanced capabilities are yours to implement
The plugin does not create an MCP server, OAuth authorization system, Agent Skills index, or Agentic Commerce integration. After you implement one, Discovery Checks can help verify its expected public endpoint or registered integration.
Markdown is disabled by default and only eligible public content is published. You control file publication and can review existing virtual or physical responses before replacing them. The plugin does not send content or telemetry to Serfe.
External services
This plugin does not require, contact, or send site content to any third-party service. Discovery Checks make bounded loopback requests only to public URLs on your own website, such as robots.txt, llms.txt, sitemaps, Markdown responses, and any same-origin agent endpoint that you explicitly configure. Each URL must be publicly reachable without authentication; private, password-protected, or login-restricted sites cannot be validated by these checks.
The optional “Is It Agent-Ready” button opens that external website in a new browser tab only after you select it. The plugin does not make that request or transmit data to the service; review that website’s own terms and privacy policy before using it.
Translations
Serfe AI Discovery includes bundled local translations for Spanish (Spain), Spanish (Argentina), and Portuguese (Brazil). They work immediately when you select one of those WordPress site languages.
When an approved WordPress.org language pack is available, WordPress uses that community-maintained translation in preference to the bundled fallback. You can help translate the plugin or suggest improvements through the official translation projects: