Is your site agent-ready?

It measures how easily AI agents and answer engines (ChatGPT, Perplexity, Claude, Google AI Overviews) can discover, read, and act on your website. Our free checker scores 0–100 across four areas — Discoverability, Content, Bot Access Control, and API/Auth/MCP/Skills — based on emerging standards like robots.txt, sitemaps, Markdown negotiation, llms.txt and the .well-known discovery files.

AEO is optimizing your content and site so AI answer engines cite you directly in their responses, rather than just ranking you in a list of blue links. It combines structured, extractable content with machine-readable discovery signals so an LLM can find, trust and quote your pages.

SEO optimizes for classic search rankings. GEO (Generative Engine Optimization) and AEO optimize for being surfaced and cited inside AI-generated answers. They overlap — good structure and crawlability help all three — but AEO/GEO add machine-readability (llms.txt, Markdown negotiation, schema, .well-known discovery) that traditional SEO ignores.

We fetch your site and run 17 equal-weight checks: robots.txt and sitemap validity, Link headers, DNS-AID, Markdown content negotiation, AI-bot rules, Content Signals and Web Bot Auth, and the .well-known files (api-catalog, OAuth metadata, auth.md, ARD ai-catalog, A2A agent card, MCP server card, Agent Skills index, WebMCP) — plus bonus signals like llms.txt and JSON-LD structured data. Each check passes or fails; the share you pass becomes your 0–100 score and Level 0–3.

llms.txt is a plain-text/Markdown file at your site root that summarizes your key content and links for large language models — like robots.txt, but for LLMs. It helps AI systems ingest your most important pages efficiently. It's quickly becoming a baseline agent-readiness signal, so yes, most sites should publish one.

It means returning a clean Markdown version of a page when a client sends the header Accept: text/markdown, while still serving HTML to browsers. Agents parse Markdown far more reliably than rendered HTML, so supporting it makes your content dramatically easier for LLMs to read and quote accurately.

Standardized JSON/Markdown documents under /.well-known/ that let agents discover your capabilities: api-catalog (RFC 9727) lists your APIs, mcp/server-card.json advertises an MCP server, agent-skills/index.json lists agent skills, and the OAuth metadata files describe authentication. Serving these lets agents interact with your site programmatically.

auth.md is a Markdown file at your site root that tells agents how to authenticate and register for your APIs, paired with an agent_auth block in /.well-known/oauth-authorization-server. It's the emerging convention (from WorkOS) for letting AI agents securely obtain scoped credentials without a human filling in a sign-up form.

MCP (Model Context Protocol) is an open standard for exposing tools and data to AI models. An MCP server card at /.well-known/mcp/server-card.json advertises your server's name, transport endpoint and capabilities so agents can discover and connect to your tools automatically.

Add explicit User-agent groups in robots.txt for crawlers like GPTBot, ClaudeBot, PerplexityBot, Google-Extended and CCBot, and layer on Content-Signal directives to declare preferences for AI training, search and AI input. Our checker flags whether you've declared any preference at all.

Content Signals are directives (e.g. Content-Signal: ai-train=no, search=yes, ai-input=no) that declare how AI systems may use your content. They give you granular control beyond simple allow/disallow, separating training from search from answer-generation use.

A soft-404 is when /sitemap.xml returns HTTP 200 but serves an HTML page (usually your app shell) instead of real XML — so crawlers and agents can't parse it. Fix it by generating a valid /sitemap.xml with the correct application/xml content type and referencing it from robots.txt.

Publish answer-first, well-structured content with clear headings, FAQs and JSON-LD; make it machine-readable via Markdown negotiation and llms.txt; and ensure discovery through a valid sitemap and .well-known files. Fastlook generates AI-citable pages and implements these signals so answer engines can find, trust and quote you.

Indirectly, yes. Valid sitemaps, structured data, fast crawlable pages and clear content help both classic Google ranking and AI answer inclusion. Agent-readiness adds machine-readability on top — so you win in traditional search and in AI answers at the same time.

Fastlook implements every standard this checker grades — llms.txt, Markdown-for-Agents, .well-known discovery, MCP, auth.md and structured data — and continuously publishes AI-citable content optimized for answer engines. Run the free scan, create a free account, and we take your score toward Agent-Native.

Yes — the agent-readiness scan is completely free and needs no account. You only sign up when you want Fastlook to implement the fixes and start improving your AI visibility.