llms.txt Generator & Checker — and the honest truth about who actually reads it

llms.txt is a proposed standard (from Answer.AI, published at llmstxt.org) for a single markdown file at your site root that gives AI tools a curated map of your content. Think of it as a robots.txt for meaning rather than permission: instead of telling bots what they may crawl, it tells them what actually matters and where to find the clean version.

The format is deliberately small. In order:

• An H1 with the project or site name. This is the only element the spec strictly requires.
• A blockquote summary (>) — one line of key context.
• Zero or more H2 sections, each a markdown link list: - [name](url): optional note.
• An optional section literally named ## Optional — its links can be skipped when a shorter context is needed.

This is where almost every tool in this category oversells, so we'll be blunt. As of 2026, the evidence says llms.txt is not an AI-search or ranking lever:

• Google says it doesn't use it. Google's John Mueller stated plainly that no AI system Google runs uses llms.txt, comparing it to the long-dead keywords meta tag; Google Search does not support it and has no plans to.
• Most files go unread. Independent server-log analysis (Ahrefs, across ~137,000 domains) found roughly 97% of llms.txt files received zero requests — and where requests did happen, the top requester was other SEO audit tools, not AI crawlers.
• The big AI crawlers honor robots.txt, not llms.txt. OpenAI's and Anthropic's crawlers read your robots.txt; they don't consume the llms.txt standard to decide what to fetch.

So who does use it? Coding and IDE agents and documentation tools — Cursor, Mintlify, GitBook and similar — treat llms.txt as a token-saving index into docs. That's a genuine, if niche, benefit, and it's the honest reason to keep one.

Bottom line: publishing llms.txt will not get your brand mentioned by ChatGPT or ranked in Google's AI answers. If that's the goal, you need a different approach — see the AI Visibility section below.

A common point of confusion: llms-full.txt is not part of the llms.txt spec. It's a convention (popularized by Mintlify) for a second file that concatenates your entire documentation into one markdown blob for direct ingestion. The index (llms.txt) points at pages; the full file is the pages. Tools that present llms-full.txt as "the standard" are overreaching — keep your llms.txt a lean index and only publish a full dump if a doc tool you use asks for one.

Four rules keep an llms.txt useful to the tools that do read it:

1. Lead with the H1 and a real summary. The first line is your site name; the blockquote is one honest sentence about what you do. That's the context an agent reads first.
2. Curate — don't dump. The most common quality complaint about generated files is over-inclusion: every URL, bloated to hundreds of lines. Link the pages that matter (docs, key product pages, guides), grouped under sections.
3. Use real link syntax and keep it reachable. Every item is - [name](url), and every link should return 200 — broken links mislead the agents that follow them.
4. Serve it as text. The file must return text/plain or text/markdown — not your app's HTML shell. An HTML page at /llms.txt is the #1 real-world failure.

Our generator applies all four automatically: it reads your sitemap and each page's own title and description, curates out the noise, and emits a spec-clean file — deterministically, with no AI guessing.

1. Enter your domain and click Check llms.txt. We fetch /llms.txt (and /llms-full.txt) and check whether it's present and served as markdown.
2. Read the honest verdict. If it's present, we validate the structure against the spec — grading only the H1 as required — and sample your same-site links for reachability.
3. Generate a clean one. Absent or weak? Generate a curated starter from your sitemap, then copy or download it and save it at your site root.