The llms.txt Generator writes a spec-valid file, not a dump.

An llms.txt generator is a tool that writes llms.txt, the proposed markdown file at a website’s root that lists its most useful pages for large language models and AI agents. The RankX AI llms.txt Generator reads your sitemap, keeps about fifty pages, writes each note from your own meta descriptions and checks the finished file against the spec.

llms.txt, named for the large language models (LLMs) it serves, was proposed by Jeremy Howard of Answer.AI in September 2024 and revised as version 2 in August 2026. The file is deliberately small: one H1 with the site name, which is the only required element, an optional blockquote summary, optional prose, and then H2 sections whose entries are markdown links with optional notes. It is a suggestion to AI agents, not a control, and no AI search engines document reading it. The longer story is in what llms.txt is, and whether you need one.

Not for search visibility, on any evidence published so far. Ahrefs studied 137,210 domains in 2026: 28% publish an llms.txt, and 97% of those files received no requests at all in May 2026. Of the requests that did arrive, the largest share, about 22%, came from SEO audit tools, and AI retrieval bots made 1.1%. No AI bot asked for a file that did not exist. Where llms.txt does work is documentation, where coding agents follow it to the right page.

No. Google’s guide to generative AI features in Search, last updated 10 July 2026, lists llms.txt files among things you can ignore and says you do not need AI text files to appear in Google Search, including its generative AI features, because Google Search itself does not use them. Google’s John Mueller had earlier compared llms.txt to the keywords meta tag. Generate one for the cheap upside if you want it, but expect nothing from Google rankings or Google AI Overviews.

Because the demand is real, the cost is nothing, and an honest generator beats a hyped one. More than a quarter of studied domains publish the file, documentation platforms build one automatically, and the file doubles as a clean, human-readable map of your best pages that coding agents and documentation tools already use. What RankX AI will not do is sell llms.txt as an AI visibility tactic, because the evidence says it is not one. The generator exists so that if you want to generate the file, you get a correct, curated one in about a minute.

The spec is stricter than most generators respect. A valid file has exactly one H1 naming the site, and that H1 is the only required block. After it come, in order, an optional blockquote summary, optional prose with no further headings, and then H2 sections containing markdown lists where each entry is a link in the form of a name and URL, optionally followed by a colon and a note. No H3s anywhere, and each H2 section exists to hold its list of links. A built-in llms.txt validator checks this generator’s own output against those rules before handing it to you.

Less than it did. An H2 named Optional marks secondary links an agent can skip when it needs a shorter context, and the llms.txt Generator puts its overflow there. Version 2 of the proposal, published in August 2026, keeps Optional as a useful convention but removes its mechanical meaning, because the tool that once expanded a file into context and omitted that section is no longer part of the proposal.

A full file usually means llms-full.txt, a documentation-platform convention rather than part of the llms.txt specification, which never mentions it. The difference between llms.txt and llms-full.txt is size: platforms such as Mintlify and Fumadocs build the full one by concatenating every documentation page into one markdown file, so an agent can ingest everything in a single fetch. To create one yourself, export each page as markdown and join them in order. It suits docs sites, not marketing sites, so this tool generates llms.txt only and is not an llms-full.txt generator.

No. A claim circulates that Anthropic honours a Commercial-Use directive inside llms.txt, sometimes expanded into a whole invented directive set with Training-Data and Citation-Required fields. It appears in no version of the specification and in no vendor documentation from Anthropic or anyone else, and it traces back to a single SEO blog post that cites no source. Be careful: AI search summaries repeat it as fact. If you want to control training use, robots.txt tokens are the documented mechanism.

Because curation is the entire point of the format. An llms.txt that mirrors your whole sitemap is just a worse sitemap, and the specification asks for a file small enough to fit in an agent’s context, with the detail behind the links. The generator ranks pages, keeps roughly the best fifty, orders sections commercial pages first and blog last, and writes each link’s note from the page’s own meta description. Dumping every URL is the measured weakness of most competing generators.

The generator reads your sitemap, or discovers it from robots.txt, considers up to 100 of its URLs and fetches at most 60 pages, one request at a time, while honouring your robots.txt. It extracts each page’s title and meta description, groups pages into sections, ranks them, and writes a spec-valid file with your homepage’s description as the blockquote summary. With no sitemap it falls back to your homepage alone. You can edit anything before you publish it, and a repeat run inside 24 hours returns the cached result instantly.

Serve it at the root of your site as /llms.txt, next to robots.txt. The specification also allows a file at a subpath, such as /docs/llms.txt, which then covers the pages under that path, and agents should use the most specific file that applies. Serve it as plain text or markdown, keep it in version control like any other content, and update it when your key pages change.

They answer three different questions. robots.txt says what crawlers may fetch, is an internet standard, RFC 9309, and the major AI vendors document how their bots treat it. sitemap.xml lists every URL you want search engines to index. llms.txt points AI agents at your best pages first, is a proposal with no committed consumers, and enforces nothing. A site can sensibly have all three. The AI Crawler Access Checker tests the first; this generator writes the third.

Yes, and the page above carries a full one. The smallest valid llms.txt is a single line, an H1 such as "# Example Site". A useful one adds a blockquote summary on the next line, then H2 sections such as "## Services", each holding a list whose items are a page name in square brackets, its URL in round brackets, then a colon and a short note, and ends with an "## Optional" section for pages an agent can skip.

Update your llms.txt file whenever the pages it lists change: a new service, a retired product, a moved URL. A stale llms.txt points the few agents that read it at the wrong pages, which is worse than no file at all. Keep it in version control with the site, regenerate after significant content changes, and check it quarterly otherwise. In RankX AI, the daily site signal watch re-reads the file and records when it disappears or its quality score drops.