Paste your website URL and the generator reads your homepage and sitemap, picks your key pages, and writes a spec-faithful llms.txt in under a minute. Free, no signup. Edit any line before you copy or download it.
Your llms.txt will appear here.
llms.txt is a proposed web standard: a Markdown file at your site's root (yoursite.com/llms.txt) that tells AI systems what your site is about, which pages matter most, and how to reference them. Where robots.txt tells crawlers what they may access and sitemap.xml lists every URL, llms.txt is a curated briefing written for large language models: a title, a one-line summary, and an annotated list of your most useful pages.
robots.txt
Access rules. Tells crawlers, including GPTBot and ClaudeBot, which paths they may fetch. Says nothing about meaning.
sitemap.xml
An inventory. Lists every indexable URL with a last-modified date so search engines can find and refresh pages.
llms.txt
A briefing. One curated Markdown page that says what the site is, who it is for, and which pages best represent it, written for language models.
AI engines like ChatGPT, Perplexity, and Google's AI features answer buyers with a handful of sources. Retrieval works on fragments, and models routinely misread what a company does from scattered signals. An llms.txt file gives every AI crawler one canonical, unambiguous summary of your site, written by you. It costs nothing, takes minutes, and removes the guesswork from how models describe your product. It is one signal among many, not a magic ranking switch, but it is the cheapest one you will ever ship.
The proposed spec is deliberately simple Markdown, in this order:
Upload the file to your web root so it resolves at yoursite.com/llms.txt, exactly like robots.txt. On most stacks that is a one-line static file; on hosted platforms without root file access (Webflow included) it takes a small proxy or edge worker. Some teams also publish llms-full.txt, a longer version that inlines full page content rather than links.
Honest answer: adoption is ahead of confirmation. Anthropic, Mintlify, Zapier, and thousands of documentation sites publish one; no major engine has formally committed to reading it, and Google has been publicly lukewarm. Our view at Arobis AI, as a team that works on generative engine optimization daily: it is a two-minute, zero-risk hedge on an emerging convention, worth shipping precisely because the downside is nothing and the upside is being legible to every AI system that does look for it. What reliably moves AI visibility is the harder work: the entities, citations, and answer-ready content that our AEO agency builds. Ship llms.txt, then work on the signals engines already read.
At your website's root, so it resolves at yoursite.com/llms.txt, the same location pattern as robots.txt. Subdirectory placements are not part of the proposed spec and AI crawlers will not look for them there.
llms.txt is a curated index: title, summary, and annotated links. llms-full.txt inlines the complete content of your key pages into one large file so models can ingest everything without following links. Most sites start with llms.txt; documentation-heavy products often add the full version.
Some AI crawlers fetch it and the developer-tools ecosystem has adopted it widely, but no major engine has formally committed to using it, and Google has downplayed it. Treat it as a low-cost hedge on an emerging convention rather than a proven ranking input.
No. robots.txt controls crawler access, sitemap.xml enumerates URLs for indexing, and llms.txt explains meaning: what your site is and which pages best represent it. They answer different questions and coexist at your web root.
On its own, marginally at best. AI engines decide what to recommend based mostly on third-party evidence: reviews, citations, comparisons, and how clearly the wider web describes you. llms.txt makes you easier to read; it does not make you more recommended. Pair it with real visibility work to change what engines actually say.
Paste your URL and click Fill from my website. The generator fetches your homepage for the site name and description, reads your sitemap.xml (or your homepage links when there is no sitemap), ranks the URLs, and fetches each page's title and meta description. It keeps the 40 most representative pages, lists pricing, product, docs and similar pages under Key pages, and moves blog posts, tags and legal pages to Optional. Every field stays editable before you copy or download.
The crawler reads only public pages, follows at most five redirects, caps every download, and refuses private or internal addresses. Nothing is stored: the result goes back to your browser and is discarded. If a site blocks bots, the generator says so and you can fill in the fields by hand.
Run the free checker to see how ChatGPT, Gemini, Claude, and Perplexity describe your product today, or go deeper with a free AI visibility audit from the team behind this tool.
Run the Free AI Visibility Checker