llms.txt Generator
Create an llms.txt file that gives answer engines a clean map of your site: what you do, and which pages are worth citing. Fill in your summary and links, then copy the result. Everything runs in your browser.
Rendered as the blockquote directly under the title. This is the line an answer engine is most likely to quote, so lead with what you do.
Start a section with ##. Then one link per line as Title | URL | description. The description is optional
but tells an engine why the page matters.
# Example > Example is a hosted analytics tool for small product teams. It tracks events, funnels, and retention, and ships a free tier with no credit card. ## Core pages - [Home](https://example.com/): What the product does and who it is for - [Pricing](https://example.com/pricing): Plans, limits, and what each tier includes - [Docs](https://example.com/docs): API reference and integration guides ## Guides - [Getting started](https://example.com/guides/getting-started): Setup walkthrough in under ten minutes
Save the result as llms.txt at the root of your domain, so it
resolves at yourdomain.com/llms.txt. Everything runs in your
browser and nothing is uploaded.
This checked one thing. Check the whole page.
The free Discoverability report scores any page for Google and for AI answer engines, names every issue it finds, and gives you the fix for each one.
Free · No signup · Takes under a minute
What llms.txt is for
llms.txt is a Markdown file at the root of your domain that tells large language models what your site is and which pages matter. Where robots.txt says what a crawler may fetch, llms.txt says what is worth reading, in a format a model can parse without wading through nav, cookie banners, and footers. The convention comes from llmstxt.org and is read by a growing number of AI tools.
The summary does the heavy lifting
The blockquote under the title is the line most likely to be quoted back when someone asks an assistant what your product does. Lead with what you actually do and who it serves, name the concrete capabilities, and skip the positioning language. A vague summary produces vague citations.
Annotate every link
A bare list of URLs tells a model nothing about which page answers which question. The optional description after each link is what lets an engine pick your pricing page for a pricing question and your docs for an integration question. Sections group related pages so the structure itself carries meaning.
What a real llms.txt example looks like
A working llms.txt opens with an H1 naming the site, a blockquote holding the one-paragraph summary, then H2 sections such as Docs, Products, or Pricing, each a Markdown list of links with a short description after a colon. Sites like Anthropic, Stripe, and Zapier publish theirs at /llms.txt — reading two or three before writing your own is the fastest way to calibrate. The generator on this page emits exactly that llmstxt.org structure, so the output doubles as a template.
Where to place it and how to serve it
Serve the file at the root of your domain — https://yourdomain.com/llms.txt — as plain text or Markdown, the same convention robots.txt follows. Subdirectory locations are not checked. One gotcha our site audits catch constantly: single-page-app hosting that answers every unknown path with the HTML shell, which means /llms.txt returns your app instead of the file and reads as absent to any crawler. After publishing, curl the URL and confirm you get raw Markdown back, not HTML.
It is a hint, not an instruction
llms.txt is a young convention with no guaranteed support, and adoption varies by provider. Treat it as cheap insurance rather than a control surface: it costs one file to publish and may improve how accurately you are described, but it does not restrict how models use your content. For that you need robots.txt rules or authentication. Google has said it does not use llms.txt, so publish it for the AI assistants that do read it, not as a Google ranking lever.
Or generate it from your real site, automatically
Hand-writing the link list means guessing which pages matter. Essel's $10 whole-site audit builds an llms.txt for you from the actual crawl: up to 60 verified, live URLs grouped by section and ordered by how heavily your own site links to them — plus a drift check that flags dead links and missing top pages in any llms.txt you already publish. The free single-page scan at /seo-score is the way in.
Frequently asked questions
llms.txt is a Markdown file at the root of a domain that describes a site for large language models. It carries a title, a one-paragraph summary, and sections of annotated links to the pages worth reading. The convention is defined at llmstxt.org and is intended to give models clean context instead of rendered HTML.
Put llms.txt at the root of your domain, so it resolves at https://yourdomain.com/llms.txt. That is the only location tools check, the same convention robots.txt follows.
No. robots.txt controls which URLs a crawler is permitted to fetch, while llms.txt describes what your site is and which pages are worth citing. They solve different problems and you can publish both. Neither one restricts how a model uses content it has already collected.
Support is not universal and no provider guarantees it, so treat llms.txt as a low-cost hint rather than a ranking lever. The reason to publish one is that it is a single file: if a model does read it, you control the summary it quotes instead of leaving that to whatever it scrapes from your homepage.
State what the product does, who it is for, and the specific capabilities that distinguish it. Name platforms, formats, and features explicitly. This paragraph is the most likely thing to be quoted verbatim when someone asks an assistant about you, so concrete beats clever.
A minimal llms.txt is an H1 with the site name, a blockquote holding a one-paragraph summary, and one or more H2 sections (like Docs or Products) containing Markdown links, each with a short description after a colon. The generator on this page produces exactly that structure, so filling in the form gives you a working example for your own site.
llms.txt is the curated index — a summary plus annotated links — while llms-full.txt inlines the full content of those pages into one large Markdown document so a model can read everything without following links. Most sites only need llms.txt. Add llms-full.txt when your documentation is the product and you want assistants to answer from the complete text.
Yes. Essel's whole-site audit crawls up to 100 pages and generates an llms.txt from the results: live, verified URLs grouped by section and ordered by internal link weight, so the file reflects what your site actually treats as important. It also flags drift in an existing llms.txt — links that died or top pages it omits. The audit is $10 flat, and the free scan at /seo-score is the starting point.
The other half
This tool is one step. Essel runs the pipeline.
Essel researches the keywords you can win, writes GEO-optimized articles grounded in your brand, and publishes them on a cadence — autonomously, so you don't have to.
Try Essel free14-day trial · no card to start