Quick answer
llms.txt is a Markdown file at the root of a website (/llms.txt) that gives AI assistants and agents a short summary of the site and a curated list of links to its most useful pages. It is a proposal published at llmstxt.org, not an official standard. Playxoft's free llms.txt Generator builds a valid file from a form, a list of URLs or a pasted sitemap.xml, and validates existing files against the format.
- llmstxt.org format
- Import from a sitemap.xml
- Group pages by path
- Optional section
- Validator with line numbers
- Download llms.txt
How to use the llms.txt Generator
- 1Enter the site's Name and a one-to-three sentence Summary.
- 2Paste your URLs or sitemap.xml under Add pages from URLs or a sitemap and press Add to sections.
- 3Rename the link titles, add short notes, and keep legal pages and archives in the Optional section.
- 4Copy or download the file and publish it at https://your-site.com/llms.txt.
- 5To check an existing file, switch to Validate, paste it, and click a line number to jump to each problem.
What is llms.txt?
llms.txt is a proposal by Jeremy Howard of Answer.AI, first published on llmstxt.org on September 3, 2024, with a second version in August 2026. The idea: when an AI assistant or coding agent needs information from a website, a full HTML page with navigation, scripts and ads is a poor fit for its limited context. A short Markdown file that summarizes the site and points to the pages worth reading gives it what it needs in a few hundred words.
The file lives at /llms.txt, or at a subpath such as /docs/llms.txt to cover one part of a site; where several apply, the most specific one wins. The proposal also suggests offering clean Markdown versions of pages at the same URL with .md added, such as page.html.md.
It is different from robots.txt, which says what crawlers may access, and from a sitemap, which lists every page for search engines. llms.txt is a curated overview, read on demand when an agent needs answers.
What is the llms.txt format?
A file has these parts, in this order:
- An H1 with the name of the site or project. It is the only required part.
- A blockquote with a short summary and the key facts needed to understand the rest.
- Optional notes: paragraphs or lists, but no headings.
- H2 sections of file lists. Each item is a Markdown link, optionally followed by a colon and notes.
# Acme Analytics
> Acme Analytics is a privacy-friendly web analytics service with a REST API.
Every API request needs an API key in the Authorization header.
## Docs
- [Quick start](https://acme.example/docs/quick-start.md): Add the script and see your first page view
- [Events API](https://acme.example/docs/api/events.md): Send custom events from a server
## Optional
- [Changelog](https://acme.example/changelog.md)A section named Optional holds secondary links an agent can skip when it needs a shorter context. Version 2 of the proposal keeps it as a convention and adds two ways for pages to point at the files: <link rel="alternate" type="text/markdown"> for a page's Markdown version and rel="describedby" for its llms.txt, as link elements or an HTTP Link header.
Does Google or ChatGPT use llms.txt?
Be realistic about what it does. Google's guidance on AI features says: "You don't need to create new machine readable files, AI text files, or markup to appear in these features" (Google Search Central). An llms.txt will not raise your rankings or get you into AI Overviews.
Which AI products fetch llms.txt on their own, and how they weigh it, is not settled or well documented. Where it clearly helps is with agents and assistants that are pointed at a site, such as a coding agent reading library documentation. Adoption is real: the llmstxt.org site notes that OpenAI, Anthropic and Google's Gemini API publish llms.txt files for their developer docs, documentation platforms generate them automatically, and Chrome's Lighthouse has an llms.txt audit that flags a file returning a server error (a missing file counts as not applicable).
It also does not control AI crawling. To allow or block AI crawlers, use robots.txt: the robots.txt generator writes those rules.
llms.txt vs robots.txt vs sitemap.xml: what is the difference?
| File | Written for | What it does | Format |
|---|---|---|---|
| robots.txt | Crawlers | Says which paths may be fetched (RFC 9309) | Plain-text rules |
| sitemap.xml | Search engines | Lists every URL you want indexed (sitemaps.org) | XML |
| llms.txt | AI assistants and agents | Summarizes the site and links the pages worth reading | Markdown |
They work together. A sitemap is a good starting point for an llms.txt: paste it into the generator and it groups the URLs into sections by path, ready for you to trim. To build the sitemap itself, use the sitemap generator.
Some documentation platforms also publish an llms-full.txt with the full Markdown text of every page in one file. That is a platform convention, not part of the llmstxt.org format.
How do I write a good llms.txt?
- Lead with facts in the summary. What the product is, who it is for, and anything an agent would otherwise get wrong.
- Curate. Link the 10 to 50 pages that answer real questions, not every URL on the site.
- Write the notes. "Events API: send custom events from a server" tells an agent whether to follow a link; a bare title does not.
- Link Markdown where you can. If your docs can serve
.mdversions, link those. - Use absolute URLs, so the links still work when the file is read on its own.
- Move secondary pages to Optional: changelogs, legal pages, archives.
- Keep it current. Regenerate it when your docs change, and check it with the validator.
Publish it at https://your-site.com/llms.txt and make sure it loads without a login.
Frequently asked questions
Last updated by the Playxoft engineering team.