What is llms.txt, and how do you create one?

llms.txt is a Markdown file at the root of a website, at /llms.txt, that gives AI tools a short summary of the site and a curated list of links to its most useful pages. Jeremy Howard proposed it in September 2024, and Anthropic, OpenAI and Google publish one for their developer docs.

By PloyThe Ploy teamUpdated October 21, 2026

Where llms.txt stands today

Who publishes it, who checks for it and what Google says about it.

  • AI developer docs

    Published

    Anthropic, OpenAI and Google publish llms.txt files for their developer documentation, at platform.claude.com, developers.openai.com and ai.google.dev/gemini-api/docs.

  • Lighthouse

    Audited

    Lighthouse’s agentic browsing audits include an llms.txt check. It flags a server error, and a missing file counts as not applicable.

  • Proposal

    Proposal

    Jeremy Howard published the proposal at llmstxt.org on September 3, 2024, and updated it to version 2 on August 10, 2026.

  • Google Search

    Not needed

    Google’s guidance on AI features says you don’t need new machine-readable files, AI text files or markup to appear in AI Overviews and AI Mode.

llms.txt template
# Example Co

> Example Co makes scheduling software for clinics. This file lists
> the pages that explain the product, pricing and setup.

Example Co is available in the US and Canada.

## Product

- [Pricing](https://example.com/pricing): Plans, limits and what each includes
- [Features](https://example.com/features): What the product does

## Docs

- [Getting started](https://example.com/docs/start): Set up an account

## Optional

- [Blog](https://example.com/blog): Product news and guides

The format is one H1, a summary and lists of links

The proposal defines a small Markdown structure, in this order. Only the H1 is required. The file usually lives at the site root, but it can also sit under a subpath such as /docs/llms.txt.

  1. One H1

    The name of the site or project.

  2. A blockquote summary

    A short description with the key facts an AI tool needs to understand the rest of the file.

  3. Optional details

    Paragraphs or lists without headings that add context, such as regions served or how to read the links.

  4. H2 sections of links

    Each section is a Markdown list of [name](url) links, each optionally followed by a colon and a short note.

  5. An Optional section

    Links under an H2 named Optional can be skipped when an AI tool needs a shorter context.

ploy.ai/llms.txt is a working example

Ploy’s own file follows the format: an H1, a two-sentence summary and eight H2 sections, from What Ploy does to Company information, with 173 links.

Ploy generates it on each request from the site’s page lists, blog and library, so a new post appears without anyone editing the file. It also has a section on connecting AI assistants through the Ploy MCP.

The homepage points to the file with an HTTP Link header, </llms.txt>; rel="service-doc", one of the discovery signals the agent readiness audit checks.

ploy.ai/llms.txt (abridged, captured October 21, 2026)
# Ploy

> Ploy is an AI-powered web marketing automation platform for growth teams. It helps teams build, optimize, and publish marketing sites, then use analytics, visitor identification, answer-engine optimization, and workflows to turn site activity into growth.

## What Ploy does

Ploy helps growth teams run their marketing website as an ongoing growth channel: build and ship pages, measure traffic, identify visiting companies, improve visibility in AI answers, and run repeatable marketing workflows.

## Use Ploy from your AI assistant (Ploy MCP)

Ploy has a remote MCP server, the Ploy MCP, at https://ploy.ai/api/mcp. ...

- [Ploy MCP overview](https://ploy.ai/mcp): What you can do with Ploy from Claude, ChatGPT and other assistants, grouped by the permissions you grant.
- [MCP server card](https://ploy.ai/.well-known/mcp/server-card.json): Machine-readable endpoint, transport and auth details.

## Product

- [Pricing](https://ploy.ai/pricing): Free, Starter, Pro, and Enterprise plans.
- [Ploybooks](https://ploy.ai/ploybooks): Step-by-step marketing workflows Ploy runs end-to-end.
- [Answer Engine Optimization](https://ploy.ai/features/answer-engine-optimization): Tools for improving AI-search visibility and tracking citations.
...

Create an llms.txt file in five steps.

  1. Choose the pages that matter

    List the pages an AI tool needs to explain your product: pricing, product and feature pages, docs, comparisons and key guides. Leave out tag pages, legal pages and thin archives, or move them under Optional.

  2. Write the H1 and summary

    Use your site or company name as the H1, then a one- or two-sentence blockquote that says what you do and who it is for.

  3. Group links under H2 sections

    Use plain section names such as Product, Docs or Comparisons. Write each entry as a Markdown link followed by a colon and a one-line note.

  4. Serve it at /llms.txt

    Publish the file at your site root so it returns a 200 status with a plain text or Markdown content type. Lighthouse flags a server error.

  5. Keep it current

    Generate the file from your CMS or page list, or review it whenever you add a key page. An llms.txt generator can draft one from your sitemap, but trim the result to the pages that matter.

llms.txt, robots.txt and sitemap.xml do different jobs

All three sit at the site root, and none replaces another. robots.txt controls what crawlers may fetch, sitemap.xml lists what exists, and llms.txt picks what is worth reading.

Questionllms.txtrobots.txtsitemap.xml
PurposeSummarize the site and point AI tools to key pagesTell crawlers which URLs they may fetchList the URLs search engines should crawl
FormatMarkdownPlain-text rules such as User-agent and DisallowXML
Controls accessNoYes, for crawlers that follow itNo
StatusCommunity proposalIETF Proposed Standard, RFC 9309 (2022)Sitemaps protocol supported by major search engines

For AI agents that take actions on a page rather than read it, see What is WebMCP?

Frequently asked

llms.txt questions

We have not found public evidence that llms.txt improves search rankings or citations in AI answers. It is a low-cost way to give AI tools a clean map of your key pages, so we treat it as groundwork rather than a growth lever.
llms-full.txt is a separate convention some documentation sites use for a single file that contains the full text of their pages. It is not part of the llmstxt.org proposal. The proposal instead suggests offering Markdown versions of pages at the same URL with .md added.
No. llms.txt only points AI tools to content. To allow or block specific crawlers, use robots.txt rules for their user agents.
llms.txt helps AI tools find and read your content. WebMCP lets a web page register tools that an AI agent in the browser can call, such as running a calculator. A site can use both.
Load the URL and confirm it returns a 200 status, starts with an H1 and lists working links. Ploy’s agent readiness audit reviews your robots.txt rules, sitemap and Link headers, and can add one to your fix plan.

Sources

  1. The /llms.txt filellmstxt.org (Jeremy Howard)
  2. AI features and your websiteGoogle Search Central
  3. llms.txt auditChrome Lighthouse
  4. RFC 9309: Robots Exclusion ProtocolIETF
  5. Anthropic developer docs llms.txtAnthropic
  6. OpenAI developers llms.txtOpenAI
  7. Gemini API docs llms.txtGoogle
  8. ploy.ai llms.txtPloy
Related guides:What is WebMCP?AI agent readiness auditAnswer Engine Optimization

See how AI agents read your site with an agent readiness audit.

Ploy checks whether AI assistants can find, read and act on your site, then fixes the gaps it can and reports the rest. Each run uses AI credits.