Skip to content
Aback Launch
Guides9 min read

llms.txt Explained: the Format, a Full Example, llms-full.txt and Whether It Works

What llms.txt is, the exact format with a full llms.txt example for a SaaS product, how llms-full.txt and ai.txt differ, an honest look at whether AI engines read it in 2026, and how to make one in ten minutes.

Devvrat Hans

Founder

November 19, 2026
llms.txt Explained: the Format, a Full Example, llms-full.txt and Whether It Works

llms.txt is a Markdown file at the root of a website (yoursite.com/llms.txt) that gives large language models a short summary of the site and a curated list of links to its most useful pages, so an AI tool can understand the site without crawling and parsing every page. It was proposed by Jeremy Howard of Answer.AI in September 2024, and it is still a proposal: no major AI search engine has committed to reading it. You can write one in a few minutes with the free llms.txt generator.

That second sentence matters, because a lot of advice online treats llms.txt like a ranking switch for ChatGPT. It is not. This guide explains what the file is, the exact format with a complete llms.txt example, what llms-full.txt adds, how llms.txt differs from robots.txt and ai.txt, what the evidence says about whether it works, and when it is worth ten minutes of your time (usually, yes, for cheap and specific reasons).

What is llms.txt?

Web pages are built for people and browsers: navigation, cookie banners, scripts, footers. An AI model with a limited context window wastes most of it on that clutter. The llms.txt proposal, published at llmstxt.org, suggests a simple fix: put a clean, human written Markdown file at a known address that says "here is what this site is, and here are the pages worth reading". Our llms.txt glossary entry has the one line definition.

The file is aimed at the moment an AI system is reading your site to answer a question, not at training. Think of it as a sitemap with a cover letter: a sitemap lists every URL for search engines, while llms.txt picks the dozen that matter and says what each one is for.

The llms.txt format

The proposal defines a small, strict structure, in this order:

  1. An H1 with the name of the project or site. This is the only required part.
  2. A blockquote summary: one or two sentences with the key facts someone needs to understand the rest of the file.
  3. Optional Markdown details: paragraphs or lists with more context, but no headings.
  4. H2 sections of links, each a list where every item is a Markdown link, optionally followed by a colon and a short note.
  5. An "Optional" section, if you want one: links an AI tool can skip when it is short on space.

A full llms.txt example for a SaaS product

# Acme

> Acme is invoicing software for freelancers and small agencies. It sends invoices, follows up on late payments automatically, and accepts card and bank payments. Plans start at $12 per month with a 14 day free trial.

Acme works in 30 currencies and integrates with Stripe, QuickBooks and Xero. It is not accounting software: it does not file taxes or run payroll.

## Product

- [How Acme works](https://acme.com/how-it-works): the invoice, reminder and payment flow in five steps
- [Pricing](https://acme.com/pricing): plans, limits and what the free trial includes
- [Integrations](https://acme.com/integrations): Stripe, QuickBooks, Xero and Zapier

## Docs

- [Getting started](https://acme.com/docs/start.md): create an account and send a first invoice
- [Automatic reminders](https://acme.com/docs/reminders.md): schedules, templates and how to pause them
- [API reference](https://acme.com/docs/api.md): REST endpoints for invoices, clients and payments

## Comparisons

- [Acme vs FreshBooks](https://acme.com/compare/freshbooks): who each tool is for

## Optional

- [Changelog](https://acme.com/changelog): release notes since 2024
- [About](https://acme.com/about): team and company history

A few things make this example good rather than just valid:

  • The summary is quotable. It says who the product is for, what it does and what it costs, in words an assistant can repeat without guessing.
  • It says what the product is not. "It is not accounting software" heads off a common wrong answer.
  • Every link has a note. The note tells the model whether a page is worth fetching for the question at hand.
  • The docs link to .md versions. The proposal also suggests serving a clean Markdown copy of important pages at the same URL with .md added. If you do not have those, link the normal pages.

llms.txt vs llms-full.txt

llms.txt is an index: short, with links. llms-full.txt is a different idea that grew up alongside it, mostly on documentation sites: one large Markdown file that contains the full text of every important page, so a tool can load everything in one request with no link following. Documentation platforms such as Mintlify generate both files automatically, which is why you see llms-full.txt on so many developer docs sites.

llms.txtllms-full.txt
ContentsSummary plus curated linksThe full text of the pages
Typical sizeA few kilobytesHundreds of kilobytes or more
Part of the original proposalYesNo, a later convention
Best forAny siteDeveloper docs that people paste into coding assistants

For a marketing site or a small SaaS, llms.txt alone is enough. Add llms-full.txt if you have real documentation that developers will want to load into an AI coding tool in one go.

llms.txt vs robots.txt vs ai.txt

Three files at your site's root, three different jobs. They are often confused, and they do not replace each other.

FileJobWho reads itEnforced?
robots.txtSays which paths each crawler may fetchGoogle, Bing, GPTBot, ClaudeBot, PerplexityBot and most reputable crawlersHonored by reputable crawlers; a formal standard (RFC 9309)
llms.txtSummarizes the site and points to the best pagesSome AI tools and agents, when they look for itNothing to enforce: it is guidance
ai.txtStates which content AI may use for trainingVaries; there are competing versionsNo: a stated preference

robots.txt is the only one with teeth: if you block GPTBot there, OpenAI's training crawler stays out. The robots.txt guide covers the AI crawler tokens in detail. An llms.txt file cannot grant or deny access; if robots.txt blocks an AI crawler, a link in llms.txt does not change that.

ai.txt is a separate proposal about permissions, not content. The best known version, from Spawning in 2023, lists which kinds of media (text, images, code, audio, video) AI models may train on. It is not legally binding and support is thin, but it is a clear place to state your position. The free ai.txt generator builds the file plus matching robots.txt rules for the major AI crawlers.

Does llms.txt work?

Honestly: not as a visibility lever, at least not yet. Here is where things stand in November 2026.

  • Google says in its Search documentation on AI features that you do not need to create new machine readable files or AI text files to appear in AI Overviews or AI Mode. Google Search does not use llms.txt to rank or cite pages.
  • OpenAI documents its crawlers (GPTBot, OAI-SearchBot, ChatGPT-User) and tells site owners to control them with robots.txt. Its crawler documentation does not mention llms.txt.
  • Anthropic and Perplexity likewise document robots.txt controls for their crawlers and have not said their answer engines rank or cite pages based on llms.txt.
  • Server logs tell the same story. Site owners who check them mostly report few requests for /llms.txt from the big AI crawlers, compared with the pages themselves.

So an llms.txt file will not, on its own, get you cited by ChatGPT or Perplexity. What gets you cited is the same as ever: pages that answer the question clearly, that crawlers can reach, and that other sites mention. How to get cited by ChatGPT and Perplexity covers that work, and generative engine optimization (GEO) is the name for the whole practice.

Why it can still be worth adding

  • Agents and coding tools do read it when pointed at it. A developer who pastes your docs URL into an AI coding assistant, or an agent told to "read acme.com and set it up", benefits from a clean map. This is where llms.txt is used most today.
  • It costs ten minutes and cannot hurt. Google's guidance is that you do not need it, not that you should avoid it. Nobody penalizes a site for having one.
  • Writing it is good practice. Forcing yourself to state what the product is, in two sentences, and to pick the ten pages that matter, improves your home page and docs too.
  • It is a cheap bet on the future. If an engine does start reading it, you are already there.

Do not spend a week on it, and do not let anyone sell it to you as a ranking service. Write it, ship it, keep it current when pricing or features change.

How to create your llms.txt

  1. Write the summary first. One or two sentences: what it is, who it is for, what it costs. This is the sentence you want an assistant to repeat about you.
  2. Pick 8 to 20 pages. The ones that answer the questions people ask: pricing, how it works, docs, comparisons, integrations. Skip the blog archive and legal pages unless they are central.
  3. Write a short, factual note for each link. No marketing adjectives.
  4. Build the file. The free llms.txt generator takes your product name, a one sentence summary and your key pages (one per line as Title | URL | note) and gives you a valid file to copy.
  5. Upload it to the root so it loads at https://yoursite.com/llms.txt, served as plain text.
  6. Validate it. The free llms.txt validator reads /llms.txt from your domain and checks that it exists, is plain text, follows the llmstxt.org structure (title, summary, sections, links) and that its links are well formed.

The most common failure the validator catches is an HTML page at /llms.txt: your server returned its 404 page with a 200 status, so the file "exists" but contains your site's not found page. The second is a file with no H1, or with the summary as plain text instead of a blockquote.

To see the rest of your AI readiness (crawler access in robots.txt, llms.txt, schema and page structure) in one report, run the free GEO checker.

FAQ

What is llms.txt used for?

It gives AI tools a short, clean summary of your site and links to its most useful pages, so they can understand it without parsing every page. Today it is used mostly by AI coding assistants and agents that are pointed at a site, rather than by AI search engines.

Does ChatGPT read llms.txt?

OpenAI has not said that ChatGPT uses llms.txt to find, rank or cite pages, and its crawler documentation does not mention the file. ChatGPT may read it if a user or tool points it at the URL, like any other page.

Is llms.txt the same as robots.txt?

No. robots.txt controls which paths crawlers may fetch and is honored by reputable crawlers. llms.txt cannot allow or block anything; it only summarizes the site and points to good pages. You need robots.txt; llms.txt is optional.

What is llms-full.txt?

A single Markdown file with the full text of a site's important pages, so an AI tool can load all of it at once. It is a convention that grew up around documentation sites, not part of the original llms.txt proposal.

Where do I put llms.txt?

At the root of your domain, so it loads at https://yoursite.com/llms.txt. Serve it as plain text, not HTML.

Is there a free llms.txt generator?

Yes. The llms.txt generator on Aback Launch builds a valid file from your product name, summary and key pages, and the llms.txt validator checks the live file once it is uploaded.

Written by

Devvrat Hans

Founder

Share
Newsletter
Get the Monday launch email
The week's new launches and winners, in one short email.

By subscribing, you agree to get our newsletter by email. Unsubscribe any time in one click. See our Privacy policy and Terms.

Launch your startup on Aback Launch

Get a launch page that search engines and AI assistants can read, a place in the weekly launch list, and a review by a real person. Free to join.