llms.txt is a proposed plain-text file — placed at the root of your website, like yoursite.com/llms.txt — that gives AI systems a curated, markdown-formatted map of your most important content. The idea is simple and reasonable. Whether it actually does anything in 2026 is a different question, and the honest answer is: not much, not yet.

I'm David Campbell, founder of Nerd Stack. We get asked about llms.txt constantly — usually right after a client reads a blog post promising it's the key to ChatGPT visibility. This guide explains what the file is, where it came from, what the evidence actually shows, and our honest recommendation on whether to bother. It's part of our AI search series alongside our complete AEO guide.

What Is llms.txt, Exactly?

llms.txt is a convention, not a law. It proposes that websites publish a file at /llms.txt containing a clean, markdown summary of the site — what the business does, plus links to the pages that matter most — so a large language model can understand the site without crawling and parsing every cluttered HTML page.

The reasoning is sound. A modern web page is mostly noise to an LLM: navigation, cookie banners, scripts, ads, footers. llms.txt is meant to be the signal — a hand-curated table of contents written for machines. Think of it as an executive summary of your website, addressed to AI.

Where Did llms.txt Come From?

The llms.txt proposal was introduced in September 2024 by Jeremy Howard, co-founder of Answer.AI and a well-known figure in the machine-learning community. The specification is maintained as a community project at llmstxt.org.

It's worth being precise here: llms.txt is a community convention — not an official standard ratified by a standards body, and, critically, not something the major AI companies have committed to using. It exists because a respected researcher proposed a tidy idea and parts of the web adopted it. That's the beginning of how good standards happen. It's also how a lot of well-intentioned ideas quietly fade.

What Does an llms.txt File Look Like?

It's just markdown. A minimal example for a web design business:

# Nerd Stack

> Denver web design and development studio building custom websites, e-commerce, and web apps for businesses across the U.S.

## Core Pages

- [Services](https://www.thenerdstack.com/services): Web design, redesign, maintenance, CMS, and custom web apps
- [Pricing](https://www.thenerdstack.com/pricing): Transparent project-based pricing
- [Work](https://www.thenerdstack.com/work): Client case studies and results

## Guides

- [Answer Engine Optimization](https://www.thenerdstack.com/posts/answer-engine-optimization-aeo-guide-2026): The complete AEO playbook
- [Web Design Cost](https://www.thenerdstack.com/denver-web-design-cost): What websites cost and why

An H1 with the site name, an optional blockquote summary, then markdown sections of annotated links. The spec also allows an expanded llms-full.txt containing complete page content. The format is genuinely easy — which is the strongest argument in its favor.

Does Anyone Actually Use It?

This is where the honest version of the article diverges from the hype. Two questions matter: are websites publishing llms.txt, and are AI engines reading it?

Publishing: adoption is real but modest. An SE Ranking study of 300,000 domains found roughly a 10% adoption rate — common among developer-facing software companies, rarer everywhere else.

Reading: this is the problem. Server-log analyses through 2025 and 2026 consistently show AI crawlers almost never request the file. One widely discussed 90-day measurement found that out of tens of thousands of AI-bot requests to a site, only a fraction of one percent went to /llms.txt. The bots overwhelmingly crawl ordinary HTML and ignore the curated file.

The major platforms have been blunt about it. At Google's Search Central Live in July 2025, Gary Illyes said Google does not support llms.txt and has no plans to. John Mueller compared it to the old keywords meta tag — a signal controlled entirely by the site owner, and therefore one search engines learned to ignore decades ago. As Search Engine Roundtable reported, even a brief appearance of an llms.txt file in Google's own developer docs in December 2025 was walked back within the day, with Google clarifying it was not an endorsement. OpenAI and Anthropic have likewise not committed to using it for retrieval or citation.

In fairness, there's a counter-argument. Search Engine Land has pushed back on the "dead on arrival" framing, noting that standards take years to gain traction and that a low-cost, low-risk file is a reasonable bet on the future. That's fair — but it's a bet on the future, not a tactic for today.

llms.txt vs. robots.txt vs. sitemap.xml

These three files get confused constantly. They are not the same thing:

  • robots.txt — tells crawlers what they're allowed to access. Broadly respected, and it's where you allow or block AI crawlers like GPTBot. This one genuinely matters.
  • sitemap.xml — lists every URL you want indexed, with last-modified dates. Search engines actively use it. This one genuinely matters.
  • llms.txt — proposes a curated content summary for LLMs. Aspirational, and largely unused by the engines today.

If you only have time for one of these, it should never be llms.txt. Make sure your robots.txt isn't accidentally blocking AI crawlers, and make sure your sitemap is accurate and current.

So Should Your Website Have an llms.txt File?

Our honest recommendation for 2026: add one if it costs you almost nothing, and expect almost nothing from it.

It's a reasonable, cheap, low-risk hedge. If the standard gains traction, you're ready. If it doesn't, you've lost an hour. What you should not do:

  • Don't pay an agency a premium for "llms.txt optimization." It's a 30-minute task. Anyone selling it as a major service is selling you the hype.
  • Don't expect it to lift your ChatGPT visibility. The evidence simply isn't there. The things that get you cited are real content, structured data, and earned mentions — see our guide on getting cited by ChatGPT and AI search.
  • Don't let it replace fundamentals. An llms.txt file on a slow, JavaScript-rendered site with no structured data is a tidy note pinned to a locked door.

How to Create an llms.txt File

If you'd like the hedge, here's the entire process:

  1. Create a plain-text file named llms.txt.
  2. Start with an H1 of your business name, and an optional one-line blockquote summary.
  3. Add markdown sections — "Core Pages," "Guides," "Contact" — each a short list of links with a few words describing each.
  4. Link only what matters. The point is curation: eight to fifteen of your best, most useful URLs — not your whole sitemap.
  5. Place it at your domain root so it resolves at yoursite.com/llms.txt.
  6. Keep it current. Revisit it whenever your important pages change. A stale llms.txt is worse than none.

That's it. No special tooling, no plugin required.

Frequently Asked Questions

Does llms.txt help with SEO?

No. It has no effect on Google rankings — Google has explicitly said it doesn't use the file — and no measurable effect on traditional SEO. It's aimed at AI systems, and even there the evidence of impact is weak.

Will llms.txt help me show up in ChatGPT?

There's no reliable evidence that it does. AI crawlers overwhelmingly ignore the file and read your ordinary HTML instead. To improve AI visibility, focus on answerable content, schema markup, and earned mentions.

Could llms.txt matter more in the future?

Possibly. Standards take years, and if the major AI platforms adopt it, an existing file becomes an asset. That's the only real reason to add one now — as a cheap hedge, not a current tactic.

Should I block AI crawlers instead?

That's a separate, strategic decision made in robots.txt, not llms.txt. Most businesses that want to be cited by AI should allow crawlers like GPTBot, ClaudeBot, and PerplexityBot. Blocking them removes you from AI answers entirely.

Do I need llms-full.txt as well?

No. llms-full.txt — a file containing your full page content — is optional and even less used than the basic version. For almost every small business it's unnecessary.

Bottom Line

llms.txt is a good idea with, so far, little evidence behind it. It's the kind of thing worth doing in the half-hour it takes — as a low-cost bet that the web might standardize on it — while understanding that in 2026 it is not moving anyone's AI visibility.

If a marketer is pitching llms.txt as the centerpiece of your AI search strategy, treat that as a useful red flag. The real work is unglamorous: fast, server-rendered pages, genuine structured data, content written as clear answers, and a reputation worth mentioning. That's what we build at Nerd Stack — and we'll happily add the llms.txt file too, in the last ten minutes of the project. Get in touch if you want it done right.

Sources: llmstxt.org — The llms.txt specification; Search Engine Journal — Google Says llms.txt Comparable to Keywords Meta Tag; Search Engine Roundtable — Google Does Not Endorse llms.txt Files; Search Engine Land — No, llms.txt Is Not the New Meta Keywords.