# llms.txt

> llms.txt is a proposed Markdown file at a site's root that gives AI tools a short site summary and a hand-picked list of links to its key pages.

- Page: https://methodcrew.com/glossary/llms-txt
- Last updated: 2026-10-28
- Publisher: [The Method Crew](https://methodcrew.com/)

## How it works

Jeremy Howard of Answer.AI proposed the format in September 2024 ([llmstxt.org](https://llmstxt.org/)). The file lives at /llms.txt and has a simple structure:

- An H1 with the site or project name, the only required part.
- A blockquote with a short summary.
- Optional paragraphs with key details.
- H2 sections listing links, each with a short note, plus an optional section that tools may skip when space is tight.

The idea is that a model with a limited [context window](https://methodcrew.com/glossary/context-window) reads one clean file instead of parsing navigation, scripts and cookie banners. Some sites also publish llms-full.txt, a single file with the full text of key pages. That is a community convention; the spec does not define it.

## How businesses use it

The main readers today are developer tools and agents that a person points at a site, for example a coding assistant asked to read a product's documentation. We have not seen any major AI search engine document llms.txt as a ranking or citation input.

Google is explicit about its own Search. Its AI optimization guide says you do not need AI text files or Markdown versions of pages, that "Google Search itself doesn't use them", and that publishing them neither helps nor harms your visibility ([Google AI optimization guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)).

So for most businesses it is a small, cheap addition, worth an hour and no more. We generate it from the same content source as the site so it never drifts from the pages, and keep it to what a buyer or agent needs: services, pricing, contact.

## Where it goes wrong

The main mistake is expecting it to move rankings or AI citations. It changes nothing in Google Search, and an agency selling llms.txt as a GEO tactic is overselling it.

Other common problems:

- Facts or prices in the file that differ from the live pages, which means publishing two versions of your own business.
- Hundreds of links with no notes, which defeats the point of a short index.
- Promotional instructions to AI models written into the file. That is a form of [prompt injection](https://methodcrew.com/glossary/prompt-injection), and it costs trust with anyone who reads it.

## Example (hypothetical)

Hypothetical: a 60-person IT services firm adds an llms.txt that lists its 8 service pages, pricing page and contact page, generated from the same CMS fields as the site. It takes about an hour, and the team sets no Google traffic goal for it, since Google Search does not read the file.

## Related terms

- [Generative engine optimization (GEO)](https://methodcrew.com/glossary/generative-engine-optimization): Generative engine optimization is the work of making your pages likely to be retrieved, quoted and cited in AI-generated answers from ChatGPT, Perplexity, Gemini and Google AI Overviews.
- [Google AI Overviews and AI Mode](https://methodcrew.com/glossary/ai-overviews): AI Overviews are AI-generated summaries shown at the top of some Google results with links to sources; AI Mode is a conversational search view built on the same systems.
- [Schema markup (structured data)](https://methodcrew.com/glossary/schema-markup): Schema markup is structured data, usually JSON-LD, that labels the facts on a page, such as organization, service, price or author, using the shared schema.org vocabulary.
- [AI visibility (share of voice in AI answers)](https://methodcrew.com/glossary/ai-visibility): AI visibility measures how often AI assistants mention or cite your business when answering the questions your buyers ask, compared with competitors over the same set of prompts.
- [Context window](https://methodcrew.com/glossary/context-window): A context window is the maximum amount of text, measured in tokens, that a language model can read and write in one request, counting instructions, documents and its reply.

## Services where this matters

- [SEO and GEO](https://methodcrew.com/services/seo-and-geo): Technical SEO plus pages written to be quoted by ChatGPT, Perplexity and AI Overviews. $3,500 to $15,000 per project.
- [Web design and development](https://methodcrew.com/services/web-design-development): Custom Next.js sites, server-rendered so search engines and AI crawlers read every page. $18,000 to $80,000 per project.

## Articles that use this term

- [How to get your company cited by ChatGPT, Perplexity and AI Overviews](https://methodcrew.com/insights/get-cited-by-chatgpt-perplexity-ai-overviews)

Full glossary: https://methodcrew.com/glossary