What is llms.txt and how to get your website cited by AI assistants

More and more people ask ChatGPT, Perplexity, Gemini or Copilot instead of typing keywords into Google. Those assistants answer with a summary and, if you are lucky, a link to your website. Generative engine optimisation (GEO) is the work of making your site easy for them to read, trust and quote. One of its simplest pieces is a file called llms.txt.

What llms.txt is

llms.txt is a plain-text Markdown file at the root of a website (for example https://example.com/llms.txt), written for language models rather than for people. It was proposed in 2024 at llmstxt.org as a simple convention:

• a title with the name of the site or company,
• a short summary in a quote block,
• sections with lists of links to the most important pages, each with one line of description,
• optionally, an llms-full.txt with the full text of those pages.

The idea is the same as a sitemap, but for meaning instead of URLs: in a few hundred words, a model learns who you are, what you do and where the reliable details are.

Does it really work?

Be realistic: llms.txt is a proposal, not an official standard, and not every AI company says it reads it. Google has said its search does not need it. But it costs almost nothing to add, some AI tools and agents do fetch it, and writing it forces you to describe your business clearly, which helps every other channel too. Treat it as one piece of a bigger picture.

The steps that matter more

1. Let AI crawlers in. Check your robots.txt: if it blocks GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot or Google-Extended, those assistants cannot read or cite you. Decide consciously which ones to allow.
2. Be one clear entity. Use the same name, description and profiles everywhere: website, LinkedIn, Instagram, App Store, Google Business Profile. Link them all from your site with sameAs in structured data.
3. Add structured data. Schema.org markup for your Organization, the Person behind it, your services, products, apps and FAQs tells machines the facts without guessing.
4. Write quotable content. Short, factual paragraphs that answer one question each (what you do, for whom, where, what the price depends on) are what assistants lift into their answers.
5. Be fast and indexable. AI search engines rely on classic indexes such as Bing and Google. A fast site with a clean sitemap, submitted to Search Console and Bing Webmaster Tools, and pinged with IndexNow when it changes, gets read sooner.
6. Keep it fresh and consistent. Out-of-date prices, old addresses or contradictory profiles make models less confident about you.

How to write your llms.txt

Keep it short and factual. A good structure:

• # Company name
• > One or two sentences: what you do, for whom, where.
• ## Key facts: legal name, location, contact, languages.
• ## Services: one line per service with a link.
• ## Products or ## Projects, with links.
• ## Main pages: about, contact, legal.

You can see a real example in franamadis.com/llms.txt, which this website generates automatically from the same data as its structured data, so the two never contradict each other.

Measuring results

There is no “Search Console for ChatGPT” yet. In practice, ask the main assistants the questions your clients ask (“who builds React Native apps in Spanish?”, “what does company X do?”) every few weeks, check whether they describe you correctly and link to you, and watch referral traffic from chatgpt.com, perplexity.ai or copilot in your analytics.

Summary

llms.txt is a quick win, but AI visibility comes mostly from fundamentals: open access for crawlers, consistent identity, structured data and clear answers. We build all of this into every website we make. See our websites, SEO and AI search service or write to us for an audit of how AI assistants see your business today.