What Is llms.txt? – Quick Answer
An llms.txt file is a simple, Markdown-formatted text file placed in a website’s root directory (e.g. example.com/llms.txt) that gives large language models (LLMs) a concise, structured overview of a site’s most important content – key pages, documentation, API references, policies. It was proposed following the pattern of robots.txt and sitemap.xml, but with one crucial difference: llms.txt is not an official web standard, has no IETF or W3C ratification behind it, and not every major AI provider has confirmed that it actually uses it.
Where Did llms.txt Come From?
The llms.txt proposal was published by Jeremy Howard, co-founder of Answer.AI and fast.ai, on September 3, 2024, on the Answer.AI blog and on llmstxt.org, which has since become the reference point for the standard. The problem it aimed to solve: LLMs have finite context windows, and converting a full webpage’s HTML (navigation, ads, JavaScript) into LLM-friendly text is laborious and error-prone. This was especially relevant for developer documentation, where models need fast access to the right API references while writing code.
What Is llms.txt Actually Used For?
The practical function of llms.txt is narrower than much of the content written about it suggests. It is not a crawling directive (like robots.txt), nor a complete content index (like sitemap.xml) – it’s more of a context map that an AI system can use when answering a specific question about a given page, or when a developer tool (e.g. an IDE integration) loads documentation into its context.
Typical use cases include:
- Developer documentation: API references and SDK descriptions in a quick, clean format – this is the most common and best-evidenced use case.
- On-demand queries: when a user explicitly asks an AI search tool about a brand or service, and the system checks whether an llms.txt exists.
- Retrieval workflows: some AI tools (e.g. Claude Desktop, Claude.ai) factor the file into their own retrieval logic when a user targets a specific domain.
What it is not good for: it has no effect on Google’s organic rankings, it doesn’t replace robots.txt or sitemap.xml, and on its own it doesn’t guarantee that any AI model will read it at all.
llms.txt vs. robots.txt vs. sitemap.xml
| File | Purpose | Audience | Official standard? |
|---|---|---|---|
| robots.txt | Access control – what crawlers may index | Search engine bots (Googlebot, Bingbot) | Yes (RFC 9309) |
| sitemap.xml | Full site map – fast, efficient discovery | Search engine crawlers | Yes (de facto industry standard) |
| llms.txt | Curated context summary of the most important content | LLMs and AI tools (partial, voluntary support) | No – community proposal |
What Does an llms.txt File Look Like?
The file strictly follows Markdown syntax, since that’s the “native language” of LLMs. The recommended structure is:
- H1 heading – the brand or site name (the only required element)
- Blockquote summary – a one- or two-sentence description of what the site is about
- Context paragraphs – short, plain-text background information
- H2 sections – link collections grouped by topic
- Link lists – in the format
- [Page name](URL): short description
A simplified example:
# AdOps Digital
> Full-service digital marketing agency specializing in SEO, PPC, social media, and AI-driven automation.
## Documentation
- [Services](https://ad-ops.hu/en/services): SEO, PPC, social media, and AI agent development
- [Case Studies](https://ad-ops.hu/en/seo-case-studies): Real client results with numbersWhich AI Systems Actually Use llms.txt?
This is the most important question, and it’s also where a lot of misinformation circulates. As of 2026:
- Anthropic (Claude): publicly confirmed – Claude Desktop and Claude.ai take the file into account in their retrieval workflows.
- Perplexity: publicly confirmed – it fetches llms.txt and uses it to prioritize page selection.
- OpenAI: no official confirmation, only observed correlations in some teams’ retrieval patterns.
- Google: a clear no. Google’s June 2026 Search Central documentation states plainly that Google Search does not use llms.txt – not for rankings, not for AI Overviews. Gary Illyes confirmed in July 2025 that Google doesn’t support it and has no plans to; John Mueller compared it to the discredited keywords meta tag.
How Widespread Is llms.txt in 2026? – The Numbers
An SE Ranking study covering 300,000 domains found an adoption rate of just 10.13% – roughly one in ten sites has an llms.txt file, after nearly a year and a half of industry discussion. More importantly: independent analyses found that 97% of llms.txt files were never fetched by any AI system, and statistical models found no measurable citation effect between having the file and appearing in AI search results.
We cannot fully verify this, since individual AI providers’ crawl logs aren’t public – the figures above are third-party estimates, not official statistics.
Is It Worth Creating an llms.txt File? – AdOps's Take
The reality sits between two extremes. On one hand: llms.txt is not a ranking factor, doesn’t guarantee AI Overview visibility, and current data shows most such files simply aren’t read by anyone. Anyone selling it as a silver bullet is misleading you. On the other hand: creating one takes minimal effort (half an hour to an hour for a well-structured site), it doesn’t hurt, and if support expands in the future (as it already has with Anthropic and Perplexity), your site will be ready.
Our practical recommendation: don’t put this at the top of your priority list. Far more impact comes from building content with clear structure, direct answers, and original data – that’s what Google AI Overview and other AI search tools actually reward. In that process, llms.txt is at best a cheap add-on step, not a substitute for solid AEO/GEO fundamentals.
Step by Step: How to Create an llms.txt File
- Create a plain text file named
llms.txt. - Add your site or brand name as an H1 heading (the only required element).
- Write a one- or two-sentence blockquote summary of what your site does.
- Group your most important pages under H2 sections (e.g. Services, Documentation, Case Studies), and give each link a short description.
- Validate the Markdown syntax, then upload the file to your domain’s root directory so it’s accessible at yourdomain.com/llms.txt.
If you want to publish more comprehensive content as well, some teams also create a supplementary llms-full.txt file containing the full, condensed text content – though this is mainly worthwhile for large documentation sites (developer tools, API providers). For building a well-structured, AI-friendly content strategy more broadly, it’s worth reading how to optimize your website for AI search engines.
Want to appear more often in AI answers?
Request a quote!
Frequently Asked Questions
Does llms.txt replace sitemap.xml?
No. Sitemap.xml serves full-site discovery for search engine crawlers, while llms.txt is a narrow, curated excerpt for AI systems – the two exist in parallel, complementing each other.
Is llms.txt a Google ranking factor?
No. Google has officially confirmed that Search does not use llms.txt – not in traditional results, not in AI Overview.
How long does it take to create an llms.txt file?
For a mid-sized, well-structured website, roughly half an hour to an hour, since the format is simple Markdown and requires no development work.