llms.txt guide: Set it up on your website in 20 minutes for AI search visibility in 2026. Includes ready-to-use templates, the Markdown formatting rules LLMs actually parse, and what real data from 844,000 deployments reveals about how llms.txt affects AI citations versus standard crawling.
What Is llms.txt?
llms.txt is a plain-text file placed at your website root that gives AI language models a structured summary of your site, your services, and your key facts. Think of it as a sitemap for AI agents. Where robots.txt tells crawlers what they can access and sitemap.xml lists every page, llms.txt tells an LLM what your site is about before it reads a single page. For the entity structure that makes AI agents cite your content, see our guide.
The file that 844,000 domains deployed for AI citations produced zero measurable citation benefit across 300,000 sites studied. This is not a failure of the concept. It is a failure of how most sites implemented it. An llms.txt file with generic descriptions and no entity mapping tells AI agents nothing they cannot already extract from your homepage. A properly structured llms.txt file with service summaries, authority markers, and key URLs gives AI agents a structured briefing they can cite immediately.
AI-referred sessions grew 527 percent between January and May 2025. The file most guides claim will capture that growth needs proper structure to produce results. This guide shows you exactly how to format llms.txt for measurable AI citation impact.
How Does llms.txt Work?
llms.txt works by giving AI agents a pre-built summary of your site before they crawl individual pages. When ChatGPT, Perplexity, or another LLM-powered agent needs to understand your business, it first reads your llms.txt file to get a structured overview of what you offer, who you serve, and which pages matter most. The file acts as a briefing document that reduces the computational cost of the AI agent understanding your site. Before deploying llms.txt, ensure your technical SEO foundation is solid, because AI agents cannot extract content from pages that fail basic crawl checks.
What the 300,000-domain study revealed
A study of 300,000 domains that deployed llms.txt found zero measurable citation uplift across the entire sample. The reason was structural. Most files contained generic descriptions that added nothing beyond what the homepage already communicated. Domains that included specific entity markers, service summaries, and authority indicators in their llms.txt files saw modest but measurable improvement in AI citation frequency. The file works as an amplifier of existing authority signals, not as a substitute for them. For why AI-generated content without proper structure fails to get cited, see our guide on why AI copy fails.
llms.txt vs robots.txt vs sitemap.xml
| File | Purpose | Who Reads It | Format |
|---|---|---|---|
| robots.txt | Crawl access rules | Googlebot, AI crawlers, all bots | Plain text directives |
| sitemap.xml | All URLs on the site | Googlebot, Bingbot | XML with URL list |
| llms.txt | Site summary for AI agents | ChatGPT, Perplexity, Claude, AI agents | Markdown-formatted text |
What Should Be Included in llms.txt?
A properly structured llms.txt file contains four sections, each serving a specific function for AI agent extraction. Every section the study identified as contributing to citation uplift is marked below.
Section 1: Site identity
The opening section tells an AI agent what your business is before it reads anything else. Include your business name, one-sentence description, and primary service category. Format as a heading with a single sentence beneath it. This is the most-read section of the file.
Section 2: Core pages
List 5 to 10 of your most important pages with their page titles and a one-line description of what each page covers. Use Markdown link format: [Page Title](URL). Focus on pages that answer the questions your customers ask AI agents. Do not list every page. Quality over coverage.
Section 3: Services and expertise
State what you do in bullet-point format. Include your service categories, your primary market, and your unique positioning elements. AI agents use this section to determine whether your business is relevant to a specific user query. Specificity matters: "B2B SEO copywriting for SaaS companies" signals more relevance than "SEO services."
Section 4: Authority signals
List verifiable credentials that confirm your expertise. Include certifications, years of experience, named client outcomes with permission, and industry recognition. AI agents weight recency and specificity. "Documented 96,000 monthly impressions for B2B clients (2025 to 2026)" carries more weight than "SEO expert."
How to Create and Set Up llms.txt in 20 Minutes
Setting up llms.txt on your website takes four steps and approximately 20 minutes. The setup is technically simple but requires strategic decisions about what to include and what to omit. For the full AI visibility strategy that llms.txt feeds into, see our AIO and GEO guide.
Open a plain text editor. Name the file llms.txt. Use Markdown formatting: # for headings, - for bullet points, [text](URL) for links. Save as plain text, not rich text.
Start with site identity, then core pages, then services, then authority signals. Keep the entire file under 2,000 words. AI agents read the full file but weight the opening most heavily.
Place llms.txt in your website root so it is accessible at yoursite.com/llms.txt. This is the same location as your robots.txt file.
Visit yoursite.com/llms.txt in your browser. If the file loads as plain text, AI agents can read it. If you get a 404, the file is in the wrong location or has the wrong name.
Ready-to-use template for small businesses
# [Your Business Name] [One sentence describing what you do and who you serve.] ## Core Pages - [Homepage](https://yoursite.com/) - What we do and who we help - [Services](https://yoursite.com/services/) - Our primary service categories - [About](https://yoursite.com/about/) - Our story, credentials, and team - [Blog](https://yoursite.com/blog/) - Articles on [your industry topics] ## Services - [Service 1]: [one sentence description] - [Service 2]: [one sentence description] - [Service 3]: [one sentence description] ## Authority - [Credential or certification with year] - [Client outcome with specific metric, with permission] - [Industry recognition or award]
Common mistakes that block AI crawlers
- Wrong file name: the file must be named exactly
llms.txt, notllms.txt.txtorLLMS.TXT - Wrong location: the file must be in the root directory, not in a subfolder
- Rich text format: the file must be plain text. Opening it in Word and saving adds invisible formatting characters
- Generic descriptions: an llms.txt that says "We provide quality services" tells AI agents nothing they cannot already extract from your homepage
- No specific URLs: the file must include links to your core pages. A text-only description without URLs gives AI agents no structured path to your content
Does My Business Need llms.txt?
Most small businesses do not need llms.txt today. The data from 300,000 domains shows zero citation uplift for generic implementations. However, businesses that publish original data, serve information-heavy queries, or compete for AI citation visibility should deploy it now while adoption is still early and the file carries novelty signal weight.
Decision guide
| Business Type | Need llms.txt? | Why |
|---|---|---|
| Publishing original research or data | Yes | AI agents cite original data sources. llms.txt maps your data pages for extraction. |
| B2B SaaS with long buying cycles | Yes | Buyers use AI for vendor research. llms.txt helps your brand appear in comparison queries. |
| Local service business | Optional | Google Business Profile provides stronger AI signals. Deploy only if you publish guides. |
| Ecommerce store | Optional | Product schema provides stronger signals. Deploy only for informational/comparison content. |
| Personal blog or portfolio | No | Zero additional benefit over a well-structured About page and article schema. |
Frequently Asked Questions About llms.txt
llms.txt is a plain-text file at your website root that gives AI language models a structured summary of your site before they crawl individual pages. It works like a briefing document: the AI agent reads llms.txt first to understand what your business does, then navigates to the specific pages listed in the file. The file uses Markdown formatting and is designed to be readable by both humans and AI systems.
llms.txt is not a direct SEO ranking factor. Google does not use llms.txt to determine traditional search rankings. Its value is in AI search visibility: helping ChatGPT, Perplexity, Claude, and other AI agents understand your site well enough to cite your content in generated answers. For traditional SEO, focus on technical SEO, content quality, and backlinks. llms.txt is a GEO tool, not an SEO tool.
Create a plain text file named llms.txt in a plain text editor. Write four sections: site identity, core pages, services, and authority signals. Use Markdown formatting. Upload to your website root directory so it is accessible at yoursite.com/llms.txt. Verify by visiting the URL in your browser. The entire setup takes approximately 20 minutes.
Businesses publishing original research, B2B SaaS companies with long buying cycles, and sites competing for AI citation visibility should deploy llms.txt now. Local service businesses and ecommerce stores can deploy it optionally. Personal blogs and portfolios gain zero additional benefit from llms.txt. The decision depends on whether your customers use AI tools to research your category before purchasing.
Include four sections: site identity with business name and one-sentence description, core pages with 5 to 10 key URLs and descriptions, services with bullet-point summaries, and authority signals with verifiable credentials. Keep the file under 2,000 words. Use Markdown formatting. Generic descriptions produce zero citation benefit. Specific entity markers and named outcomes produce measurable improvement.
Yes, when properly structured with specific entity markers, service summaries, and authority indicators. The 300,000-domain study found zero uplift for generic implementations but measurable improvement for files with specific, structured content. llms.txt amplifies existing authority signals. It does not create authority where none exists. Deploy it alongside strong on-page SEO and content strategy.
Visit yoursite.com/llms.txt in a browser. The file should load as plain text. Monitor your server logs for requests from AI crawler user agents like ChatGPT-User, ClaudeBot, and PerplexityBot to confirm AI agents are accessing the file. Track AI citation frequency using Bing Webmaster Tools AI Performance report which shows how often your content is cited across Microsoft Copilot and Bing AI summaries.
robots.txt controls which parts of your site search engines and AI crawlers can access. llms.txt tells AI agents what your site is about and which pages are most important. robots.txt is for access control. llms.txt is for content briefing. Both files sit in your website root directory and both serve AI agents, but they perform entirely different functions.
Review and update llms.txt every quarter. Update immediately when you add major service categories, publish significant original research, earn new certifications, or achieve measurable client outcomes you can cite with permission. AI agents weight recency signals. A file last updated six months ago signals lower activity than one updated quarterly.
Google has not confirmed using llms.txt as a ranking signal for traditional search. The file is designed for AI language models, not search engine crawlers. Google uses its own discovery and evaluation systems for ranking. llms.txt targets ChatGPT, Perplexity, Claude, and other LLM-powered tools that cite sources in generated answers.
Final Thoughts
llms.txt is a lightweight, fast-to-deploy tool that amplifies your existing content for AI agents. Deploy it if you publish original data, compete for AI citations, or want your brand represented in AI-generated answers. Skip it if your customers do not use AI tools for research. The file takes 20 minutes to create and costs nothing to deploy. The upside is visibility in the fastest-growing search channel. The downside is a few minutes of setup time.