llms.txt: The One File That Lets AI Cite Your Business

Photo: Blue Screen of Death in Silver Black Laptop - Credit to http://homedust.com/ by Homedust, CC BY 2.0. Cropped.
Key takeaways
- llms.txt is a plain Markdown file at yoursite.com/llms.txt that summarises your business and links to your most important pages for AI systems.
- The format is simple: one H1 with your name, a short summary in a blockquote, then sections of Markdown links with one-line descriptions.
- It is a low-cost, low-risk addition. It does not replace SEO, but it helps AI tools describe you accurately, and Lighthouse now checks it.
What is llms.txt?
llms.txt is a plain-text file placed at the root of your site (yourdomain.com/llms.txt). It gives language models a concise, structured summary of your business: who you are, what you offer, and which pages matter. Think of it as the curated briefing your homepage wishes it could hand directly to an AI that is about to answer a customer's question about you.
Why it matters now
Answer engines do not scroll your site the way a human does. They sample, summarise and decide what to cite. Without a clear machine-readable summary, they fall back on whatever third-party directory or cached snapshot they happen to find — which is often out of date. A correct llms.txt lets you control the story AI tells about your business.
What a good llms.txt contains
- Your company name, founding year, location and founder.
- Your primary offering in one line (no jargon).
- Disambiguation — what you are NOT, so stale descriptions get corrected.
- Key facts worth citing: years of experience, apps shipped, published books, client outcomes.
- Verified profile links (LinkedIn, etc.) and your main page URLs.
Does your site pass? Check free
Run your URL through the Kavenacc AEO / SEO Auditor. It checks for an llms.txt, an AI-crawler-friendly robots.txt, schema and GEO citation signals, then scores you 0–100 with a fix list. Kavenacc's own llms.txt is public at kavenacc.co.ke/llms.txt — use it as a template.
Common mistakes
Three errors sink most llms.txt files: (1) leaving in old positioning from a previous business model, (2) writing for humans instead of models (no clear structure), and (3) forgetting to link the file from anywhere. The auditor's "Machine-readable context" dimension flags all three.
Make it part of a system
llms.txt is one tile in a larger AEO floor: robots.txt for AI crawlers, schema for structure, and citation-rich content. Kavenacc ships all of them as standard in the OpsCommand build — so your site is citeable from day one, not after a six-month retrofit.
The format, exactly
The llms.txt proposal (llmstxt.org) defines a small Markdown structure. A valid file looks like this:
# Your Company Name > One or two sentences: what you do, for whom, where, and your price range. Contact: email · phone · city. ## Main pages - [Services](https://example.co.ke/services): What you offer and starting prices. - [Pricing](https://example.co.ke/pricing): Packages and payment options. - [Contact](https://example.co.ke/contact): How to reach you. ## Optional - [Blog](https://example.co.ke/blog/): Guides and articles.
The rules that matter:
- Exactly one H1 (a single line starting with
#) holding your business name. - A blockquote summary (a line starting with
>) directly after it. - Sections with H2 headings (
##) containing Markdown links in the form[Title](URL): description. - An "Optional" section for secondary links that can be skipped when an AI has little room.
Common mistakes that make the file fail checks
- Bare URLs instead of Markdown links. Tools that validate the file, including Google Lighthouse's agentic browsing audit, look for
[text](url)links. A list of plain URLs can be reported as "File does not appear to contain any links." - Several H1 lines. Comment-style lines starting with
#at the top of the file count as extra headings. - Serving it as HTML. The file must be plain text at
/llms.txt, not a web page. - Out-of-date facts. If prices or services change on your site but not in llms.txt, AI tools may repeat the old version.
- Blocking AI crawlers in robots.txt. An llms.txt file is pointless if robots.txt tells GPTBot, ClaudeBot or PerplexityBot to stay away.
What to put in it for a Kenyan SME
- Your legal or trading name, town and the counties or countries you serve.
- Your main services with starting prices in KES or USD.
- Payment methods (M-Pesa, card, bank), delivery or service areas, and opening hours.
- Links to your service pages, pricing, contact page, FAQ and best guides.
- Verified social profiles, so AI tools don't confuse you with a similarly named business.
- A disambiguation note if your business has changed what it does, so old directory listings are not repeated as current.
How llms.txt fits with robots.txt, sitemaps and schema
| File | Purpose | Audience |
|---|---|---|
| robots.txt | Which crawlers may visit which paths | All crawlers |
| sitemap.xml | Complete list of indexable URLs | Search engines |
| Structured data (schema) | Machine-readable facts on each page | Search engines and AI systems |
| llms.txt | A curated, human-readable summary and reading list | Large language models and AI agents |
They work together rather than replacing each other. Kavenacc's own llms.txt is a working example you can compare against.
Does it actually help?
Be realistic: llms.txt is a young proposal, and AI companies have not all said how, or whether, they use it. What it reliably does is give any AI agent that reads it an accurate, compact summary of your business, and it takes less than an hour to write. It also counts towards Lighthouse's agentic browsing checks. The bigger wins for AI visibility still come from clear pages, consistent facts and being mentioned by other trusted sites. Treat llms.txt as one inexpensive part of that system.
Writing the summary line
The blockquote under your title is the most important part of the file, because it is what an AI is most likely to repeat. Write it as one or two factual sentences that answer: who are you, what do you do, for whom, where, and roughly what does it cost? For example: "Kavenacc Digital Solutions is a Nairobi studio that builds websites, web apps and mobile apps with built-in AI features, launched in 7 days from $499, for businesses in Kenya and East Africa." Avoid slogans and superlatives; AI systems are increasingly cautious about repeating marketing claims, and plain facts are easier to quote accurately.
A step-by-step way to create yours
- List your ten most important pages: homepage, each core service, pricing, contact, about or founder, FAQ, and your best two or three guides.
- Write a one-line description for each that says what a reader will find there, ideally including a price, place or number.
- Group them under H2 headings such as "Main pages", "Services", "Guides" and "Optional".
- Add key facts as short bullet points: founding year, location, service area, payment methods, verified social profiles.
- Save as UTF-8 plain text named
llms.txtand upload it to the root of your site. - Test it: open yoursite.com/llms.txt in a browser, then run a Lighthouse report, which includes an llms.txt check in its agentic browsing audits.
- Set a reminder to update it whenever prices or services change.
WordPress, Wix and other platforms
On WordPress you can upload the file to the site root with your host's file manager, or use an SEO plugin that supports llms.txt. Hosted builders that don't allow root files sometimes offer a setting or app for it; if yours doesn't, it's worth asking support, as the request is becoming common. If your site is a JavaScript app, make sure the server returns the text file directly rather than routing /llms.txt to the app's homepage, a surprisingly common mistake.
Optional: llms-full.txt
Some sites also publish llms-full.txt, which contains the full text of key pages in Markdown, so an AI can read everything in one request without crawling. It's useful for documentation-heavy sites. For most small businesses, a well-written llms.txt with good links is enough.
More questions answered
Is llms.txt an official standard?
Not yet. It is a community proposal, published at llmstxt.org, that has been adopted by many sites and tools, and Google's Lighthouse now checks for it. Because it is just a text file, it carries no risk, so adopting it early costs little.
Should I list every page?
No. llms.txt is a curated summary, not a sitemap. Include your most important pages and your best guides, and leave the full list to sitemap.xml.
How long does it take for AI tools to notice the file?
There is no fixed timeline. AI agents that read llms.txt fetch it when they visit your site, so a correct file takes effect as soon as it is live. Model training and search indexes update on their own schedules, which is why consistent facts across your whole site matter more than any single file.
Related articles



Frequently asked questions
Is llms.txt the same as robots.txt?
Do I need llms.txt if I already have good SEO?
Where does llms.txt go?
Check your own site free
Before you optimise, see where you stand. Kavenacc's free AEO / SEO Auditor scores any website across 13 dimensions — llms.txt, robots.txt, schema, citations and more — and hands you a prioritized fix list.
Run the free AEO / SEO Auditor →