An llms.txt file is a plain Markdown file placed at the root of a website, at yoursite.com/llms.txt, that gives AI tools a short, curated summary of the site and links to its most useful pages. For a local or small business it is a cheap, low-risk addition. It is not a ranking factor, and no major AI search product has confirmed relying on it, so it should come after the basics.
What is llms.txt, and who proposed it?
The llms.txt idea was proposed in 2024 by Jeremy Howard, co-founder of Answer.AI and fast.ai. The reasoning is simple: language models have limited context windows, and websites are full of navigation, scripts, cookie banners and ads. A clean Markdown file that lists what a site is about and where the key information lives could help AI systems find the right content.
The proposed format has a few parts:
- An H1 heading with the name of the site or business.
- A short blockquote summary of what the business does.
- Optional paragraphs with extra context.
- H2 sections containing lists of links, each with a short description.
- An optional section for less important links that a tool can skip.
Some sites also publish an llms-full.txt file with the full text of key pages. That is common for software documentation, less so for local businesses.
Do AI search engines actually use llms.txt?
This is the honest part. At the time of writing, adoption is strongest among developer tools and documentation sites, and some AI coding assistants and documentation platforms can read these files. Google has said that it does not use llms.txt for Search, and Google representatives have compared it to the old keywords meta tag. Other large AI search providers have not publicly confirmed that their systems rely on it.
So llms.txt is best treated as a small, optional addition. It costs very little, it does no harm, and it may become more useful if adoption grows. It will not, on its own, get a dental clinic or a bakery mentioned in ChatGPT or Google’s AI Overviews.
What matters more for local AI visibility?
AI assistants answering local questions, such as “best physiotherapist near Kadıköy” or “plumber open on Sunday in Leeds”, usually draw on sources they can trust: business profiles, review platforms, maps data and well-structured websites. For a small business, these foundations carry far more weight:
- A complete and accurate Google Business Profile, with correct category, hours, services, photos and regular review responses.
- Consistent name, address and phone number across directories, social profiles and the website.
- Apple Business Connect and Bing Places listings, since some assistants use those maps sources.
- Reviews on the platforms customers actually use in your sector.
- A website with clear service pages, prices or price ranges where possible, an FAQ, and LocalBusiness structured data.
- A robots.txt that does not accidentally block crawlers you want to reach your site.
If these are in place, llms.txt is a reasonable finishing touch.
How do you write a simple llms.txt file?
For a small business, the file can be under a page long. Here is the structure, described in plain terms, for a fictional family-run bike repair shop in Izmir.
Heading and summary
Start with an H1 line containing the business name. Below it, a one- or two-sentence blockquote: what the business does, where, and for whom. For example, a bike repair and rental shop in Alsancak serving commuters and tourists, open six days a week.
Key facts paragraph
Add a short paragraph with facts an assistant might need: service area, opening hours, languages spoken, booking method and whether walk-ins are accepted. Keep it consistent with your Google Business Profile.
Link sections
Under an H2 such as “Services”, list links to each service page with a one-line description: repairs, rentals, bike fitting. Under “Visit”, link the contact page and the booking page. Under “Optional”, link the blog or news pages.
Upload and check
Save the file as llms.txt, upload it to the root of your site so it opens at yourdomain.com/llms.txt, and check it loads as plain text in a browser.
How do you add it on common website platforms?
- WordPress: some SEO plugins, including Yoast and Rank Math, have added llms.txt generation features; otherwise upload the file to the site root with your host’s file manager.
- Shopify, Wix and Squarespace: root file access is limited, so check the platform’s current support, apps or help documentation.
- Custom sites: your developer can place the file in the public root folder.
Update it whenever you add a service, change hours or move pages. An outdated llms.txt is worse than none.
What mistakes should small businesses avoid?
- Treating llms.txt as a replacement for a proper website or business profile.
- Stuffing it with keywords or location lists; it is meant to be a clear summary.
- Listing information that contradicts your other listings.
- Paying a large fee for a file that takes an hour to write.
- Expecting traffic changes in analytics because the file exists.
How Alien Road applies llms.txt in GEO work
In a typical engagement for a local business, discovery starts by asking several AI assistants the questions customers ask, such as “who repairs bikes near Alsancak”, and noting which businesses are named and which sources are cited. That shows where information is missing.
Setup puts the foundations first: business profiles, directory consistency, service pages, FAQs and LocalBusiness schema. Only then do we write an llms.txt file, drawn from the same verified facts, and add it to the site root.
For measurement we repeat the same set of assistant questions monthly, record mentions and citations, and track Google Business Profile actions such as calls, direction requests and website visits.
Reporting shows which changes coincided with new mentions. We are clear with clients that llms.txt is a small, optional element and that most gains come from the profile and website work.
Where to start this week
Open your Google Business Profile and fix anything incomplete: hours, services, categories, photos. Then write a half-page llms.txt with your name, a two-sentence summary, key facts and links to your main pages, and upload it. Finally, ask ChatGPT, Gemini and Perplexity the question your customers would ask, and save the answers. You now have a baseline to compare against next month.
Keep reading: GEO – AIO – AEO · GEO