llms.txt for Small Business: What It Does and Doesn't Do

TL;DR

llms.txt is an optional Markdown file at /llms.txt. It was proposed in September 2024 and updated to v2 on August 10, 2026, and no standards body has adopted it.
Google says Search and its AI features don't use llms.txt, and that keeping one won't harm or help your Google visibility.
OpenAI, Anthropic and Perplexity each document a search crawler that robots.txt controls. None of those crawler docs says llms.txt decides who gets recommended.
Do four checks first: crawler access, your Google Business Profile, matching name, address and phone everywhere, and markup that matches the page.
If you add the file, copy business details word for word from your Google Business Profile so the two can never disagree.

Does my small business need an llms.txt file?

Probably not yet, and it should never be the first thing you do. llms.txt is an optional proposal, not a standard. Google says Search and its AI features don't use it. I also found no AI platform that says it uses the file to pick which businesses to recommend. It costs little, so add it last, once crawler access, listings and site basics are solid.

That's my short answer. The rest of this guide covers what the file is, what each platform's own docs say, the four checks I'd run first, and a template you can copy that will never contradict your Google listing.

What is llms.txt, in plain English?

It's a short Markdown file at the root of your website that sums up who you are and points to your most useful pages. The proposal suggests adding a /llms.txt Markdown file to websites, offering brief background information, guidance, and links to more detailed Markdown files. Think of it as a cover sheet for your site that both machines and people can read.

The idea came from the developer world. The original background pointed to coding tools that pull in information about programming libraries and APIs from website documentation. That origin matters. The file was built for documentation sites, not for a plumber, dentist or bakery.

What changed in v2 (August 2026)

Most small business guides still describe the 2024 version. Jeremy Howard published the proposal on September 3, 2024, and the page now shows a modified date of August 10, 2026. The v2 page describes itself as a proposal to standardize on an /llms.txt file that helps agents use a website . In other words, the wording now leans toward AI agents that browse and act, not toward search rankings.

The v2 background says thousands of sites publish an llms.txt file, documentation platforms generate one automatically, Chrome's Lighthouse checks for one in its agentic browsing audits, and OpenAI, Anthropic and Gemini publish llms.txt files for their own developer docs. Look closely at that last part. Those companies publish the file for their own docs. That doesn't mean they read yours to decide who gets recommended.

One more detail most guides skip: the proposal makes no particular recommendation for how to process the file, since that depends on the application. Nobody is required to read it.

What goes in the file

The format is simple. It starts with an H1 heading with your name, which is the only required part. After that come an optional blockquote summary, optional plain paragraphs, and any number of H2 sections with lists of links, each followed by a short note if you want one. An "Optional" section at the bottom holds lower-priority links that a reader can skip.

How is llms.txt different from robots.txt and sitemap.xml?

robots.txt controls access. sitemap.xml lists your pages. llms.txt only summarizes. It can't open or close anything on your site.

File	Its job	Can it block a crawler?
robots.txt	Tells crawlers what they may fetch	Yes
sitemap.xml	Lists pages you want found	No
llms.txt	Short guide to your best pages	No

If you only remember one thing from this article, make it this: your AI crawler policy lives in robots.txt and in your CDN or firewall settings, not in llms.txt. I cover robots.txt in more depth in how to allow AI crawlers in robots.txt.

What do Google, OpenAI, Anthropic and Perplexity actually say?

Google says outright that it doesn't use the file. OpenAI, Anthropic and Perplexity document crawlers that robots.txt controls, and none of their crawler docs says llms.txt is used to choose businesses. Here is each one, from their own docs where I could find them.

Google (Search, AI Overviews, AI Mode)

Google's AI optimization guide has a myth-busting section on this. It says you "don't need to create new machine readable files, AI text files, markup, or Markdown" to appear in Search, including its AI features, because Google Search itself doesn't use them. It also says it's fine to create and maintain llms.txt files for other services or systems that use them.

Google went further in June 2026. An update published June 15 to its Search Central documentation added that maintaining llms.txt files won't harm, nor help, visibility or rankings in Google Search.

Google's separate AI features and your website doc covers the basics that do matter. To show as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible to appear in Search with a snippet. Traffic from these features shows up in the Search Console Performance report under the "Web" search type. The same doc tells businesses to keep their Business Profile information up to date. For a local business, that one line matters more than the whole llms.txt debate. It also says there's no special schema.org structured data you need to add.

So why does Chrome check for the file? That's a different product with a different goal. Lighthouse has an experimental Agentic Browsing audit that checks how a site handles llms.txt. Its documentation says that without the file, agents "may spend more time crawling the site" to understand its structure. When someone asked Google's John Mueller about the mismatch, he said that if an AI platform needs a file, give it to them, but that no AI sites actually require or use llms.txt, so the question is hypothetical. More on Google's AI surfaces in how to get found in Google AI Overviews and Google AI Mode for local businesses.

OpenAI (ChatGPT)

OpenAI's crawler doc talks about robots.txt, not llms.txt. OAI-SearchBot is used to surface websites in ChatGPT's search features, and sites that opt out of it won't be shown in ChatGPT search answers, though they can still appear as navigational links. OpenAI recommends allowing OAI-SearchBot in robots.txt and allowing requests from its published IP ranges.

Training and search are separate settings. Each setting is independent, so a site can allow OAI-SearchBot to appear in search results while disallowing GPTBot to opt out of model training. See how to get found in ChatGPT for the rest of the picture.

Anthropic (Claude)

Anthropic's crawler help article lists separate bots. ClaudeBot collects training data, Claude-User fetches pages when Claude users ask questions, and Claude-SearchBot indexes content for search results. Each has its own robots.txt user-agent. Anthropic says blocking Claude-SearchBot keeps its system from indexing your content, which may reduce your site's visibility and accuracy in user search results. Anthropic says all three bots honor robots.txt, including Claude-User. More in how to get found in Claude.

Perplexity

Perplexity's help center says PerplexityBot only crawls content in compliance with robots.txt. According to its crawler documentation, PerplexityBot isn't used to crawl content for AI foundation models, so blocking it protects nothing from training and removes you from Perplexity's results. Be aware that OpenAI and Perplexity warn that robots.txt rules may not apply to ChatGPT-User and generally don't apply to Perplexity-User , which are fetchers that run when a user asks. More in how to get found in Perplexity.

Gemini app, Grok and Apple

I found no primary source from Google's Gemini app team, xAI or Apple saying they read llms.txt to pick businesses, so I'm not going to guess. If that changes, the source will be the platform's own documentation, not a vendor blog. For what is known, see Gemini, Grok and Apple Business Connect setup.

Does anything actually read llms.txt?

Some tools do, mostly AI coding assistants. Log data suggests AI search bots rarely ask for the file. Treat everything in this section as vendor or anecdotal evidence, not settled fact.

Ahrefs server logs (vendor data). Ahrefs log data across about 137K domains, shared on X, found that most valid llms.txt files got zero requests in May. One former Ahrefs researcher sums it up as 97% of published files getting zero requests, with Claude Code as the main consumer rather than search bots.
Who does read it. The bots people hope are reading it for AI search, like OAI-SearchBot and PerplexityBot, barely show up. The main consumers are Claude Code and other coding agents.
A counterpoint. The team at Evil Martians posted that their logs show Claude fetching llms.txt. That's one site's anecdote, but it's a reminder that "rarely" isn't "never."
Adoption among top cited sites (vendor study). Web Gerek reports that SE Ranking's January 2026 study found none of the top 20 most AI-cited media companies had an llms.txt file, and adoption was about 10% across 300k domains, mostly smaller and developer-focused sites.

Put plainly: the sites that get cited the most don't seem to need the file, and the bots that power AI search rarely ask for it. That's why I rank it last.

What should a small business fix before adding llms.txt?

Four things come first, because AI tools and customers can actually see them. Each check below fixes a real gap. llms.txt fixes none of them.

Check	Where to look
1. Search crawlers allowed	robots.txt and CDN or firewall
2. Profile complete	Google Business Profile
3. Same name, address, phone	Google, Bing, Apple, directories
4. Markup matches page	LocalBusiness structured data

1. Let the search crawlers in

Open yourdomain.com/robots.txt. Look for rules that name OAI-SearchBot, Claude-SearchBot, Claude-User or PerplexityBot, or a blanket rule that blocks everything. A simple allow looks like this:

User-agent: OAI-SearchBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: Claude-User
Allow: /

User-agent: PerplexityBot
Allow: /

Training crawlers like GPTBot and ClaudeBot are separate settings. Whether you allow them is your call, and it doesn't change the search settings above. For Google's AI features, your normal Googlebot rules are what count.

Then check your host, CDN or firewall. A robots.txt allow means nothing if your firewall blocks the same agent, and Perplexity publishes WAF setup guidance for Cloudflare and AWS because this is a common failure. If your dashboard has a "block AI bots" switch, make sure it isn't stopping the search crawlers you want.

2. Finish your Google Business Profile

Google's own AI features doc tells businesses to keep their Business Profile information up to date. That means the right primary category, accurate hours, service area, a clear description, real photos and replies to reviews. This is the source your llms.txt should copy from, so get it right first.

Other assistants may lean on other sources. A summary of BrightLocal research posted on X says ChatGPT barely used Google Business Profiles and that Yelp was a leading source. That's a vendor study relayed secondhand, but it's a good reason not to stop at Google. I compare the two in ChatGPT vs Google for local search.

3. Make your name, address and phone match everywhere

Use the same business name, address and phone number on Google, Bing Places, Apple Business Connect, Yelp and the directories in your industry. If your llms.txt says one thing and a directory says another, the file has only added a conflict. The same BrightLocal summary also reported that results changed a lot from one repeat search to the next. Consistent facts give every engine the same story each time it looks.

4. Match your structured data to the page

Google says no special schema is needed for its AI features, and an Ahrefs study shared on X (vendor data) found schema had no meaningful effect on ChatGPT citations. I still treat LocalBusiness markup as basic hygiene. If you have it, the name, phone, hours and service area in it should match what the page shows. More in LocalBusiness schema for AI search.

For a fuller list, use my AI visibility audit checklist.

How do I write an llms.txt for a small business?

Keep it short and copy every business detail word for word from your Google Business Profile. That way the file can never contradict your listing. If a detail changes on Google, change it in the file the same day.

Here's a template for a made-up home service business. Swap in your own details. The bracketed parts tell you where each detail comes from.

# Example Plumbing Co.

> Licensed residential plumber serving
> [service area, as on Google Business Profile].
> Free estimates.

Name, phone, hours and service area below
match our Google Business Profile.

## Services
- [Drain cleaning](https://example.com/drain-cleaning/):
  Clogs, camera inspections, hydro jetting
- [Water heaters](https://example.com/water-heaters/):
  Repair and replacement, tank and tankless
- [Request an estimate](https://example.com/estimate/):
  How to book and what to expect

## About and contact
- [About us](https://example.com/about/):
  Who we are, licensing, service area
- [Contact](https://example.com/contact/):
  [Phone, exactly as on Google], [hours]

## Optional
- [Reviews](https://example.com/reviews/):
  Customer reviews
- [FAQ](https://example.com/faq/):
  Common questions and answers

Tips for writing yours:

Use absolute URLs. A crawler reading the file on its own has no base URL to work from.
List 5 to 10 pages that answer real customer questions: services, pricing (if public), about, contact, FAQ.
Skip thin pages, tag archives, old promos and anything behind a login.
Only list prices that are on the page. A price that exists only in llms.txt will go stale.
Keep it short. Twenty to thirty lines is plenty for most small businesses.
Write for humans too. If a customer opened it, would it make sense? That's a good test.

How do I publish and test it?

Put it at the root of your domain so it loads at yourdomain.com/llms.txt. That's the location the proposal names, and it needs to be public with no login.

WordPress. You can use an SEO plugin like All in One SEO, or upload the file by FTP to your public_html directory.
Static sites. For Astro, Next.js, Hugo and similar, drop the file in the public or static folder and it will be served at the root.
Hosted builders. Some platforms need a workaround, so check your platform's documentation for serving a static file at the root, or contact support. If yours can't do it, skip the file. You lose very little.

To test it:

Open yourdomain.com/llms.txt in a private browser window. Make sure it loads as plain text and reads clearly.
Click every link. Each one should load a live page, with no redirects to a 404.
Compare the name, phone, hours and service area against your Google Business Profile, line by line.
If you're curious, run Chrome's Lighthouse agentic browsing audit, which checks for the file. Remember that's an agent readiness check, not a search ranking signal.
Set a reminder to review the file whenever your hours, services or phone change.

Can llms.txt hurt my business?

Not in Google, by Google's own account. The real risk is the file going stale and contradicting your listing.

Google's June 2026 update says the file won't harm or help your Google Search visibility. The risks I'd watch for are practical ones:

Contradictions. An old phone number, old hours or an old service area in the file conflicts with your listing.
Exposure. Listing pages you'd rather not have quoted, like old price sheets or staging URLs.
Wasted money. Paying a vendor who sells the file as a citation fix.

What mistakes should I avoid?

Treating it as your bot policy. It can't block anything. Set crawler rules in robots.txt and check your CDN.
Doing it before the basics. A perfect file on top of a half-finished Google listing fixes nothing.
Stuffing it. Every page on your site is not a reading list. Pick the pages that answer real questions.
Writing marketing copy. Plain facts beat slogans. "Licensed plumber, free estimates" is better than "the best in the business."
Believing guarantees. No platform has said the file earns citations. Anyone promising that is promising something they don't control.

Should you add one? A quick decision list

Add one if all three of these are true:

You can edit files at your site root without paying a developer.
Crawler access, your Google Business Profile, and your name, address and phone are already clean.
You're fine with the fact that nothing is promised.

Wait if any of these are true:

Your Google listing is incomplete, or a wrong phone number is floating around somewhere.
You haven't checked whether your CDN or firewall blocks AI search crawlers.
Someone is selling you the file as a way to get cited.

Add it now, no matter what else is true, if you publish developer docs or an API. That's the use case the file was built for.

What actually helps a small business show up in AI answers?

Pages on your own site that clearly answer what buyers ask, consistent listings, and being mentioned on third-party sites AI tools already cite. Here is what the studies point to. Most are vendor studies, so read them as signals, not rules.

Your own website still matters for local searches. A BrightLocal study of 200,875 local AI searches, relayed on X, found individual business sites made up most of the unique cited domains and drew about 42% of citations (vendor study).
Roundups and lists get cited a lot. Ahrefs reported that "best X" listicles were 43.8% of the page types ChatGPT cites (vendor study, shared on X). For a small business, being fairly included on those lists can matter more than any file on your own site.
Clear, fact-rich writing helps. The academic GEO paper found that adding citations, quotations and statistics improved how visible a source was in generative engine answers in its tests.
Each engine is different. Results vary by platform and from one run to the next, so check each one on its own. Here's how to check if AI recommends your business.

If a competitor keeps showing up and you don't, read why ChatGPT recommends my competitor. For the bigger picture, start with what AI search visibility is and GEO vs SEO for local businesses.

How I approach this at Get Found with AI

I help businesses across the United States get found, cited and recommended in ChatGPT, Claude, Gemini, Grok, Perplexity and Google's AI features. When a site has an llms.txt, I check that it matches the Google Business Profile. When it doesn't have one, I add it near the end of the list, if at all. The work that comes first is crawler access, complete listings, matching name, address and phone, markup that matches the page, and pages that answer real customer questions.

I don't promise placement, rankings, citations or timelines, because nobody controls those. You can see how the process works on how it works and what's included on services. If you're comparing providers, how to choose an AI search optimization agency and what AI search optimization costs will help. More guides are on the blog.

How I put this together

On October 22, 2026, I searched "llms.txt for small business" (US) and read the three top results:

How to write an llms.txt file for a small business site (DEV Community). This is a tight how-to on the format, with a good tip about absolute URLs. It doesn't cover whether the file works, what Google says, or crawler access.
llms.txt Guide for Small Business (OnePoint Solutions, updated October 4, 2026). It has useful platform-by-platform setup steps. It says the file tells AI assistants what your business does, but it doesn't put that claim next to Google's statement that Search doesn't use the file. It also suggests adding prices, which can drift out of sync with your pages.
What Is an llms.txt File, and Does Your Small Business Website Need One? (David Martin Design). It's honest that the file isn't a ranking hack. It was published before Google's May and June 2026 guidance and before the v2 update, so it misses both.

I also read Web Gerek's honest guide, which has the SE Ranking data I cite above. None of these pages covered the v2 update from August 10, 2026, Google's "won't harm (nor help)" wording, or the split between search crawlers and training crawlers. None gave a priority order for a small business or a template tied to the Google Business Profile. I checked platform claims against Google, OpenAI, Anthropic and Perplexity's own documentation wherever I could. Where I relied on a secondary report or a vendor study, I say so.

Frequently asked questions

What is an llms.txt file?

It is a plain Markdown file at yourdomain.com/llms.txt. It gives a short summary of your business and links to your most useful pages. Jeremy Howard proposed it in September 2024 and updated it to v2 on August 10, 2026. It is a proposal, not an adopted standard.

Does Google use llms.txt for AI Overviews or AI Mode?

No. Google's AI optimization guide says you don't need AI text files or Markdown to appear in Google Search, including its AI features, because Google Search doesn't use them. Google also says keeping one won't harm or help your visibility.

Will llms.txt get my business recommended by ChatGPT, Claude or Perplexity?

I found no statement from OpenAI, Anthropic or Perplexity saying they use llms.txt to choose which businesses to recommend. Their crawler docs point to robots.txt for access. No file can promise citations, and I can't promise any placement.

Is llms.txt the same as robots.txt?

No. robots.txt tells crawlers what they may fetch, and the major AI search crawlers follow it. llms.txt is a summary and reading list. It cannot allow or block anything.

Can llms.txt hurt my website?

Google says it won't harm or help your Google visibility. The real risk is a file that goes stale and contradicts your Google Business Profile, like an old phone number or wrong hours. Copy details word for word from your profile, or skip the file.

What should I fix before adding llms.txt?

Four things: let search crawlers like OAI-SearchBot, Claude-SearchBot and PerplexityBot in through robots.txt and your CDN, complete your Google Business Profile, make your name, address and phone match everywhere, and make sure any LocalBusiness markup matches the visible page.

Should I pay someone to make an llms.txt file?

Not as a way to get cited. You can write one yourself in under an hour with the template in this article. Be wary of anyone who sells the file as a citation or ranking fix, because no platform has said it works that way.

llms.txt for Small Business: Do You Need One in 2026?