What Is llms.txt? Does Your Singapore Business Website Need One?

Laptop on a Singapore office desk showing a plain text file of links, like an llms.txt file

Quick answer

llms.txt is a plain text file at the root of a website that lists its most useful pages for AI tools. It is a 2024 proposal, not an official standard. Google ignores it and no major AI company says it relies on one, so treat it as optional.

This guide is for Singapore business owners and marketing managers who have heard that “every website needs an llms.txt now” and want a straight answer before paying anyone to set one up. You will learn what the file is, who actually reads it today, how it differs from robots.txt, and the short list of things that do make AI assistants like ChatGPT, Gemini and Google’s AI Overviews more likely to mention your business.

We look at this from the working side. Adssential Marketing is a Singapore digital growth agency that builds websites and runs SEO and GEO (generative engine optimisation) work for local businesses. We test how AI assistants describe our clients, so we care more about what moves those answers than about the latest file format.

What is llms.txt?

llms.txt is a Markdown file placed at yourdomain.com/llms.txt that gives large language models a short, curated map of your website. Instead of an AI tool having to crawl hundreds of pages, the file points it to the handful that explain your business best.

The idea came from Jeremy Howard, who published the proposal on llmstxt.org on 3 September 2024. The reasoning is simple: AI tools work within limited context, and a normal web page is full of menus, scripts and adverts that get in the way. A clean list of links with one-line descriptions is easier for a model to use.

The proposal sets out a loose format:

  • An H1 heading with the name of the site. This is the only required part.
  • A short summary in a blockquote under the heading.
  • Sections under H2 headings, each holding a list of links in the form [page name](URL): one-line description.
  • An “Optional” section for secondary pages that a tool can skip when it is short on space.

The same proposal also suggests offering clean Markdown copies of important pages (for example page.html.md), so a tool can read the content without the page design around it.

Here is what a simple llms.txt for a Singapore service business could look like:

# Tan & Co Aircon Services

> Aircon servicing, chemical washes and repairs for homes and offices in Singapore. Same-week appointments, island-wide.

## Services
- [Aircon servicing](https://example.sg/aircon-servicing/): routine servicing plans and what each visit includes
- [Chemical wash](https://example.sg/chemical-wash/): when a chemical wash is needed and how long it takes
- [Aircon repair](https://example.sg/aircon-repair/): common faults, diagnosis and repair process

## About
- [About us](https://example.sg/about/): company background, team and service areas
- [Contact](https://example.sg/contact/): phone, WhatsApp and booking form

## Optional
- [Blog](https://example.sg/blog/): maintenance tips and buying guides

Example only. The business and URLs are invented to show the format.

Does Google use llms.txt?

No. Google has said plainly that Google Search ignores llms.txt, and that creating one will neither help nor harm your rankings or your visibility in AI Overviews.

Google’s documentation on AI features and your website says you do not need to create new machine-readable files, AI text files or special markup to appear in AI Overviews or AI Mode. Its AI optimisation guide for site owners goes further and names llms.txt directly: you can create one for other services if you like, but Google Search ignores it.

Google’s own staff said the same thing earlier. John Mueller compared llms.txt to the old keywords meta tag, a field where a site owner simply claims what the site is about (Search Engine Journal, 17 April 2025). Gary Illyes was reported as saying Google does not support llms.txt and is not planning to (Search Engine Roundtable, 24 July 2025).

📊 Field data and verified research

Google’s guidance: “You don’t need to create new machine readable files, AI text files, or markup to appear in these features.” The same page says there are no extra requirements for AI Overviews or AI Mode beyond normal Search eligibility. Source: Google Search Central, AI features and your website (linked above), opened 15 Nov 2026.

For a Singapore business that gets most of its search traffic from Google, this matters most. Google’s AI Overviews and AI Mode draw on pages that are already indexed and eligible to appear in normal search. If your site is not doing well in Google today, an llms.txt file will not change that.

Do ChatGPT, Claude and Perplexity read llms.txt?

None of them has said in its official crawler documentation that it uses llms.txt to decide what to cite. What they do document is robots.txt, which is how you allow or block their crawlers.

We checked the crawler pages of the three companies most Singapore businesses ask us about. OpenAI’s crawler documentation, Anthropic’s help page on its crawlers and Perplexity’s bot guide explain their bots and how robots.txt controls them. None of them says its crawler reads llms.txt or gives it any weight.

There is a twist that causes much of the confusion. All three companies publish an llms.txt for their own developer documentation, as an index for programmers who feed docs into coding tools. That is a sensible use of the file, but it is not the same as saying their search products rank or cite websites because of it.

Some site owners have seen AI bots fetch their llms.txt in their server logs; one case reported by Search Engine Roundtable showed OpenAI requesting the file regularly. A fetch is not the same as a ranking or citation benefit, and the largest study so far suggests most files are hardly ever requested at all.

📊 Field data and verified research

Ahrefs studied 137,210 websites in its web analytics data and found that about 28% published an llms.txt file. Of those files, 97% received no requests at all in May 2026, and named AI bots made up only about a fifth of the requests that did happen. Source: Ahrefs, llms.txt study, 15 June 2026.

llms.txt vs robots.txt: what is the difference?

robots.txt tells crawlers what they may and may not visit, and the major AI companies say they follow it. llms.txt only suggests what is worth reading, and nobody is obliged to follow it.

If you want to control whether your content is used to train AI models, or whether it can appear in AI search answers, robots.txt is the tool that actually works today. Each company uses separate crawler names, so you can, for example, allow a search bot while blocking a training bot.

Crawler (user agent) Company What it is for Blocking it in robots.txt means
OAI-SearchBot OpenAI Finding pages for ChatGPT search Your pages will not be shown in ChatGPT search answers (OpenAI says changes take about 24 hours)
GPTBot OpenAI Collecting content to train its AI models Your content should not be used for model training; search is controlled separately
Google-Extended Google Not a crawler but a control token for Gemini training and grounding Google says it does not affect your inclusion or ranking in Google Search
ClaudeBot Anthropic Collecting content that may be used for training Your content is excluded from that collection
Claude-SearchBot Anthropic Improving search results in Claude Anthropic says blocking it may reduce your visibility in Claude’s answers
PerplexityBot Perplexity Indexing pages for Perplexity’s search answers Your pages are less likely to appear in Perplexity answers

Sources: the OpenAI, Anthropic and Perplexity crawler pages linked above, and Google’s list of common crawlers, all opened 15 Nov 2026. User-initiated fetchers such as ChatGPT-User and Perplexity-User work differently and may not follow robots.txt in the same way.

One problem is worth checking before anything else, because it is the opposite of what people worry about. A security plugin, a firewall rule or an over-cautious robots.txt quietly blocks the search bots of AI assistants, and the business never shows up in that assistant’s search answers. Check robots.txt first; it has a far bigger effect than any llms.txt.

Should your Singapore business create an llms.txt?

For most small and medium businesses, llms.txt is a low-priority extra: harmless, cheap if your SEO plugin builds it, and unlikely to change how AI assistants talk about you. Spend the effort elsewhere first.

Here is how we would weigh it for different situations:

Your situation Our view on llms.txt
Local service business (clinic, contractor, F&B, tuition) Optional. Fix your Google Business Profile, service pages and reviews first
B2B services with long buying cycles Optional. Clear service pages, case studies and consistent company facts matter far more
Software company or anything with developer documentation Worth doing. Developers paste docs into AI coding tools, and a clean llms.txt with Markdown pages helps them
Large content site (hundreds of articles, a knowledge base) Reasonable to add if your CMS generates it automatically and keeps it current
You already use Rank Math or another plugin that builds it Fine to switch on, as long as it only lists pages you want public and indexed

The file has no known downside if it is accurate. The real cost is a stale file: a list of old URLs, discontinued services or pages that now redirect. If you add one, someone has to keep it current, the same as your sitemap.

💡 From our experience at Adssential

The businesses we see named by AI assistants did not get there with a special file. They got there because their websites say clearly, in their own words, what they do, where they work and what they know from experience. AI tools keep looking for original, first-hand knowledge they cannot invent themselves, and a business that writes down what it actually knows from the ground is the one they pick up.

💡 From our experience at Adssential

For every client on one of our GEO packages, we add an llms.txt file. It is optional, as this article explains, but we would rather not leave any component out. We want to cover every aspect and support AI crawlers in every way we can, and a clear, easy-to-read summary of a website and its key information is something any business should want to give them.

How do you create an llms.txt file on WordPress?

On WordPress the simplest route is your SEO plugin; otherwise it is a plain text file uploaded to the root of your site. Either way, it takes minutes, not days.

Using Rank Math

Rank Math includes an llms.txt module. According to Rank Math’s own guide, you switch on the “LLMS Txt” module in the Rank Math dashboard, then edit it under General Settings. The file is empty until you configure it. You choose which post types and taxonomies to include (with a default limit of 100 items per type), and posts set to noindex are left out automatically. There is also a field for extra lines such as a short company summary or key links. Rank Math itself notes that llms.txt is a proposal, not an official standard.

By hand

  1. Write the file in a plain text editor using the format above: name, one-line summary, then sections of links.
  2. Keep it short. Your main service pages, your about and contact pages, and a few genuinely useful guides are enough.
  3. Save it as llms.txt and upload it to the root folder of your site, so it opens at yourdomain.com/llms.txt.
  4. Open the URL in a browser to confirm it loads as plain text, not as a 404 page.
  5. Add a reminder to review it whenever you add, remove or rename a service page.

What not to put in it

Anything in llms.txt is public, exactly like a public web page. Do not list staging sites, login areas, customer portals, internal documents or any page that shows personal data. If a page is not fit to appear in Google, it does not belong in llms.txt either. PDPC’s Guide to Data Protection Practices for ICT Systems (revised 20 July 2026) is a useful checklist for keeping systems that hold personal data out of public reach, and it reminds organisations that they remain accountable for personal data handled by the vendors they engage.

What actually helps AI assistants recommend your business?

What works is the same thing that earns trust from Google: clear, crawlable pages that answer real questions, consistent facts about your business across the web, and proof that you have done the work. Google’s own guidance says there are no special tricks for AI Overviews beyond solid SEO.

These are the changes we would make before anyone touches an llms.txt:

  1. Let the right bots in. Check robots.txt and your firewall so Googlebot, OAI-SearchBot, Claude-SearchBot and PerplexityBot are not blocked by accident.
  2. Give every service its own clear page. AI tools answer questions like “who does aircon chemical washes in Tampines”. A page that states the service, the area, the price range and the process can be quoted; a vague home page cannot.
  3. Answer the questions buyers actually ask, in plain sentences near the top of each page. Assistants lift short, self-contained answers.
  4. Keep company facts identical everywhere: name, address, phone, opening hours and services on your website, Google Business Profile, social profiles and directories.
  5. Show first-hand proof: real examples, results you can stand behind, photos of your own work and reviews from real customers.
  6. Earn mentions on other trusted sites, such as industry associations, local media and partners. AI answers often lean on what others say about you, not only on what you say about yourself.
  7. Use structured data that matches the visible page, such as Organization and LocalBusiness details, so search engines read your facts without guessing.

💡 From our experience at Adssential

Two of our SEO clients show how this carries over. SWC Construction, a Singapore waterproofing specialist, and GWS Living Art, a green wall and green roof specialist, each had 80 industry keywords tracked, and within about six months almost 80% were ranking on the first page of Google. When we asked ChatGPT for the best waterproofing contractor in Singapore, it put SWC Construction first on its shortlist, and when we asked for the best green roof contractor, it put GWS Living Art first. AI answers vary between runs, so we treat this as what we observed, not a fixed position, but neither result needed an llms.txt file; both came from strong SEO foundations.

If you want to see where your own site stands, our GEO audit checklist walks through 20 checks you can run yourself, and our guide on how ChatGPT, Gemini and Google AI Overviews pull from different sources explains why one assistant may mention you while another does not.

How do you check whether AI assistants already mention you?

Ask them the questions your customers ask, write down the answers, and repeat the test every month. It is the only honest way to measure AI visibility for a small business today.

  1. List 10 to 15 real questions a buyer would type, such as “best interior designer for a 4-room HDB flat” or “how much does a corporate website cost in Singapore”.
  2. Ask each one in ChatGPT, Gemini, Copilot and Perplexity, and look at Google’s AI Overview for the same search.
  3. Note whether your business appears, what is said about it, and which sources are cited.
  4. Check the cited sources. If a competitor keeps appearing because of a directory listing or a news article, that tells you where to earn your own mention.
  5. Repeat on the same day each month and compare. Answers change between runs, so look for a trend rather than one result.

If you prefer a second pair of eyes, our free website and GEO audit checks how search engines and AI assistants currently see your site.

Frequently asked questions

Is llms.txt an official standard?

No. It is a proposal published by Jeremy Howard in September 2024 on llmstxt.org. No standards body has adopted it, and Google has said Google Search ignores the file.

Will an llms.txt file improve my Google rankings or AI Overviews?

No. Google says creating an llms.txt file will neither help nor harm your visibility or rankings in Google Search, which includes AI Overviews and AI Mode. Those features use pages that are already indexed and eligible for normal search.

What is the difference between llms.txt and robots.txt?

robots.txt tells crawlers which parts of a site they may visit, and OpenAI, Google, Anthropic and Perplexity document that they follow it. llms.txt only suggests which pages are most useful, and no major AI company has said it relies on it.

Can llms.txt stop AI companies from using my content?

No. To limit AI training or AI search use, add rules for the relevant crawlers in robots.txt, such as GPTBot for OpenAI training, OAI-SearchBot for ChatGPT search and Google-Extended for Gemini training.

Does Rank Math create llms.txt automatically?

Rank Math has an llms.txt module, but the file stays empty until you switch the module on and choose what to include. It lists the post types and taxonomies you select and leaves out posts set to noindex.

Is it risky to add an llms.txt file to my website?

There is no known penalty for having one. The main risks are listing pages that should not be public and leaving the file out of date, so keep it short and review it whenever your services or pages change.

The bottom line

If an llms.txt file costs you two minutes in your SEO plugin, switch it on and keep it tidy. If someone is charging you to build one as an “AI visibility” fix, put that money into clearer service pages, consistent business facts and real proof of your work instead.

That is the work our GEO and AI search optimisation team does every week, alongside the SEO foundations that AI assistants draw on. If your website needs restructuring to make your services clear, our web development team can help with that too.

You Might Be Interested

Common Web Design Mistakes to Avoid (and How to Fix Them)

Common mistakes to avoid in Singapore web design when building a website. Websites are cornerstone of a successful online presence. However, many individuals and businesses often make critical mistakes in their web design endeavors.

Categories