Last updated: October 2026
llms.txt is a proposed plain markdown file at the root of your site that gives AI agents a short summary and a list of links to your most useful pages. It is cheap to make and does no harm, but there is no public evidence that ChatGPT, Perplexity or Google use it to decide which brands to cite.
Key takeaways🔗
- llms.txt was proposed by Jeremy Howard in September 2024 at llmstxt.org. It is a community proposal, not a web standard.
-
The format is simple: one H1 with your name, an optional one-line summary, and H2 sections that list links with short notes.
-
Google says Google Search ignores llms.txt, and a study of nearly 300,000 domains found no link between having the file and getting cited by AI.
-
Make one if it takes you 20 minutes and helps agents or developers find your docs. Do not expect it to move your AI visibility on its own.
-
You can draft one in seconds with the free llms.txt generator.
What is llms.txt?🔗
llms.txt is a markdown file you place at yoursite.com/llms.txt so that large language models and AI agents can find the important parts of your site without crawling every page. The idea comes from the llms.txt proposal, written by Jeremy Howard, first published on September 3, 2024 and last modified on August 10, 2026.
The reasoning is practical. Web pages are full of navigation, scripts, cookie banners and layout code. A model with a limited context window wastes space on all of that. A short, clean markdown file that says "here is who we are and here are the pages that matter" is easier for a model to use.
The proposal also suggests a second piece: serving a clean markdown version of each important page at the same URL with .md added (for example /pricing.md). The llms.txt file then links to those markdown versions.
It helps to be clear about what llms.txt is not:
- It is not robots.txt. It does not allow or block any crawler.
-
It is not a sitemap. A sitemap lists every URL for search engines. llms.txt is a short, curated reading list.
-
It is not an official standard from the IETF, W3C or any AI company. The spec is maintained in public on GitHub and is open for community input.
What does an llms.txt file look like?🔗
An llms.txt file follows a fixed order, according to the spec:
-
An H1 with the name of the project or site. This is the only required part.
-
A blockquote with a short summary of what the site is.
-
Optional paragraphs or lists with more detail.
-
Zero or more H2 sections, each holding a list of links in the form
<a href="url" class="text-blue-600 hover:text-blue-800 transition-colors">name</a>: optional note.
A section named ## Optional has a special meaning by convention: it holds secondary links an agent can skip when it needs a shorter context.
Here is a short example for a made-up US payroll software company:
# Acme Payroll
> Acme Payroll is payroll and HR software for US small businesses with 5 to 100 employees. Plans start at $40 per month plus $6 per employee.
Acme runs payroll in all 50 states, files federal and state payroll taxes, and supports W-2 employees and 1099 contractors.
## Product
- <a href="https://acmepayroll.com/pricing.md" class="text-blue-600 hover:text-blue-800 transition-colors">Pricing</a>: Plans, per-employee fees, what each tier includes
* <a href="https://acmepayroll.com/features.md" class="text-blue-600 hover:text-blue-800 transition-colors">Features</a>: Payroll, tax filing, PTO tracking, benefits
* <a href="https://acmepayroll.com/integrations.md" class="text-blue-600 hover:text-blue-800 transition-colors">Integrations</a>: QuickBooks, Xero, Gusto migration
## Help
- <a href="https://acmepayroll.com/docs/start.md" class="text-blue-600 hover:text-blue-800 transition-colors">Getting started</a>: Set up your first pay run
* <a href="https://acmepayroll.com/docs/state-taxes.md" class="text-blue-600 hover:text-blue-800 transition-colors">State tax guide</a>: Rules by state
## Optional
- <a href="https://acmepayroll.com/blog" class="text-blue-600 hover:text-blue-800 transition-colors">Blog</a>: Payroll tips for small business owners
* <a href="https://acmepayroll.com/about" class="text-blue-600 hover:text-blue-800 transition-colors">About</a>: Company history and team
Real examples are easy to find. Stripe's docs, Cloudflare's developer docs and Anthropic's Claude platform docs all publish one. Cloudflare also publishes a much larger llms-full.txt with the full text of its docs. Notice the pattern: most well-known adopters are developer documentation sites, where coding agents read docs constantly.
Do AI engines actually read llms.txt?🔗
As of October 2026, no major AI search engine has publicly said it uses llms.txt to choose sources or recommend brands, and the public evidence points the other way. Here is what is on the record:
| Source | What it says | Date |
|---|---|---|
| John Mueller, Google (via Search Engine Roundtable) | "FWIW no AI system currently uses llms.txt." He suggested checking your server logs to see for yourself. | June 2025 |
| Google Search Central AI optimization guide | You don't need new machine-readable files or AI text files to appear in Google Search. Creating llms.txt "will neither harm nor help" your visibility, "as Google Search ignores them." | Updated July 2026 |
| Chrome Lighthouse llms.txt audit | Lighthouse checks for the file in an agentic browsing category. A missing file is "Not Applicable." It only flags a server error when fetching the file. | May 2026 |
| SE Ranking study | Of nearly 300,000 domains, 10.13% had an llms.txt file. No measurable relationship with how often AI systems cite a domain. Removing the variable made their model slightly more accurate. | 2026 |
Two things are worth separating here.
First, AI search answers. Google is explicit that its Search features, including AI Overviews and AI Mode, ignore the file. OpenAI, Perplexity and Anthropic have not published statements saying their answer engines use llms.txt for ranking or citation. The SE Ranking data found no correlation. If your goal is "get ChatGPT to recommend us," llms.txt is not the lever.
Second, agents and developer tools. This is where the file has a plausible job. A coding assistant that needs your API docs, or a browsing agent trying to understand your site, can be given your llms.txt link and follow it to clean markdown. Chrome added a Lighthouse audit for it under agentic browsing, which signals some interest in agent readiness, but the audit treats the file as optional. That is a different use case from earning citations in answers.
So the honest answer: llms.txt is a low-cost courtesy to agents, not a ranking signal. Anyone selling it as a shortcut to AI citations is ahead of the evidence.
Should you create an llms.txt file?🔗
Create one if it is quick for you and your site has content that agents or developers would want to read in clean form. Skip it, or put it at the bottom of the list, if your time is better spent on the things that do move AI visibility.
It makes the most sense for:
- Software companies with docs or an API. The proposal itself notes that coding agents now fetch library docs routinely to get API calls right. A clean index helps them.
-
Sites with dense reference content, such as pricing, specs, policies or knowledge bases, where a markdown version is easier to parse than a heavy page.
-
Teams that already generate markdown, since docs platforms such as Mintlify and GitBook can produce llms.txt automatically (both are listed as tools on llmstxt.org).
It matters less for a local business, a small ecommerce store, or a brand whose main goal is being named in "best X for Y" answers. For those goals, the work that counts is the same work covered in our guides on how to get cited by ChatGPT and content strategy for AI citations: clear pages that answer buyer questions, comparison content, mentions on sites AI engines already cite (including Reddit threads), and clean structured data like the schema markup we cover separately.
TrendlyAI publishes its own llms.txt at trendlyai.com/llms.txt. I treat it the way I would treat a good README: useful for anyone, or any agent, that wants the facts in one place, but not something I count on to change how often we get cited.
How do you create an llms.txt file?🔗
You can write an llms.txt file by hand in a text editor in under half an hour. Here is the process:
-
Write the H1 and summary. Use your brand name exactly as it appears elsewhere. In the blockquote, say what you do, who it is for, and one or two concrete facts (price range, location, main product). Keep it factual. Marketing language helps no one here.
-
Pick 5 to 20 pages. Choose the pages a buyer or developer would need: pricing, product or service pages, docs, comparison pages, policies. This is a curated list, not your whole sitemap.
-
Group them under H2 headings such as Product, Docs, Pricing, Company. Add a short note after each link that says what is on the page.
-
Put secondary links under
## Optionalso agents with less room can skip them. -
Save it as
llms.txtat your root, so it loads athttps://yoursite.com/llms.txt. Serve it as plain text. -
Optionally add markdown versions of the key pages (
/pricing.mdand so on) and point the links at them. This is the part of the proposal that does the most work for agents, and also the part that takes the most effort. -
Keep it current. A file with old prices or dead links is worse than no file. Add it to whatever checklist you use when pricing or product pages change.
If you would rather not start from a blank file, the free TrendlyAI llms.txt generator drafts one from your site that you can edit and upload. Read the draft before you publish it, the same way you would read any copy that represents your company.
How you upload depends on your platform. On WordPress, you can upload the file to your site root with your host's file manager or SFTP. On Next.js, Astro and most static site generators, drop it in the public folder. On hosted builders that do not allow root files, you may not be able to add one at all, which is another reason not to treat it as essential.
How do you know if it is doing anything?🔗
The direct way to check is your server or CDN logs. Filter requests for /llms.txt and look at the user agents. If no AI crawler ever requests the file, it is not being read by that crawler, whatever anyone claims. This is the same check John Mueller suggested.
The indirect way is to watch your AI visibility over time and see whether anything changes after you publish. Be careful with that comparison: AI answers vary from one ask to the next, and you are probably shipping other changes in the same weeks. A brand mention tracker shows whether ChatGPT, Perplexity, Claude and Google AI Overviews name you for the prompts you care about, which is the outcome that matters. Our guide to measuring AI search traffic in GA4 covers the click side.
If you want a quick baseline before and after, the free AI visibility check runs 5 prompts on ChatGPT and Perplexity in about 30 seconds with no signup. For the bigger picture on what does move citations, start at the GEO hub.
Frequently asked questions🔗
Is llms.txt an official standard?🔗
No. llms.txt is a public proposal by Jeremy Howard, published at llmstxt.org in September 2024 and maintained openly on GitHub. No standards body has adopted it, and no major AI search engine has publicly committed to using it for ranking or citations.
Does Google use llms.txt?🔗
Google Search does not. Google's Search Central AI optimization guide says Google Search ignores llms.txt and similar files, so they will neither help nor hurt your visibility in Search, AI Overviews or AI Mode. Separately, Chrome's Lighthouse tool has an optional llms.txt audit for agent readiness.
Will llms.txt help my brand get cited by ChatGPT or Perplexity?🔗
There is no public evidence that it will. SE Ranking's study of nearly 300,000 domains found no measurable relationship between having llms.txt and AI citation frequency. Content that answers buyer questions and mentions on sources AI engines already cite are far better bets.
What is the difference between llms.txt and robots.txt?🔗
robots.txt tells crawlers which URLs they may or may not fetch. llms.txt does not control access at all. It is a short markdown summary with links to your most useful pages, meant to help AI agents read your site efficiently.
What is llms-full.txt?🔗
llms-full.txt is a convention some sites use alongside llms.txt. Instead of links, it contains the full text of the docs in one large markdown file. Cloudflare's developer docs publish one. It is not part of the core llms.txt format described on llmstxt.org.