llms.txt: what it is and whether you need one
The Doily team
Reviewed by the Doily team, Oct 2026 Updated 4 min read
Quick answer
llms.txt is a proposed markdown file at /llms.txt that gives AI agents a short, curated map of your site. Google says Google Search ignores it, so it neither helps nor harms your Google visibility. It is optional, cheap to publish, and most useful for documentation that agents read on demand.
Why it matters
The file gets sold as a shortcut to AI citations. The evidence is thinner than the pitch. Google has said in its documentation that Google Search does not use llms.txt, and added a clarifying note on Jun 15, 2026 after community questions.
That does not make the file useless. The proposal's author describes it as a way to give agents a clean entry point into documentation, and some documentation sites publish one. Whether it fits your site depends on what agents need from you.
Key takeaways
- Google Search ignores llms.txt. Keeping one will not change your Google rankings or AI Overviews visibility either way.
- The format is a proposal, not a standard that search engines enforce. The spec requires only an H1 with the site name.
- It suits documentation, product references, and sites agents read to complete tasks. A local service business gets little from it.
- Do the basics first: indexed pages, direct answers, and open access for search bots.
Step by step
-
Decide whether it fits your site
If you run software documentation, an API reference, or a large knowledge base, an agent-readable index can help tools that fetch pages on demand. If you run a local business or a small brochure site, put the time into your pages and listings first. Be honest about which one you are.
-
List the pages an agent should read first
Pick 10 to 30 pages: what you do, pricing or terms, key how-to pages, policies, and contact details. Write a 1-line note for each that says what the page contains. Leave out thin pages, duplicates, and anything outdated.
-
Write the file in the specified order
Start with an H1 that names the site, which is the only required part. Add a blockquote with a short summary, optional detail paragraphs, then H2 sections that hold lists of links with short notes. A section named Optional marks secondary links an agent can skip.
-
Point to clean versions of pages
The proposal also suggests offering a clean markdown version of each page at the same URL with
.mdadded. Do this only where you can keep the versions in sync. Linking to the normal HTML page is acceptable if you cannot. -
Publish it at the root and check access
Serve the file at yourdomain.com/llms.txt with a 200 status and a plain text or markdown content type. Make sure robots.txt and your firewall do not block it. Open it in a browser to confirm it renders as text.
-
Keep it current
A file that links to deleted or renamed pages is worse than none. Add it to your monthly page review and update it whenever you publish or retire a key page.
-
Measure it honestly
Check server logs for requests to /llms.txt to see which agents fetch it. Do not expect a ranking change, and do not report one to a client unless the logs and your tracking both support it.
An example file
Example data for a fictional business. The structure follows the proposal.
# Example Plumbing
> Licensed plumbing and drain service for homes in Hartford County.
> Emergency calls, repairs, and installs. Quotes are free.
Prices are listed as ranges and confirmed after inspection.
## Services
- [Drain cleaning](https://example.com/drain-cleaning): Methods, timelines, and price ranges
- [Water heater repair and install](https://example.com/water-heaters): Tank and tankless options
## Company
- [Service area](https://example.com/service-area): Towns we cover and response times
- [Contact](https://example.com/contact): Phone, hours, and request form
## Optional
- [Blog](https://example.com/blog): Maintenance advice and seasonal remindersKeep the file short enough that an agent can read it in full, and let the linked pages carry the detail.
How Doily helps
Doily's AI visibility tracking records whether engines mention and cite your business, so you can judge any change, including an llms.txt file, against your own prompt results.
Questions
Does Google use llms.txt?
No. Google says it does not use llms.txt or similar AI text files for Google Search, including its generative AI features. Creating one will neither help nor harm your visibility there, and it is fine to keep one for other services that use them.
Do ChatGPT, Claude, or Perplexity use my llms.txt?
Their crawler documentation, checked Oct 11, 2026, does not list llms.txt as an input for search or citations. OpenAI and Perplexity do point readers of their own documentation sites to an llms.txt index.
How is llms.txt different from robots.txt?
robots.txt tells crawlers what they may access. llms.txt is guidance an agent reads on demand when it needs information about a topic. One controls access and the other describes content.
Sources
Primary sources, checked Oct 11, 2026.
