No, you do not need an llms.txt file to be found in Google or cited in its AI features. Google says you don't need new machine readable files or AI text files to appear in AI Overviews and AI Mode. llms.txt is a proposal, so treat it as a cheap extra for documentation-heavy sites, not a ranking lever.
What is llms.txt?
llms.txt is a plain Markdown file, usually placed at /llms.txt, that gives AI agents a short, curated guide to your site. The proposal at llmstxt.org comes from Jeremy Howard of Answer.AI and was first published on September 3, 2024.
The format is simple. The only required part is an H1 with your site name. After that you can add a short summary in a blockquote, then H2 sections that list links, each with an optional note. An "Optional" section marks links an agent can skip when it needs less context.
It is built to sit alongside robots.txt and sitemap.xml, not replace them. Robots.txt says what crawlers may fetch. A sitemap lists what exists. llms.txt tries to say which pages matter most and what they are about.
Is llms.txt a standard?
No. The llmstxt.org page itself calls it a proposal and says the specification is open for community input. It also reports real use: thousands of sites publish one, and some documentation platforms generate it for you. Adoption is not the same as proof that any search or answer engine reads it when choosing what to cite.
What does Google say about it?
Google's page on AI features in Search is direct. To appear as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to show with a snippet. There are no additional technical requirements. Google also says you don't need to create new machine readable files, AI text files or markup for these features, and that there is no special schema you must add.
The page does not mention llms.txt by name. The sentence about AI text files is the closest statement, and it points the same way: the basics decide whether you appear.
When is an llms.txt file worth adding?
Use this short test before spending time on it.
- Do you have documentation, an API or a long list of guides? A curated link list is most useful when there is a lot to choose from. A five-page marketing site gains little.
- Do developers use coding assistants to read your docs? These are the agents the proposal was written for, since they fetch pages on demand and benefit from a clean index.
- Is your core content already indexable? If key pages are blocked, thin or missing from Google, fix that first. A file cannot rescue them.
If you answered yes to the first two, it costs an hour and carries little risk. If not, spend that hour on the work in the next section.
What to do before llms.txt
- Put the answer first. Open each page with a direct answer in a couple of sentences. Snippets and AI answers both quote the clearest passage.
- Keep important content as text. Google lists this among the fundamentals for its AI features, along with letting Google crawl the page and using internal links.
- Check robots.txt. A blocked page cannot be crawled. Our guide on allowing ChatGPT and Perplexity in robots.txt covers the user agents.
- Match markup to visible text. Structured data should describe what is on the page, nothing more. See whether FAQ schema is still worth adding.
How to write one if you decide to
Start with your site name as the H1. Add one sentence saying what the site is. List your five to fifteen most useful pages under a heading such as "Guides", each with a link and a short note on what it answers. Put anything secondary under "Optional". Save it as plain text at the root of your domain and open it in a browser to confirm it loads.
Keep it current. An outdated index that points at removed pages is worse than none, so add it to whatever checklist you use when you publish or retire a page.
A worked example
Say you run a developer tool with a docs site of 80 pages. Your llms.txt might have the tool name as the H1, a one-sentence summary, a "Getting started" section with three links, an "API reference" section with the main endpoints, and an "Optional" section holding changelogs and legacy pages. An agent that reads it learns where to begin and what it can skip.
Now say you run a three-page site for a local service. The file would list your home page, services and contact page, which any crawler finds in seconds anyway. The effort is better spent writing one clear page for each question your customers ask.
Frequently Asked Questions
Does llms.txt improve Google rankings?
There is no evidence in Google's documentation that it does. Google says no extra files or markup are needed for its AI features.
Where does the file go?
At the site root as /llms.txt in most cases. The proposal also allows one at a subpath, such as a docs folder, where the most specific file applies.
What must the file contain?
Only an H1 with the project or site name. The summary, notes and link sections are optional.
Can llms.txt replace robots.txt or a sitemap?
No. It is designed to work alongside them. Robots.txt controls crawler access and the sitemap lists your URLs.
Is it risky to add one?
The downside is small. The real cost is keeping it accurate, so only publish one you will maintain.