Fix · AI access

Does llms.txt help?

Short answer: no provider documents reading it, Google says you do not need one, and it is harmless. Here is what it is and what to do instead.

What it is

llms.txt is a proposed convention: a Markdown file at the root of a site that summarises the site and lists its most useful pages, written for language models rather than browsers. It was put forward in 2024 by a developer-tools community, and some documentation sites publish one. It is not a standard and no standards body has adopted it.

What the providers say

  1. Google. Its guide to AI features states: "You don't need to create new machine readable files, AI text files, or markup to appear in these features." Google's AI Overviews and AI Mode draw on the normal search index.
  2. OpenAI, Anthropic, Perplexity. Each publishes a page describing its crawlers and how to control them with robots.txt. As read on 22 September 2026, none of those pages mentions llms.txt.

So the honest position is: there is no documented mechanism by which the file changes whether you are found, understood or cited. There is also no documented harm.

What to do with the hour

  1. Make the HTML say who you are. One Organization or LocalBusiness node with name, address, telephone and a register identifier. The identity fix.
  2. Make sure the answer crawlers are allowed. Search crawlers and training crawlers are different tokens. The access fix.
  3. Put the answer in the first paragraph. A machine quotes what it can lift. The answer-shape fix.

If a developer has already written one, leave it. If a vendor is charging you for one, ask them which provider documents reading it.

Prove it. This is one fix where the test is a question: ask the vendor or the article proposing the file to name the provider page that says it is read. As at the date above, we could not find one.

What it will not do

This page will be wrong the day a provider documents support. We re-read the provider pages monthly and will change the answer here, with the date, when that happens.

Sources: Google Search Central, "AI features and your website"; OpenAI crawler documentation; Anthropic crawler support article; Perplexity crawler guide. All read 22 September 2026.