llms.txt is a plain-text file at the root of a website (yourdomain.com/llms.txt) that gives AI assistants a short description of the business and a list of the pages worth reading, in a form a language model can take in at once. It was proposed in September 2024.1 Some AI tools read it; the big search-backed assistants have not committed to it.
For a contractor it is ten minutes, cannot hurt, and is not a ranking signal or a way to get cited on its own.
What it looks like
A heading with the business name, a one-paragraph description, then sections of links with a line each. Ours is at /llms.txt; a roofer's might read:
# Stormline Roofing > Roof repair, replacement and storm-damage insurance claims in Schenectady County, NY. Licensed and insured, GAF-certified, since 2004. ## Services - [Roof replacement](https://example.com/roof-replacement/): asphalt and metal, with the warranty explained - [Storm damage and insurance claims](https://example.com/insurance-claims/): the five steps, who meets the adjuster ## Contact - Phone: (518) 555-0100, same-day inspections
Who reads it
Developer tools and some smaller assistants fetch it; it is popular with documentation sites. OpenAI, Google and Anthropic have not said their search products use it, and their crawlers read your ordinary pages regardless. So the honest position in late 2026 is: it is cheap insurance for a future where more assistants read it, and it is a useful discipline, because writing it forces you to state in two sentences what the business does and where.
What matters more
The things the AI answers guide lists first: allowing the AI crawlers by name in robots.txt (see whether blocking hurts), a plain statement of trade, towns, license and reviews on the first screen, LocalBusiness structured data, dates and an author on pages, and a Google Business Profile that matches the site. The free audit checks for llms.txt as one line of its answer-engine section, after those.
Does Google use llms.txt?
Google has not said it does, and its AI Overviews work from the ordinary Googlebot index. Treat it as a file for other assistants and for the future, not as a Google signal.
Is it the same as robots.txt?
No. robots.txt tells crawlers what they may fetch; llms.txt tells a model what to read first and what the site is. A site needs both to be right, and a robots.txt that blocks AI crawlers makes the llms.txt pointless.
Sources
- Answer.AI, September 2024: The /llms.txt file, the proposal and format.
- Google Search Central: Overview of Google crawlers and fetchers.
See this on your own site.
The free audit measures your site on a real phone: load time, the first screen, the call button, what Google's AI says about you next to the competitors. Written, back within one business day. No call.