llms.txt explained: what it is and whether to add one

llms.txt is a plain markdown file served at the root of a site. It was proposed in September 2024 by Jeremy Howard of Answer.AI. The idea is to give language models a short, curated map of your site so they do not have to wade through a navigation-heavy HTML page to find what matters.

The format

The proposal asks for a file at /llms.txt containing, in order: an H1 with the site or project name, a blockquote summarizing it, optional paragraphs of context, then H2 sections holding lists of links, each with a short description.

# Example Co

> Example Co makes invoicing software for freelancers.

## Docs

- [Getting started](https://example.com/docs/start): Set up your first invoice
- [Pricing](https://example.com/pricing): Plans and limits

## Optional

- [Changelog](https://example.com/changelog)

Does it affect citations?

As of this writing we are not aware of any major AI search vendor confirming that llms.txt is used to decide which pages to retrieve or cite. Some documentation tools and coding assistants do read it, but that is not the same as a ranking input. Treat any claim that llms.txt "boosts AI visibility" as unproven.

Why add one anyway

• It takes about ten minutes.
• It forces you to decide which pages are your best answers, which is useful regardless of any crawler.
• Audit tooling is starting to look for it. Google Lighthouse's agentic browsing audit checks for llms.txt, so its absence can show up in reports.
• It is harmless. A static text file cannot hurt your rankings.

What matters more

If you only have time for one thing, make sure search crawlers can reach your pages, put direct answers high on the page under clear headings, and mark up your entities with JSON-LD. Those are documented, durable inputs. llms.txt is a cheap extra on top of them, not a substitute.

Our own llms.txt lives at /llms.txt, and the free Cited Lens audit will tell you whether yours is detected.