llms.txt: The robots.txt for Language Models
One small file tells AI systems what your site is and which pages matter. What llms.txt is, what it can and can't do, and how to write yours.
Published
llms.txt is a plain-text file at the root of your website (yoursite.com/llms.txt) that gives AI systems a short, human-readable summary of what your site is and links to the pages that matter most. Think of it as a cover letter for language models: where robots.txt tells crawlers what they may access, llms.txt tells AI tools what’s worth reading.
It’s simple to create, costs almost nothing to maintain, and fits naturally into the work of making a site easy for AI to understand. It’s also worth being clear-eyed about what it is: a proposed convention, not an official standard.
Where it came from
The llms.txt format was proposed in 2024 by Jeremy Howard of Answer.AI. The reasoning: web pages are cluttered with navigation, scripts, and layout that make them hard for language models to read efficiently, and models work best with concise, well-organized text. A single curated file solves that by pointing models to the right content in a format they handle well.
Since then, it has been adopted by many documentation sites, developer tools, and a growing number of businesses.
What it can do, and what it can’t
What it can do:
- Give AI tools a clear, accurate summary of your business in your own words
- Point them to your most important pages: services, pricing, locations, FAQs
- Reduce the chance an AI system misunderstands what you do
What it can’t do (yet):
- Guarantee that any AI company reads it. Major AI providers haven’t committed to using llms.txt in their crawlers or answers, and Google has said it doesn’t use it for search.
- Replace the fundamentals. Clear pages, structured data, and consistent information everywhere matter far more.
- Control or block anything. For access control, you still use
robots.txt.
Our take: it’s a small, low-risk addition that makes your site easier for AI tools to use, especially tools and agents that look for it. Just don’t treat it as a ranking trick.
What goes in it
The proposed format uses Markdown:
- An H1 with your business or site name (the only required part)
- A blockquote with a one- or two-sentence summary
- Optional paragraphs with key details
- H2 sections with lists of links to important pages, each with a short description
An example for a local business
# Example Family Dental
> Family and cosmetic dental practice in Westlake Village, California,
> serving the Conejo Valley since 2005. New patients welcome; most PPO plans accepted.
Open Monday–Friday 8am–5pm and Saturday 9am–1pm.
Phone: (805) 555-0100. Address: 123 Example Rd, Suite 4, Westlake Village, CA 91361.
## Services
- [Cleanings and exams](https://www.example.com/services/cleanings): preventive care for adults and kids
- [Dental implants](https://www.example.com/services/implants): single-tooth and full-arch implants, from consultation to restoration
- [Clear aligners](https://www.example.com/services/aligners): orthodontic treatment for teens and adults
## Patients
- [Insurance and financing](https://www.example.com/insurance): accepted plans, membership plan, and payment options
- [New patient information](https://www.example.com/new-patients): what to expect at your first visit
- [Book an appointment](https://www.example.com/book): online scheduling
## Optional
- [About our team](https://www.example.com/about)
- [Patient reviews](https://www.example.com/reviews)
Tips for writing yours
- Write plainly. No marketing fluff; state facts a model can repeat accurately.
- Lead with what matters: who you are, what you do, where, and for whom.
- Link to canonical pages and keep descriptions to one line.
- Keep it current. Update it whenever your services, hours, or locations change.
- Make sure it matches your site. Contradictions undermine trust.
How we handle it
Every site we build includes an llms.txt file written in plain language, generated from the same source as the site so it never drifts out of date. It’s one part of a larger picture that includes structured data, answer-first content, and fast, accessible pages. See our AI SEO and agent-ready services.