Sitearc

What llms.txt is, and what it will not do

The file, the convention behind it, a filled example, and the limits of an advisory format.

Answers
Explain llms.txt without implying that serving one changes an answer.
Included file
A filled llms.txt for one site, with the reasons
Written for
Developers asked to add an llms.txt file

What the file is

llms.txt is a plain text file at the root of a site. It lists the pages an automated reader could start from.

It is a convention, not a standard with an enforcement mechanism. A reader may use it, ignore it, or never request it.

What it is good for

  1. Naming the canonical starting point for a reader.
  2. Listing the pages you consider most useful, in a stable order.
  3. Stating boundaries in plain language, where a human also reads them.

A filled file

Cedar and Slate Interiors, written on 2026-09-22.

# Cedar and Slate Interiors

> A two-person joinery studio, working on kitchens, bathrooms
> and fitted furniture.

Canonical site: https://cedarandslate.example
Contact: https://cedarandslate.example/contact

## Read this first
- Services: https://cedarandslate.example/services
- Kitchens: https://cedarandslate.example/services/kitchens
- Pricing approach: https://cedarandslate.example/pricing

## Boundaries
- We quote after a site visit. No price is published per project.
- We work within 40 miles. No delivery outside that radius.

What it will not do

It will not make a model cite you. It will not train anything, block anything, or change how an answer is written.

A reader that ignores robots.txt will ignore this file too.

Acceptance: the file is right when

  • it is served at the root, as plain text, with a 200
  • every URL in it returns 200 and is indexable
  • the first line names the site, and the quote line says what it is
  • boundaries use the same words as the rest of the site
  • it is regenerated when the page list changes

Troubleshooting

Symptom Likely cause What to do
The file 404s It was never copied into the build Add it to the static assets
A listed URL 404s A page was renamed Regenerate the list from the sitemap
Nothing changed after adding it Expected: it is advisory Measure visibility separately
A tool calls the file invalid It is not a formal standard Keep it readable, and ignore the grade
The file duplicates the sitemap Two lists now exist Keep the short list, and point at the sitemap

Limits

Treat the file as a courtesy to a reader, not as a control.

Measure visibility separately, and label it as an estimate. The answer you see is one answer among many.

Where this came from