What is llms.txt?
llms.txt is a proposed web standard (llmstxt.org, proposed by Jeremy Howard in
2024): a markdown file at your site root that gives AI assistants a curated,
token-efficient map of what matters on your site — instead of forcing them to guess from
crawling raw HTML full of nav, cookie banners and scripts.
The format
# Your Site Name > One sentence saying what this site is and who it's for. ## Main pages - [Pricing](https://example.com/pricing): plans and what each includes - [Docs](https://example.com/docs): integration guide and API reference ## Optional - [Blog](https://example.com/blog): weekly product updates
That's the whole spec, essentially: an H1 name, a blockquote summary, and sections of annotated
links. A companion llms-full.txt can inline full page content for heavy use-cases.
Does it actually matter in 2026?
Adoption is real but uneven. Documentation-heavy products (developer tools especially) adopted it fastest, some AI dev tools fetch it natively, and several crawlers request it opportunistically. Nobody serious claims it's a ranking factor. The case for shipping one is asymmetric: it costs minutes, it's the single clearest signal of "this site is curated for machine readers," and when an assistant does fetch it, you — not a scraper's guess — control the summary it reads.
llms.txt vs robots.txt
robots.txt says what crawlers may do (permission). llms.txt
says what they should read first (curation). They're complements: policy at the gate,
map inside the door.
Generate yours in 10 seconds
Our audit drafts a spec-valid llms.txt from your sitemap and
homepage links — you edit the descriptions and serve it at /llms.txt.