What is llms.txt, and does my website need one?
A plain look at llms.txt: a proposed file for AI tools, not a ranking or citation directive, and worth publishing only if a site can keep it accurate.
llms.txt is a proposed Markdown file at a site's root that summarises the site for AI tools willing to read it. It is not a ranking or citation directive, and adoption among AI systems varies. Publish one only if the site can keep it accurate. Inciteful publishes its own, generated from its page content, and keeps it in step as the site changes.
The proposal, in plain terms
The llms.txt proposal puts a Markdown file at a site's root, normally /llms.txt. It opens with the site or project name, can add a short summary, then lists sections of labelled links to pages worth reading. It is designed to sit alongside robots.txt, a sitemap and structured data, not replace them.
A proposal is not a standard every tool follows
No specification can force a model or an AI application to read a file, and adoption varies between them. llms.txt is not a ranking signal and not a citation directive: publishing one does not commit any AI system to using it. A site should publish one only when it can keep the content accurate and in step with the pages people and crawlers can actually reach. A stale file is worse than none.
What each discovery file is actually for
robots.txt, a sitemap and llms.txt each do a different job, and none of them controls what an AI system does with the pages it can reach.
| File | Its job | Its limit |
|---|---|---|
| robots.txt | Tells compliant crawlers, including AI ones, which parts of a site they may fetch | Does not describe the site, and does not promise that an allowed page gets fetched |
| Sitemap | Lists a site's canonical, indexable URLs for discovery | Not a summary, and listing a page does not get it indexed |
| llms.txt | A short Markdown overview and a set of links, for tools that choose to read it | Adoption varies, and it is not a ranking or citation signal |
| Structured data | Describes a page's entities in a machine-readable format | Must match what is actually on the page, and earns no special treatment on its own |
Where llms.txt fits inside Inciteful
Inciteful's own site publishes an /llms.txt file, generated from its page content, so it stays in step with what the pages actually say. The AEO Boost's audit notes whether a client's site publishes an llms.txt file, as an observation alongside the crawl, not as a scored or blocking check. Neither treatment makes the file a route to a citation on its own.
What actually earns a citation, so build that first
Google's own generative-search guidance points back to ordinary foundations: content worth citing, a clear technical structure, and pages a crawler can reach and read, and it warns against content produced mainly to influence rankings or AI answers. An llms.txt file does not change any of that. It is a small signpost, not a substitute for the page itself.
What site owners ask about llms.txt
Does Google use llms.txt for AI Overviews?
Google's published generative-search guidance is about ordinary search foundations and does not name llms.txt as a requirement. Treat the file as a courtesy to tools that read it, never as something that secures inclusion on its own.
Is llms.txt the same as robots.txt?
No. robots.txt tells crawlers which parts of a site they may fetch. llms.txt is a proposed Markdown overview and link list, for AI applications that choose to read it. Both can exist on the same site and do different jobs.
Can llms.txt improve citations?
There is no reliable basis for that claim. A clear file may help a compatible tool find the right pages faster, but the pages still have to be accessible, accurate and worth citing on their own. llms.txt cannot make a weak page win a citation.
Does Inciteful check llms.txt?
The AEO Boost's audit notes whether a site publishes an llms.txt file, as an observation, not a scored or blocking check. Inciteful's own site publishes one, generated from its page content, so it stays accurate as the pages change.
Should every site publish an llms.txt file?
Only if the site can keep it accurate. A file that lists pages which have moved, or a summary that no longer matches the site, is worse than no file, because it hands a wrong answer to any tool that reads it.
What should a site put in its llms.txt file?
A short, factual summary and links to the pages that best explain the site: the homepage, key product or policy pages, and anything worth a tool reading first. Leave out anything stale, and review it whenever the site's structure or content changes.



