All posts
Guide

Should you publish llms.txt if your site has no public docs? Usually not for visibility, maybe for governance

By Janis Plume, Founder, Outbound Pros · 8 min read · 2026-11-08

Quick answer

Usually no. If your site has no public docs, publishing llms.txt is unlikely to improve AI visibility, search performance, or citations. Google states llms.txt is not used by Search, and observed adoption remains limited. Publish it only if you want a simple public note about preferred sources, scope, or usage expectations. Treat it as governance documentation, not a discovery or citation tactic.

Why is llms.txt a weak move when you have no public docs?

llms.txt makes the most sense when a site already has a library of public source material that needs a tidy front door. Think docs, policies, research hubs, product references, changelogs, or technical explainers spread across many URLs. If you do not have that layer, the file has very little to point at besides homepage, sales pages, and generic marketing copy.

That matters because citation does not come from a file existing. Citation comes from assistants being able to fetch, parse, and trust the underlying pages. A neat manifest cannot rescue weak source pages. It cannot make vague copy suddenly quotable. It cannot fix contradictions between your pricing page, help center, and blog if those pages do not exist in the first place.

This is where teams get distracted. They see llms.txt as an easy checkbox and mistake it for infrastructure. It is not infrastructure in the way crawlable HTML, clean page structure, explicit facts, and stable canonical URLs are infrastructure. It is closer to a note you leave on the front desk. Useful in some buildings, irrelevant in others.

There is also a harder reality. Google states llms.txt is not used by Search. A large domain study found adoption at 10.13% across roughly 300,000 domains, with 0% among the top 1,000 sites, and no citation lift after controls. That does not mean nobody will ever look at the file. It means you should not sell yourself a fantasy that this is moving core discoverability.

If you need the broader context on where the file does and does not matter, read our guide on whether to publish llms.txt at all.

What should you do instead if you want AI assistants to cite you?

Start with pages that contain stable, direct, quoteable facts. If you have no public docs, your first move is not an index file. Your first move is creating source pages worth indexing and citing. That usually means clear definitions, policy pages, methodology pages, product facts, service scope, or evidence pages that answer one question cleanly.

In practice, the best order is boring. Put the answer first. State the entity name near the claim. Keep qualifiers attached to the claim. Put key facts in visible HTML, not inside images, tabs, hover states, or widgets. Then make sure the page is server rendered enough that crawlers can fetch the content without needing a browser to execute code.

That last point is not theory. Verified server log work showed AI crawlers fetch JavaScript files and never run them. So if your site is hiding the important material behind client side rendering, llms.txt will not save you. The crawler can discover the file and still miss the actual facts on the page.

If you need a practical walkthrough on this failure mode, start with rendering before anything else.

See whether AI crawlers execute JavaScript or only fetch files for the operational version of that argument.

  • Create a small set of canonical fact pages before you create a manifest for them
  • Use plain, visible HTML for facts you want repeated accurately
  • Separate current facts from historical or proposed information
  • Keep one claim on one canonical URL where possible
  • Add supporting schema only where it clarifies visible page meaning, not as a substitute for weak copy

When does publishing llms.txt still make sense without docs?

There are a few cases where I would still publish it, even on a site without a real docs layer.

First, governance. Some teams want a simple public statement saying which pages represent the current source of truth, which sections are archival, and where machine readable consumption should start. That will not create visibility on its own, but it can reduce internal confusion. It also gives legal, product, and marketing a shared artifact to point at when they argue about what should be exposed.

Second, transition states. Maybe you are actively building a knowledge base, policy center, or research section, but it is not complete yet. In that case, an llms.txt file can act as a temporary map to the handful of pages that already deserve to be treated as canonical. The key is honesty. Do not stuff it with every URL on the site just to look complete.

Third, multi team sites where ownership is blurry. On those sites, llms.txt can function as a slim editorial contract. It tells people which pages should carry enduring factual statements and which pages are campaign material, announcement copy, or narrative pages that should not be relied on for strict retrieval.

SituationPublish llms.txt?Why
No docs, only marketing pagesUsually noThere is little high trust source material to map, so the file adds almost nothing
No docs yet, but a few canonical policy or product pagesMaybeIt can point to the limited pages that should be treated as current sources
Docs or research hub launching soonYes, if maintainedIt can become a tidy map once real source pages exist
Complex site with multiple teams and mixed page purposesMaybeUseful as governance documentation, not as a citation growth tactic

What are teams getting wrong about llms.txt right now?

They are using it as a substitute for editorial discipline.

The pattern is easy to spot. A site has thin service pages, contradictory product claims, no stable definitions, and key details hidden in accordions or app interfaces. Instead of fixing source quality, the team publishes llms.txt and declares itself ready for AI search. That is cosmetic work dressed up as technical strategy.

Another mistake is copying unsourced GEO folklore. You have probably seen dramatic multiplier claims attached to tables, FAQ schema, or freshness tweaks. Do not build your operating plan on numbers nobody can verify. If a tactic matters, you should be able to explain the retrieval reason in plain language. For llms.txt, the retrieval reason is weak unless the underlying pages are already strong.

One more miss, teams confuse machine friendliness with machine preference. Even if a crawler can access your file, that does not mean an assistant will prefer your pages in answers. Preference usually comes from clearer facts, better page focus, cleaner retrieval, and fewer contradictions than the alternatives. That is why aggregator pages still beat original sources so often.

Who should not follow the advice to skip llms.txt?

Do not follow the skip it advice if your site actually does have source material, but you are calling it something else. Plenty of companies say they have no docs when they really have policy pages, implementation notes, release updates, public standards, partner requirements, or detailed support articles. That is source material. In that case, llms.txt may be reasonable, especially if those pages are buried.

Also do not follow it if you are in a regulated or highly disputed information category where public source designation matters on its own. There the file can help clarify what is current, what is archived, and what should not be relied on. Again, that is a governance use, not a visibility promise.

And if you are operating a large brand site with several web teams, publishing the file may simply be worth it because the maintenance cost is low once ownership exists. I still would not expect citation lift from the file itself. I would expect organizational clarity, which can indirectly improve the pages that do earn citations.

What is the practical decision rule here?

Use a simple rule. If you cannot name the specific public pages that should be treated as your canonical source set, you are not ready for llms.txt. Build those pages first. If you can name them, and they are stable, crawlable, and maintained, then publishing the file is fine, but still secondary.

My operator view is blunt. On a site without public docs, llms.txt is rarely the next best move. It is usually a low consequence side task that gets over promoted because it feels technical and easy. The harder work is deciding what your public source of truth should be, writing it clearly, and making it retrievable without client side dependencies.

That advice fails if your real problem is internal governance rather than retrieval. In that case, llms.txt can be worth publishing earlier. It also fails if your site is tiny and the maintenance overhead of the file is effectively zero. Then the downside is small, but the upside is still mostly administrative, not visibility related.

If your main objective is pipeline from outbound execution, that is a different problem and belongs with the parent team, not here. We run managed outbound under Outbound Pros, but execution strategy lives there while this site stays focused on how AI systems find, extract, and cite source material.

If you want help deciding which public pages should become your source of truth, you can book a working session here: book a call.

Common questions

Can llms.txt hurt anything if I publish it anyway?

Usually not, if it is accurate and maintained. The risk is mostly strategic distraction. Teams publish it, feel done, and delay the source page work that actually affects extractability and citation.

Should I list my homepage in llms.txt if I have no docs?

You can, but it rarely adds value. A homepage is often broad, promotional, and constantly changing. It is not usually the best place for stable facts assistants can quote cleanly.

Is llms.txt useful for Google Search?

Google states llms.txt is not used by Search. So do not treat it as an SEO lever for rankings or indexing.

What should I build before llms.txt?

Build a small set of canonical public pages with direct answers, stable facts, clear scope, and visible HTML content. Then make sure crawlers can fetch those pages without needing JavaScript execution.

Does llms.txt help if my site is JavaScript heavy?

Not by itself. Verified evidence shows AI crawlers fetch JavaScript files and never run them. If the important page content depends on execution, the file does not fix the retrieval problem.

Last updated: 2026-11-08

Talk through your AI visibility with people who measure it

30 minutes. We will look at what assistants can actually retrieve from your site and tell you plainly what is worth fixing first.

Book a strategy call

30 minutes, no obligation. The calendar shows real availability.

Or start with the free GTM audit from Outbound Pros