All posts
Guide

Should you publish llms.txt on subdomains? Or keep it only on the main domain

By Janis Plume, Founder, Outbound Pros · 9 min read · 2026-11-14

Quick answer

Publish llms.txt on each public subdomain that contains standalone documentation or source material you want discovered. Do not rely on a root domain llms.txt to speak for docs, help, academy, or app subdomains. Also, do not expect SEO or citation lift from the file itself. Google says it is not used by Search, and a study across about 300,000 domains found 10.13% adoption, 0% among the top 1,000 sites, and no citation lift after controls.

What is the practical rule for llms.txt across subdomains?

Treat each public subdomain as its own publishing surface. If a subdomain contains content that could reasonably be cited, summarized, or used for grounding, publish a dedicated llms.txt there. If a subdomain has no public source content, skip it.

That means docs.example.com, research.example.com, or help.example.com may justify their own file. A marketing root domain file should not be your only implementation if the real factual material lives elsewhere.

The key point is scope, not magic. llms.txt is a directory hint at best. It does not fix weak pages, hidden facts, bad rendering, or contradictory claims across your estate.

Why is root only usually the wrong implementation?

Because root only assumes AI systems will infer your whole content architecture from one file on one host. That is a shaky assumption. Subdomains often operate like separate properties in crawl behavior, ownership, templates, and content quality.

If your product facts live on a docs subdomain, your policy details live on a help subdomain, and your root domain is mostly brand and sales copy, then a root llms.txt creates distance between the file and the material you actually want surfaced.

I would rather make the path obvious on the host where the useful content lives than hope an assistant follows a root file, traverses across hosts, and keeps the right page relationships intact.

  • Publish on the root domain if the root domain holds the main public source material
  • Publish on each public subdomain if that host contains independent documentation or factual reference pages
  • Skip private app subdomains, staging hosts, and thin marketing microsites with no real source content
  • Do not use llms.txt as a substitute for crawlable HTML, stable URLs, and clear on page facts

What does the actual evidence say about llms.txt?

Here is where people usually overclaim. The evidence does not support treating llms.txt as a visibility lever by itself.

Google states llms.txt is not used by Search. A study across about 300,000 domains found 10.13% adoption, 0% among the top 1,000 sites, and no citation lift after controls. That should reset expectations fast.

So why publish it at all? Because a lightweight directory can still help humans, internal teams, and some downstream systems understand where your curated public materials live. The file can be useful as housekeeping. It is just not the growth shortcut people want it to be.

This is the same problem I see with a lot of GEO advice. People take a weak signal, wrap it in certainty, then ignore the harder work of making pages extractable. The harder work is usually where the result comes from.

If you need the broader decision on whether to bother with the file at all, read should you publish llms.txt or ignore it. If your issue is that the right content sits on a JavaScript heavy host, fix rendering first with server side rendering vs client side AI visibility.

Which subdomains actually deserve their own llms.txt?

Use a simple test. Would you be comfortable with an AI assistant citing pages from that host as the canonical source for a customer question? If yes, that host may deserve its own llms.txt. If no, leave it out and fix the content first.

Subdomain typePublish llms.txt?Why
docs or help centerUsually yesContains stable how to, product, API, or policy information
research or resourcesUsually yesMay host original source material worth citing directly
root marketing siteMaybeOnly if it holds meaningful factual pages, not just sales copy
app subdomain behind loginUsually noLimited public access and weak citation value
staging or test hostNoShould stay blocked, not curated for discovery
campaign micrositeUsually noThin, temporary, and rarely canonical

When does publishing on subdomains fail?

It fails when teams confuse the file with a content strategy. If the docs subdomain has vague headings, answer buried in tabs, duplicate variants of the same claim, or pages that depend on client side rendering, publishing llms.txt there changes very little.

It also fails when your canonical facts are split across hosts without governance. One team updates pricing language on the root site, another edits feature details in docs, and support publishes exceptions in the help center. Now the model sees three versions of the truth. A fresh llms.txt on every host does not solve that contradiction.

This advice also does not fit every company. If you run a tiny site with one host and a handful of pages, do not create complexity just to look sophisticated. Publish one file at the root if you want it, then spend your energy on source clarity.

And if your most important material is outbound execution playbooks, campaign operations, or channel orchestration, that belongs with our sibling properties, not here. We run managed outbound under Outbound Pros, but the execution detail is better handled on the parent site than folded into AI search guidance.

How should you implement this without making a mess?

Keep it boring. Each file should point to the best public source materials on that host. Do not stuff every blog post, campaign page, or changelog entry into it. Curation matters more than volume.

Implementation standard I would use

  • Publish a separate llms.txt at the root of each public subdomain with durable source content
  • List only canonical pages that answer recurring questions cleanly
  • Use the same entity naming across root and subdomain files
  • Exclude staging, duplicate language variants with weak maintenance, and temporary campaign pages
  • Review whenever ownership of docs, help, or policy content changes

Notice what is missing here. No claim that the file boosts rankings. No claim that assistants will obey it consistently. No claim that adding it everywhere is automatically better. This is an organizational layer, not a performance hack.

The real leverage still comes from extractability. Put the answer near the top, keep qualifiers attached to claims, use crawlable HTML, and make one page the canonical home for each fact. If you have not done that work, llms.txt is lipstick on a content ops problem.

If you want a quick check on whether your pages are likely to be readable and quotable in the first place, use the AI Visibility Checker. If you need help fixing the underlying content architecture across site and docs, you can also book a call.

So should most teams publish llms.txt on subdomains?

Yes, if those subdomains are public, content rich, and genuinely act as source libraries. No, if they are private, thin, temporary, or operationally neglected.

My operator view is simple. Publish the file where it matches a maintained body of public source content. Do not publish it as theater. If your root domain is not where the facts live, root only is too neat to be useful.

The trade off is maintenance. Every extra host adds another file to keep aligned. If your team cannot maintain consistency between root, docs, and help, fewer files may be safer than more files with stale guidance.

Common questions

Should every subdomain have its own llms.txt?

No. Only public subdomains with meaningful source content should get one. Private apps, staging hosts, and thin campaign sites usually should not.

Can the main domain llms.txt cover docs and help subdomains?

You can mention those areas, but I would not rely on one root file to represent separate hosts. If the subdomain matters, publish a file on that subdomain too.

Will publishing llms.txt improve SEO or AI citations by itself?

Do not expect that. Google says llms.txt is not used by Search, and the cited adoption study found no citation lift after controls.

What matters more than llms.txt for AI visibility?

Crawlable HTML, clear answer first structure, stable canonical pages, and keeping qualifiers close to claims matter more in practice.

Who should ignore this advice?

Small sites with one host and little documentation can keep it simple. Teams that cannot maintain consistency across hosts should fix content governance before multiplying files.

Last updated: 2026-11-14

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