There is a file called llms.txt that has been circulating in SEO and GEO circles since late 2024. You may have seen it described as “the new robots.txt” or “the file that tells AI engines what to cite.” The reality is more nuanced, and for treatment centers and mental health clinics, the nuance matters.
Here is what llms.txt actually is, what the data says about whether it works, and why Altitude deploys it for every client anyway.
What llms.txt is
llms.txt is a plain text file placed at the root of a website that helps AI systems understand the most important content on a site. Unlike traditional HTML pages filled with menus, JavaScript, tracking scripts, navigation, and styling noise, llms.txt provides clean structured signals specifically for AI systems. The concept was proposed by Jeremy Howard of Answer.AI in 2024 and gained major traction during 2025 and 2026.
Think of it like a table of contents written specifically for AI crawlers. Where robots.txt tells search engine bots which pages they can and cannot access, llms.txt tells AI systems which pages matter most, what the site is about, and how the content is organized. It uses Markdown formatting so AI crawlers can parse it cleanly without wading through the visual and structural noise that surrounds content on a normal web page.
A basic llms.txt for a treatment center might look like this:
# Example Recovery Center
> Example Recovery Center is a CARF-accredited addiction treatment facility
> in Phoenix, Arizona. We provide detox, residential, PHP, IOP, and
> outpatient services for adults with substance use disorders and
> co-occurring mental health conditions.
## Core services
- [Detox Program](/services/detox): Medical detox with 24/7 clinical supervision
- [Residential Treatment](/services/residential): 30-90 day inpatient programs
- [PHP](/services/php): Partial hospitalization for step-down or direct admit
- [IOP](/services/iop): Intensive outpatient, 3x weekly, evening and morning tracks
## Clinical resources
- [Insurance Verification](/insurance): What we accept and how to verify
- [Admissions Process](/admissions): How to get started, what to expect
## About
- [Our Team](/about/team): Licensed clinicians, certifications, and credentials
- [Accreditations](/about/accreditations): CARF, state licensing, and complianceThat is the entire file. It sits at yourdomain.com/llms.txt and is publicly accessible to any crawler that looks for it.
What the data actually says
Here is where honesty matters.
A SE Ranking study of 300,000 domains found a 10.13% adoption rate as of late 2025. After eighteen months of industry conversation, llms.txt is on roughly one in ten sites. Crawler interest from the major AI search and answer bots is negligible: across 500 million AI bot visits analyzed across a 90-day window, fewer than 500 targeted llms.txt directly.
As of May 2026, the major AI companies, including OpenAI, Google, Anthropic, and Meta, have not publicly committed to reading or acting on llms.txt in their production search or answer systems. GPTBot occasionally fetches the file. That is not the same as confirmation that the file influences how ChatGPT sources, ranks, or cites content.
Google's John Mueller has explicitly stated that no Google Search system reads or acts on llms.txt. The strongest confirmed real-world use case is developer tooling: AI coding assistants like Cursor, GitHub Copilot, and Claude retrieve documentation in real time, and llms.txt helps them fetch the right pages with less wasted processing.
So why does Altitude deploy it for every client?
Why it still belongs in your GEO stack
Three reasons, and none of them depend on the hype.
1. Perplexity and Anthropic have confirmed support
Anthropic has publicly confirmed that Claude Desktop and Claude.ai both respect llms.txt directives in retrieval workflows. Perplexity has publicly confirmed that it retrieves llms.txt and uses it to prioritize page selection. These are two of the fastest-growing AI search surfaces. Perplexity in particular is where a meaningful share of health and wellness research is happening right now, especially among younger adults who are often the ones researching options for a family member.
2. The B2A opportunity is more important than the B2C one
The companies that understand llms.txt are not treating it as an SEO play. They are treating it as a Business-to-Agent play: the first standardized way for a brand to publish a machine-readable surface that AI agents can route on. AI agents, meaning automated systems that research, compare, and summarize options on behalf of a user, are growing rapidly. A family member who asks an AI assistant to “find me PHP programs in Denver that take Blue Cross” is triggering an agent-driven query. llms.txt is how your facility tells that agent what you offer, where, and how to learn more, without the agent having to parse your entire site.
3. Early adoption in a slow-moving sector is an advantage
In healthcare and behavioral health, adoption rates among top-100 domains remain under 10 percent. That gap is an opportunity for mid-market brands to move before their sector as a whole. Treatment centers that implement llms.txt now are positioning ahead of a standard that is still forming. The file takes under 30 minutes to create. The cost of implementing it is essentially zero. The cost of ignoring it as AI agent usage scales is harder to predict.
What llms.txt cannot do
This is the part most guides skip.
llms.txt cannot block AI systems from crawling your content. It is a navigation file, not a blocking tool. It helps AI tools find your best content. It cannot restrict any crawler or prevent any AI system from reading your site. If you want to prevent specific AI bots from accessing your site, that is a robots.txt task, where you explicitly block user agents like GPTBot, ClaudeBot, or PerplexityBot.
llms.txt also does not directly improve your Google rankings. Google has confirmed they do not use it. It will not move your traditional SEO metrics.
And creating individual Markdown copies of every page on your site, a popular but misguided implementation approach, can actually hurt you. If those Markdown files are indexable, they introduce duplicate content at scale. Duplicate content dilutes crawl budget and can suppress rankings for the original pages. A single, well-written llms.txt at your root domain is what you need. Not a parallel Markdown version of your entire site.
What a treatment center's llms.txt should include
For behavioral health providers specifically, the file should cover:
Site identity. Your facility name, location, accreditations, and a one-paragraph description written in plain language. This is what AI systems use when they attribute content to your organization. Vague descriptions produce vague attributions.
Service structure. A clear, linked list of every level of care you offer, named consistently with how you name them on your site and in your SAMHSA listing. If your site calls it “Partial Hospitalization Program” and your llms.txt calls it “PHP,” you are creating an entity mismatch that confuses AI crawlers.
Clinical resources. Your insurance verification page, admissions process, and any content that answers the questions families actually ask. These are the pages most likely to be cited in an AI-generated answer about finding treatment.
Team and credentials. Links to your clinical staff bios. AI systems that apply YMYL standards to health content look for expertise signals. Making it easy for an AI crawler to find your licensed clinicians is part of building that signal.
How Altitude handles this
Altitude automatically generates and deploys llms.txt for every client site. The file is structured around the facility's actual service lines, clinical staff, and accreditations, not a generic template. It is updated when content changes, so the AI-facing map of your site stays accurate as you add or modify pages.
More importantly, Altitude treats llms.txt as one piece of a broader entity signal strategy, not a standalone fix. It works alongside FAQPage schema, author attribution, authoritative source citation, and content freshness to build the kind of AI-readable presence that earns citations across Perplexity, ChatGPT, Claude, Gemini, and Google's AI Mode.
The file is not magic. But it is part of what separates a facility that is positioned for AI search from one that is hoping the old playbook still works.
Frequently asked questions
Does llms.txt help with Google rankings?
No. Google has confirmed it does not use llms.txt. It does not affect your traditional SEO. Its value is in AI search surfaces, specifically Perplexity and Claude-powered retrieval, and in AI agent-driven queries where an automated system is researching options on behalf of a user.
Will implementing llms.txt hurt my site?
Only if you create individual Markdown pages for every URL on your site and leave them indexable. A single llms.txt file at your root domain is safe, and the risk of implementing it is effectively zero.
How long does it take to implement?
For a well-structured treatment center site, a quality llms.txt file takes under 30 minutes to write and publish. Altitude handles this automatically for all clients.
Which AI systems actually use it?
Perplexity and Anthropic's Claude have publicly confirmed they use it in retrieval workflows. OpenAI has not officially confirmed support but GPTBot has been observed fetching the file. Google does not use it. As AI agent usage scales, the number of systems that rely on it is expected to grow.
Is this the most important GEO change I can make?
No. FAQPage schema, named clinical author attribution, authoritative source citation, and content freshness have more documented impact on AI citation rates. llms.txt is a supporting signal in a broader strategy, not a standalone solution. Altitude tracks and scores all of these together.
What is the difference between llms.txt and llms-full.txt?
llms.txt is a curated index of your most important pages with short descriptions. llms-full.txt is a companion file that contains the full text of your site's content in one Markdown document for deep AI ingestion. Most sites only need llms.txt. Documentation-heavy platforms and larger facilities with extensive clinical content may benefit from maintaining both.
For the full picture on AI search visibility for behavioral health, read our previous post on why GEO matters for behavioral health.
See your facility's AI visibility score in 60 seconds.
Altitude is the SEO and GEO platform built specifically for treatment centers and mental health clinics by The Purpose Pilot. Run a free audit and see where you stand across ChatGPT, Perplexity, Claude, Gemini, and Google's AI Mode.
Run a free audit