Emerging standards

What is llms.txt?

Shopify just shipped it to every store. Anthropic, Cloudflare, and Stripe have had it for months. 82% of Australian businesses still don't know it exists. Here's what it is, where it came from, what the data says, and whether you should care.

82%

of Australian businesses have no llms.txt file, based on StudioHawk's audit of 83 sites across retail, finance, travel, and hospitality.

What it is

llms.txt is a plain-text file that lives at the root of your website — yourdomain.com/llms.txt. It gives AI language models a clean, curated summary of what your site is about, what pages matter, and what content you want them to understand.

Think of it as the AI equivalent of robots.txt — but instead of telling crawlers where not to go, it tells AI systems where to start.

An optional companion, /llms-full.txt, contains the complete content of your site in clean markdown — a full briefing document for any AI that encounters your brand.

Where it came from

Jeremy Howard — co-founder of fast.ai and the researcher who popularised transfer learning — proposed the standard on 3 September 2024. The spec is published at llmstxt.org.

The proposal recognised a structural problem: websites are built for humans and Google. The HTML is cluttered with navigation, footers, cookie banners, and JavaScript. When an AI system tries to understand your business, it has to wade through all of that noise. llms.txt removes the noise at the source.

It remains a community convention — no IETF RFC has been filed. But convention is becoming momentum.

Who has implemented it

The list of adopters signals where this is heading:

Anthropic — publishes a concise /llms.txt plus a full /llms-full.txt dump of their documentation
Cloudflare — ships llms.txt across all developer documentation
Stripe, Vercel, Cursor, Supabase — all confirmed implementations
Shopify — auto-enabled for every store (more on this below)

SE Ranking's analysis of 300,000 domains found 10.13% had an llms.txt file as of 2025. In the world of web standards, that's fast adoption.

What Shopify just did

In May 2026, Shopify quietly auto-enabled three agent-facing endpoints on every store with no announcement, no email, no admin banner. Every Shopify storefront now natively serves:

/llms.txt — now 301-redirects to /agents.md
/agents.md — Shopify's richer format, pointing AI agents to the store's search API, catalogue, sitemap, and checkout
/.well-known/ucp — Universal Commerce Protocol, enabling AI agents to add items to carts and initiate checkouts autonomously

Shopify didn't just adopt the standard — they extended it toward a full agentic commerce protocol. AI-referred traffic to Shopify stores grew 8x year-on-year by Q1 2026, with AI-attributed orders up 13x. That growth is the context for the infrastructure decision.

When the platform that powers 4.8 million stores ships this by default, it shifts from "emerging" to "expected".

What the data says about citation impact

Here's where intellectual honesty matters. The evidence on whether llms.txt directly improves AI citations is thin:

Perplexity has confirmed they reference it
OpenAI, Google, Anthropic have not made explicit public commitments to use it for citations
SE Ranking found no statistical correlation between having an llms.txt and being cited by AI systems in their 2025 analysis
Server log analysis shows major AI crawlers do not consistently request the file during routine crawls
Presenc AI research found modest citation uplift specifically on Anthropic's Claude.ai and Perplexity

The honest take: llms.txt is not a citation shortcut. It's infrastructure positioning. The direct uplift today is modest and unproven. The signal it sends — that your site is built for the agentic web — is real, but the mechanism isn't a magic ranking factor.

What we found across 83 Australian businesses

As part of StudioHawk's AI Visibility Audit benchmark — auditing Australia's most-visited business websites — we tracked llms.txt adoption:

82%

have no llms.txt

93%

have no dedicated LLM reference page

This includes brands like Medibank, RACV, iSelect, Telstra, and major AU retailers. The gap isn't because they've evaluated it and passed — most simply haven't got there yet.

We score these signals in our benchmark data but do not include them in individual audit scores. The standard is moving fast enough that we track it — but not yet proven enough that we penalise businesses for missing it.

Our take

Implement it. Not because it's definitely working right now, but because:

It costs an afternoon to build and nothing to maintain
You'd be ahead of 82% of your Australian competitors immediately
Shopify shipping it by default means the expectation is being set at the platform level
When the AI crawlers do start reading it consistently — and they will — you'll already be indexed correctly

The businesses that win in AI search aren't the ones who waited for certainty. They're the ones who were already configured correctly when the behaviour changed.

How to implement it

A minimal llms.txt for a business website:

# [Your Business Name]

> [One sentence describing what you do and who you serve.]

## Key pages

- [Homepage](https://yourdomain.com/): Overview of services
- [Services](https://yourdomain.com/services/): Full service listing
- [About](https://yourdomain.com/about/): Company background
- [Contact](https://yourdomain.com/contact/): Get in touch

## What we do

[2-3 sentences. Plain language. No marketing fluff.]

Upload it to yourdomain.com/llms.txt and serve it as text/plain. That's it.

For a deeper implementation — including /llms-full.txt with full site content and proper schema markup — StudioHawk includes llms.txt setup as part of our AI visibility engagements.

See how your site scores on AI visibility

Free audit. No signup. Results in under 90 seconds.

Audit my site

Built by StudioHawk — Australia's largest dedicated SEO agency. Data from the StudioHawk AI Visibility Audit benchmark, May 2026.