Google confirmed on June 15, 2026 that llms.txt does nothing for Search rankings. Here's what the data shows, and how to ship a correct llms.txt file anyway.
Google does not read llms.txt. It said so plainly on June 15, 2026. If your only goal is higher Google rankings, you can close this tab. But if you're building for an AI-driven web — one where ChatGPT, Perplexity, and Claude increasingly influence what shoppers discover — keep reading, because the file still has a real, if narrow, purpose.
The concept was proposed in September 2024 by Jeremy Howard, co-founder of Answer.AI and a lecturer at the Universities of Queensland and Stanford. The problem he was solving was practical and had nothing to do with search rankings: large language models have limited context windows, and a full website — cluttered with navigation, cookie banners, JavaScript, and ad markup — is a terrible thing to feed one.
His solution: a single Markdown file at your domain root, /llms.txt, that gives an AI model a clean, human-readable overview of who you are and where your important content lives. Think of it less like robots.txt (which controls crawler access) and more like the table of contents at the front of a well-organized manual.
The original audience wasn't Google. It was developer tools — coding assistants like Cursor and Claude Code that need fast, reliable access to API documentation. When a developer loads the Anthropic or Stripe docs into an AI coding environment, having an llms.txt makes that process faster and more accurate. That's the use case it was actually built for.
Getting cited by AI search engines is a legitimate goal, but llms.txt is not the path there, at least not via Google. Here's the official timeline:
The June update also addressed people who want to keep a file anyway: maintaining one is fine, but it won't help or hurt your Google visibility. Google simply ignores it. This wasn't a policy change — Google never used it as a ranking signal. The update was retroactive context-setting to cut through hype that had outrun the facts.
One more structural point worth holding: AI Overviews and Google's AI Mode pull from the same index as regular Search. There is no separate "AI ranking" to optimize for with a special file.
The numbers aren't flattering for llms.txt advocates. Ahrefs tracked 137,210 domains with the file (a technically skewed sample, so an upper bound on adoption) and found that 97% saw zero requests in May 2026. Of the 3% that were fetched, only 19.5% of those requests came from named AI tools — GPTBot led, followed by Claude Code.
SE Ranking analyzed roughly 300,000 domains and found about 10% had published the file. When they ran a statistical model and an XGBoost classifier to check whether the file correlated with AI citations, the result was flat: no effect. Adoption rates were nearly identical across traffic tiers (low-traffic sites at 9.88%, mid-tier at 10.54%, high-traffic at 8.27%), which means high performers aren't doing something others are copying.
A 90-day server-log study by Limy.AI put the sheer scale in perspective: out of more than 500 million AI bot events, roughly 408 targeted llms.txt. That is not a typo.
As of Q1 2026, no major AI company — OpenAI, Google, Anthropic, Meta, or Mistral — has publicly committed to reading or acting on llms.txt in their production search or retrieval systems. The file is not an official standard. It has no backing from W3C or IETF, and no enforcement mechanism. AI providers do what they want with it.
Despite all of the above, there are a few legitimate reasons a Shopify seller or content operator might still want to publish one. They're narrow but real.
Shopify officially documented native llms.txt support in late May 2026. The developer docs now list templates/llms.txt.liquid and templates/llms-full.txt.liquid as supported Liquid templates, alongside a linked agents.md file. That companion file — agents.md — is effectively the operating manual for an AI tool that wants to buy something on your behalf.
Adobe Analytics reported that AI traffic to US retail sites rose 393% year-over-year in Q1 2026, and that traffic now converts 42% better than non-AI traffic. Whether llms.txt directly drives that is unproven, but the agentic buying layer is real and Shopify is building for it. Your file positions you in that architecture even if the current traffic volumes are negligible.
Publishing an llms.txt — and being thoughtful about what you include — is a low-cost way to signal which content you're comfortable with AI models referencing. You can include attribution preferences, usage guidelines, and contact information for commercial use. It won't be legally binding, but it establishes intent and creates a record of your position. That matters more as AI content law evolves.
If your store has a public API, developer documentation, or an affiliate program with technical onboarding materials, coding assistants actively benefit from a well-structured llms.txt. Claude Code and GPTBot both fetch the file (though rarely), and developer-tool AI assistants increasingly use llms-full.txt for accurate API guidance. If your audience includes developers, this is worth 30 minutes.
The format is deliberately simple — structured Markdown. Here's what the spec requires, in order:
[Page Name](URL): brief description.## Optional has special meaning — AI consumers may skip those entries when building a smaller context window.A few hard technical rules that trip people up:
text/plain, not HTML. Check the Content-Type header.To verify after publishing, visit the URL directly in your browser and confirm the Markdown content renders as plain text with an HTTP 200 response.
# Acme Supply Co.
> Acme Supply Co. sells professional-grade hand tools to independent contractors in the US, founded in 2019. Our catalog covers 400+ SKUs across plumbing, electrical, and carpentry.
## Products
- [Full Catalog](https://acmesupply.com/collections/all): Browse all 400+ tools by trade
- [Best Sellers](https://acmesupply.com/collections/best-sellers): Top 20 tools by order volume
## Key Pages
- [About](https://acmesupply.com/pages/about): Company background and sourcing standards
- [Shipping Policy](https://acmesupply.com/pages/shipping): Rates, timelines, and free shipping threshold
- [Returns](https://acmesupply.com/pages/returns): 60-day return window, no questions asked
## Blog
- [Tool Guides](https://acmesupply.com/blogs/guides): How-to articles on tool selection and use
## Optional
- [Press](https://acmesupply.com/pages/press): Media mentions and brand assets
Notice what's not in there: marketing language, vague superlatives, or every page on the site. The blockquote includes specifics — product count, customer type, founding year. That factual density is what makes the file useful to a language model rather than decorative.
Shopify does not let you place files at the domain root by default. There are three working methods:
templates/llms.txt.liquid. Add your template in the theme editor. The tricky part: your custom file completely replaces Shopify's default and does not merge. If you remove the agentic commerce discovery links to /.well-known/ucp and the MCP endpoint, you'll cut your store off from checkout orchestration Shopify has already wired in. Keep those links; change only your description content./llms.txt to that CDN URL. Quick, no dev required.Yoast SEO for Shopify added automatic llms.txt generation in March 2026. If you're already running Yoast, the setting is in the dashboard — toggling it on generates and updates the file weekly with zero developer involvement.
If you're spending time on AI visibility, the ROI ranking goes roughly like this, based on what actually influences AI citations today:
It's also worth remembering what zero-click search at 68% actually means: the queries that still drive traffic are the ones where users need something beyond the snippet. Building content that earns that click — through specificity, trust, and depth — is a more durable investment than any single file.
Ship it if you're on Shopify and want to be positioned for agentic commerce infrastructure. Ship it if you have developer-facing documentation or an affiliate program with technical onboarding. Use the blockquote to say something specific and true about your brand, and point to pages that can actually back that claim up.
Don't ship it expecting Google rankings to move. Don't ship it expecting a flood of AI-referred traffic next month. And definitely don't pay an agency to build an elaborate llms-full.txt architecture before your product pages have accurate specifications and your schema markup is clean.
The file costs about 30 minutes to do right. Do it right, then spend the remaining 30 minutes on something with a proven return.
No. Google's AI optimization guide, updated June 15, 2026, states directly that Google Search ignores llms.txt files. The file has no effect on standard rankings or Google's AI Overviews, which pull from the same index as regular Search.
As of May 2026, very few do in any meaningful volume. Ahrefs found 97% of files received zero requests. Of the 3% that were fetched, GPTBot led followed by Claude Code — primarily for developer documentation use cases, not general web retrieval.
The file must live at yourdomain.com/llms.txt — root level, no subdirectory. It should be served as text/plain with UTF-8 encoding and return an HTTP 200 status code. On Shopify, you need to use the native llms.txt.liquid template, a URL redirect from the Files section, or a Cloudflare Worker proxy.
Start with an H1 heading (your brand name), then a blockquote with a one-to-two sentence specific description of what you sell and who you sell to — include product count, customer type, and founding year. Follow with H2 sections linking to your most important pages: catalog, key policies, and blog, each with a brief description. Avoid marketing language; factual specificity is what AI models can actually use.
Only if you have extensive documentation or a large product catalog where deeper context is valuable for developer tools. For most Shopify sellers, a well-written llms.txt is sufficient. Creating llms-full.txt with indexable Markdown copies of every page can introduce duplicate content at scale, which risks diluting crawl budget.
Update it whenever your catalog, pricing, or key policies change significantly — at minimum quarterly. If you launch a new product category or run a major seasonal promotion, refresh the file within a week of that change.
The playbook
One email when we publish something worth reading. No drip sequence, no spam, unsubscribe any time.