Does LLMS.txt Do Anything for SEO? Google Says No, and the Data Agrees.
Does llms.txt help SEO? No. Google confirmed in June 2026 that llms.txt has no ranking effect. Here's what the data shows actually drives AI citations.
Does LLMS.txt Do Anything for SEO? Google Says No, and the Data Agrees.
Does llms.txt help SEO? No. In June 2026, Google confirmed through a Search Central help doc update that llms.txt files have no effect on rankings or AI citations. HTML remains the standard, and the data backs Google's position: domain authority, structured content, and brand presence drive AI visibility. A text file in your root directory does not.
If you have already added llms.txt to your site, you didn't break anything. You also didn't gain anything. Here's the full picture.
What llms.txt actually is
The llms.txt proposal originated in late 2024 as a Markdown-formatted file placed at the root of a domain (yoursite.com/llms.txt). The idea was to give large language models a clean, summarized map of a site's most important content, the way robots.txt gives search engines crawling instructions or sitemap.xml gives them a URL inventory.
A typical llms.txt file looks like this:
Clean. Readable. Reasonable on its face. The pitch was that AI systems could use this to understand a site faster than crawling and parsing HTML.
Why llms.txt gained traction
A few reasons it spread.
- It felt familiar. Robots.txt and sitemap.xml work. A similar-shaped file for AI sounded plausible.
- AI search anxiety. Marketers worried about losing traffic to AI summaries wanted any lever they could pull.
- Low cost, low risk. Adding a text file takes minutes. Even if it didn't help, it didn't seem to hurt.
- Vendor marketing. Several SEO tools added "llms.txt generators" and sold them as part of AI optimization packages.
By early 2026, a meaningful slice of WordPress sites and SaaS landing pages had an llms.txt. None of them benefited.
Why llms.txt doesn't work
Google addressed this directly in its June 2026 Search Central update on third-party SEO tools. The update named llms.txt as a tactic to ignore. Three reasons it doesn't work:
1. No major AI provider adopted it
OpenAI, Anthropic, Google, and Perplexity never committed to reading llms.txt. The proposal was put forward by a third party. Without provider adoption, the file is just a static text file sitting unread on a server.
2. AI systems already parse HTML
The systems retrieving content for ChatGPT, Gemini, and Perplexity crawl and parse standard HTML the same way Google does. They extract structured data from schema.org markup, pull entity information from clear page structure, and use authority signals to weight what they cite. They don't need a separate Markdown file.
3. Google explicitly disregards it
Google's June 2026 guidance was direct: llms.txt does not influence rankings, AI Overviews, or AI Mode citations. Google's crawlers do not give it any weight. The file is functionally invisible to Google.
What the data actually shows drives AI citations
While llms.txt was being debated, real research was tracking what makes AI systems cite a site. The most useful data point from 2026:
Sites with 32,000+ referring domains are 3.5x more likely to be cited by ChatGPT than sites with under 200 referring domains.
Translation: the same authority signal that drives Google rankings drives AI citations. Domain authority is the single biggest predictor. Not file structure. Not special markup. Not llms.txt.
A few other patterns from the data:
- Branded search volume correlates with AI citation rate. Sites with strong brand recognition get pulled into AI responses more often, even when they aren't the highest-ranking result for a query.
- Structured content (FAQ, HowTo, clear H2 sections) gets extracted more reliably. AI systems parse semantic HTML and lift quotable chunks.
- Topical depth matters more than individual page optimization. Sites with 50+ pages on a topic get cited as authorities on that topic.
- Mentions across the open web (press, podcasts, industry blogs) feed AI training data and retrieval. Brand presence outside your own domain compounds.
None of these are llms.txt. All of them are SEO fundamentals applied to the AI search context.
What to invest in instead
If you have an hour to spend on AI search visibility, none of it should go toward llms.txt. Spend it here.
Earn referring domains
Real links from real, relevant sites. Press mentions. Industry publications. Podcasts. Guest contributions. The link graph still rules AI citation rates.
Build entity clarity
Use Organization schema on your homepage. Use LocalBusiness schema if you serve a geographic area. List your services explicitly. Define your terms on dedicated pages.
Cover topics in depth
A 10-page topic cluster will outperform a single 2,000-word piece for AI citation. AI systems pattern-match on topical authority. One thin page rarely gets surfaced.
Lead with the answer
The first 100 words of every page should answer the question the page is targeting. AI systems extract that section disproportionately.
Standard schema, applied well
Article, FAQ, HowTo, Product, LocalBusiness, Organization. Use the schema types that match your content. There is no special "AI schema."
Crawlable HTML
Server-side rendering or static generation for content pages. Avoid client-side-only React or Vue for content that needs to be cited. If a bot can't render it, it can't cite you.
What about robots.txt and sitemap.xml? Are those next?
No. Robots.txt is a real standard that Google and major bots respect (it controls crawler access, not rankings). Sitemap.xml is a real signal Google uses for URL discovery. Both have decades of adoption.
Llms.txt is a proposed standard that no major provider has committed to. The distinction matters. Don't confuse "files in your root directory" with "files anyone reads."
Should I remove llms.txt if I already have one?
It's harmless to leave. It's also harmless to remove. The file has no effect either way. If you want a tidier root directory, delete it. If you want to keep the option open in case adoption changes, leave it. Don't spend more time generating, maintaining, or auditing it.
Frequently asked questions
Does llms.txt help with SEO rankings?
No. Google confirmed in its June 2026 Search Central guidance that llms.txt has no impact on rankings or AI search visibility. Google does not read the file as a ranking signal.
Does ChatGPT use llms.txt?
OpenAI has not announced support for llms.txt as a content discovery or citation signal. Like Google, ChatGPT's retrieval systems work from crawled HTML.
Does any AI tool actually read llms.txt?
No major AI provider (OpenAI, Anthropic, Google, Perplexity) has committed to reading llms.txt as a standard. A few small experimental tools have referenced it, but none of the systems driving meaningful AI search traffic use it.
What's the difference between llms.txt and robots.txt?
Robots.txt is a long-established standard that controls crawler access to a site. Major search engines and AI crawlers respect it. Llms.txt is a proposed format that no major provider has adopted. Robots.txt works; llms.txt doesn't.
What actually helps my site get cited by AI?
Domain authority (measured by referring domains), brand presence across the open web, topical depth, standard schema markup, and clear answers near the top of each page. The same signals that drive Google rankings drive AI citations.
Should I remove llms.txt from my site?
It doesn't matter. The file has no effect, positive or negative. Remove it if you want a cleaner root directory. Leave it if you don't.
Bottom line
Does llms.txt help SEO? No. Google said so in June 2026, and the citation data confirms it. AI systems cite sites with strong authority, clear structure, and topical depth. They don't cite sites with a Markdown file in the root directory. If you're allocating time toward AI search, spend it on link building, content depth, and schema. Skip llms.txt.
*Want help building an SEO foundation that earns AI citations the real way? Book a call with us.*
Related articles
- Google Reviews Are Disappearing in July 2026: What's Happening and How to Recover Legitimate Reviews
- Google's June 2026 Spam Update: What It Targets and How to Respond
- Google Search Console AI Performance Reports: How to Read Your AI Overview and AI Mode Impressions
- Why 80% of ChatGPT Product Recommendations Change When Web Search Is Enabled
- The May 2026 Google Core Update: What Changed and What to Do If You Got Hit
- Google Home Listing Ads Explained: How the LSA Real Estate Format Works in All 50 States
- All Kanopy resources
- SEO & web services
- Paid ads services
- AI outreach services