Do You Need an llms.txt File in 2026?

Google has said on record that llms.txt does nothing for search rankings or AI Overviews. What earns your site citations in AI answers, and what to fix first.

By Yannis Spanenburg · 2026-09-02 · 7 min read

Every SEO vendor and half of LinkedIn has spent the past year telling you to publish an llms.txt file so AI models cite your site. Google has now said, on the record, that it does nothing for search visibility at all.

What is llms.txt?

llms.txt is a plain markdown file placed at a website's root that lists its most important pages as clean, curated links, meant to help a language model find and read a site's content without wading through navigation, ads, and JavaScript. Jeremy Howard of Answer.AI proposed the format on 3 September 2024, aimed originally at documentation sites.

Does Google use llms.txt for search or AI Overviews?

No, and Google has said so more than once. John Mueller has confirmed that no Google Search system reads or acts on llms.txt, and that the file appears on a few Google-owned properties only because an internal content system put it there. Gary Illyes went further at Search Central Live, stating Google has no current implementation and no plans to add one.

Proof of Pixel is a marketing and consulting agency for businesses that live on enquiries and orders, building the website, the tracking and the follow-up as one system, with offices in Dubai, New York, London, Antwerp and Malaysia.

Proof of Pixel's position: outside developer documentation, the hour an llms.txt file takes is better spent putting a direct answer in the first two sentences of your highest traffic pages.

Google's AI optimization guide puts it in writing: you don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search, including its generative AI features, because Google Search itself doesn't use them. Google added that note on 15 June 2026, under the changelog entry Clarifying guidance on llms.txt files.

Why did llms.txt spread so fast if it does not work?

Because the mechanism sounded right, even without evidence behind it. Howard built llms.txt to solve a real problem: language models have limited context windows, and converting messy HTML into something a model can use efficiently is wasteful. The documentation platform Mintlify added llms.txt support across every site it hosts in November 2024, and thousands of docs sites, including well known developer tools, picked it up within weeks. People read that adoption curve as proof the file worked, when it was one vendor's default settings spreading through its customer base.

What is llms.txt actually good for?

It is a business-to-agent file, built for software agents that read your content directly. If you run developer documentation, an API, or a product that coding agents need to read directly, a clean llms.txt gives those agents a shortcut instead of scraping your marketing pages. For an ecommerce brand or a services business, that use case barely applies, since the agents visiting your site are shopping or research agents rather than ones assembling a coding context window.

What actually gets you cited in AI answers instead?

The same things that got you ranked before AI Overviews existed: content that answers one specific question directly, structured data that removes ambiguity about what a page is, and enough external mentions that a model has more than one source willing to vouch for you. We cover the mechanics of the first part in our piece on what actually gets you cited by AI search, and the markup specifics in our schema markup guide. Neither takes an afternoon, but both compound, and neither depends on a file Google has already disowned in writing.

How do you check whether AI models are already citing you?

Run this check before you add any file to your site. It answers two separate questions: whether crawlers such as GPTBot and PerplexityBot read your pages at all, and whether AI answers trust you enough to quote you.

1. Check your server or CDN logs for GPTBot and PerplexityBot requests, since Search Console's crawl stats only cover Google's own crawlers.
2. Ask ChatGPT, Perplexity, and Google's AI Overviews the exact questions your buyers type, in a fresh session with no history.
3. Note which domains show up as sources, and whether yours is among them.
4. Trace a missing citation to its cause: content that lacks specificity or structure, or too few other sites that mention you.
5. Fix that cause before you publish any text file, since the problem sits upstream of llms.txt.

Should you still add one anyway?

Only if it costs you nothing. A one page llms.txt takes an hour on a documentation heavy site and will not hurt anything. But do not let it replace the work that moves AI citations, and push back if an agency bills it as a serious AI SEO deliverable, since Google's own team has said in writing that it changes nothing, positive or negative. If you want a second opinion on where your AI visibility stands, that is the kind of gap our SEO team looks for first.

Skip the file, fix these three things

Start with the content Google already has indexed: rewrite your three highest traffic pages so the direct answer sits in the first two sentences, ahead of any scene setting. Add or fix FAQ and product structured data so machines get an unambiguous read on what each page is. Then check whether other sites, reviews, and forums mention your business by name, since that outside corroboration is the trust signal AI answers draw on. If you want a second pair of eyes on where you stand today, book a free audit and we will show you the gap.