Quick answer: llms.txt is a plain-text file at yourdomain.com/llms.txt that hands an AI model a short, hand-picked index of your site instead of making it parse your nav bar, cookie banner and template JavaScript. Real estate agents do not need one to get recommended by AI — a seven-month study of roughly 900 sites found zero requests from the crawlers behind ChatGPT, Claude or Perplexity, and Google says directly that it has no effect on Search or AI Overviews. Ship it anyway: it costs about fifteen minutes and there is no real downside.
What llms.txt actually is
Developer and AI researcher Jeremy Howard proposed the format on September 3, 2024. It is deliberately thin: one heading with your site's name, a short summary underneath, then a list of links to your best pages with a line each describing what is there. The idea borrows from robots.txt, which tells crawlers what they may access, and sitemap.xml, which tells search engines what pages exist — except llms.txt is aimed at language models with small context windows, handing them a short index rather than a maze of navigation code.
A related file, llms-full.txt, dumps your entire site into one document. That variant is not part of Howard's original proposal; it is a convention the documentation platform Mintlify popularized in November 2024. The two do different jobs — llms.txt is the index, llms-full.txt is the full dump — and confusing them is a common mistake.
Does it get real estate agents recommended by AI?
No confirmed evidence of it yet. A server-log study that watched roughly 900 websites for seven months, September 2025 through April 2026, logged 1,227 total requests for llms.txt-family files, and zero came from GPTBot, ClaudeBot or PerplexityBot. The study's own conclusion was that among everyone requesting the file, there was not a single real AI bot.
A separate study of roughly 300,000 domains, published November 2025, tested whether having the file correlated with being cited more often in AI answers. It did not — a prediction model actually got more accurate once researchers removed llms.txt as a factor.
Google's own documentation says it plainly. Google Search Central, updated July 2026, states that you do not need llms.txt to appear in Google Search or its AI features, and that keeping one for other purposes is fine because it "will neither harm nor help your site's visibility or rankings in Google Search, as Google Search ignores them." Asked about it directly, Google's John Mueller put it more bluntly for a non-developer audience: "For non-developer sites, I don't think this makes much sense… Prioritize needs before dreams."
So where does it actually get used?
llms.txt does get used, just not by the audience real estate agents are trying to reach. Coding tools like Cursor, GitHub Copilot and Claude Code are widely reported to fetch llms.txt when pointed at a documentation site, because a developer mid-task is a genuinely good match for "here is a short map of this API."
Google even added an llms.txt check to Chrome's Lighthouse tool in May 2026 — filed under a new "agentic browsing" category next to a separate standard called WebMCP, nowhere near its SEO tools. That is Google's own product team saying the same thing as its search team: this is agent tooling, not a way to get found.
Should your site have one anyway?
Yes — just size the expectation correctly. It is roughly a fifteen-minute file and there is no real downside to shipping it. If you want it easy to find rather than guessed at, mirror it at /.well-known/llms.txt too, which is five more minutes.
Ours lives at leadjens.com/llms.txt: 42 pages indexed, organized by what LeadJens does, the case study, the FAQ and every blog post, each with a one-line description and a matching page.md mirror, so any AI tool that does fetch a page gets clean text instead of a wall of navigation code. I built it because it is good hygiene, not because I expect it to move a client's visibility score. Nobody should sell you an "llms.txt setup" as an AEO project — including me.
What actually moves the needle
The signals with real evidence behind them are the less exciting ones: a complete Google Business Profile, reviews published somewhere AI can actually read them, one consistent name, title and market across every listing, and a site marked up with schema so an AI system can verify who you are. Those are covered in what makes a site AI-ready, and they are the same gaps that show up in almost every visibility check I run — the five signals that actually move it are laid out in full here.
Ship the llms.txt file this week if you want; it costs you nothing. Then spend the real hour on the fixes that have data behind them. Get a free AI Visibility Score and I will tell you which gap is actually costing you.
FAQ
Do AI crawlers actually read llms.txt?
No evidence of it yet. A server-log study that watched roughly 900 websites from September 2025 through April 2026 logged 1,227 requests for llms.txt-family files, and not one came from GPTBot, ClaudeBot or PerplexityBot, the crawlers behind ChatGPT, Claude and Perplexity. The requests that did arrive came from other tools.
Does Google use llms.txt for AI Overviews?
No. Google Search Central, updated July 2026, states that llms.txt files are not needed to appear in Google Search or its AI features, and that keeping one will neither harm nor help a site's visibility or rankings because Google Search ignores them. AI Overviews draws from Google's regular search index.
Is llms.txt the same as robots.txt or a sitemap?
No. robots.txt tells crawlers what they are allowed to access and sitemap.xml lists what pages exist. llms.txt is a short, hand-picked index written for language models with limited context: your site name, a summary, and a list of your best pages with one line describing each. Only robots.txt and sitemap.xml are honored by search engines.
Should I pay someone to set up llms.txt for my real estate website?
No. It is roughly a fifteen-minute file with no measured effect on AI visibility, so it should not be sold as an AI visibility project or priced like one. Build it yourself if you want it, then spend the real budget on the signals with evidence behind them: a complete Google Business Profile, machine-readable reviews, one consistent name and title everywhere, and validated schema.
Sources: Jeremy Howard / Answer.AI, llms.txt proposal, September 3, 2024. Mintlify, llms-full.txt convention, November 2024. Server-log study of ~900 domains, September 4, 2025 – April 13, 2026. Citation-correlation study of ~300,000 domains, published November 20, 2025. Google Search Central, "Optimizing your website for generative AI features on Google Search," last updated July 10, 2026. John Mueller, Google Search Relations, 2026. Google Chrome Lighthouse llms.txt audit, May 5, 2026. Coding-agent adoption is reported in secondary coverage and is not vendor-confirmed. Page count for leadjens.com/llms.txt read from the live file, August 14, 2026.