llms.txt is a proposed standard - a Markdown file at your site root that lists your most important pages in clean, LLM-friendly form. It was introduced by Jeremy Howard (Answer.AI) in September 2024. As of 2026 no major AI engine has confirmed it uses llms.txt for retrieval, and a 300,000-domain study found no measurable citation lift. It's low-cost to ship, but manage expectations.
A plain Markdown file at yourdomain.com/llms.txt that gives models a curated, noise-free map of your content. The premise: modern web pages are full of nav, ads, cookie banners, and JavaScript that waste an LLM's limited context window. llms.txt offers a clean index instead. Spec lives at llmstxt.org.
It's conceptually like robots.txt (for access) or sitemap.xml (for discovery) - but for LLM comprehension. For a plain-language definition, see our llms.txt overview and the glossary.
This is where most guides oversell. The verifiable facts as of 2026:
docs.claude.com.Verdict: llms.txt is not a proven citation lever today. It may help AI agents and coding tools that explicitly fetch it (common in dev-tool docs), and it costs almost nothing to publish - but it is not a substitute for crawlability, structure, and freshness.
llms.txt is structured Markdown:
> One-sentence description of what you do.
## Docs
- [Getting Started](https://site.com/docs/start): Set up in 5 minutes.
- [API Reference](https://site.com/docs/api): Full endpoint list.
## Guides
- [Pricing](https://site.com/pricing): Plans and costs.
## Optional
- [Changelog](https://site.com/changelog): Release history.Rules:
## Optional section marks lower-priority links.llms-full.txt with full page contents inline./llms.txt at the site root (and optionally /llms-full.txt).If you'd rather not hand-write it, the llms.txt generator scaffolds a valid file from your existing pages.
Spend your effort on the proven levers: allow AI search crawlers, get indexed in Bing, lead with answer-first passages, add statistics and schema, and refresh regularly. Ship llms.txt as a cheap bonus - not a strategy. See our AEO guide for the levers that actually move AI citations.
Not sure which levers your site is missing? Run a free audit of up to 50 pages - SEO AEO Specialist by MrChief audits the proven AEO levers and flags where llms.txt fits (or doesn't) for your site.
No. Google representatives (Mueller, Illyes) stated in 2025 that Google does not use or support llms.txt and has no plans to. Mueller compared it to the discontinued meta keywords tag, and Illyes confirmed the position at Search Central Live in July 2025.
Neither OpenAI nor Perplexity has committed to automatically reading it for retrieval. There's no confirmed evidence it drives citations on those engines today. It may still help coding or agentic tools that explicitly fetch the file, but that is a narrower use case.
A ~300,000-domain study (SE Ranking, Nov 2025) found no measurable citation improvement. It may help agents that explicitly fetch it, but general citation lift is unproven. Treat it as a low-cost experiment rather than a reliable visibility lever for now.
It's cheap and harmless, and useful if your audience uses coding/agentic tools that fetch it. Just don't expect it to move ChatGPT/Perplexity citations - prioritize crawlability, structure, and freshness first, then ship llms.txt as a bonus.
robots.txt controls crawler access; llms.txt is a proposed comprehension aid listing your key content in clean Markdown. robots.txt is universally honored; llms.txt is not yet used by major engines, so the two solve very different problems.