llms.txt is described as the sitemap for AI — a simple file you place at the root of your website to tell AI crawlers what is worth reading and where to find it. The pitch is compelling: in a world where LLMs decide which brands get cited, having a file that routes them to your best content should matter.
The reality is more complicated. Only 5 to 15% of websites have implemented llms.txt, and the bots that actually drive AI search citations barely touch it.
What llms.txt is
llms.txt is a plain-text file at yourdomain.com/llms.txt that documents your site's content surfaces in a format designed for AI retrieval systems. Think of it like sitemap.xml — but instead of listing every URL for a search crawler, it surfaces your key content areas and documentation for an AI agent trying to understand what you do. The format is Markdown-based and human-readable.
What the data says about llms.txt brand visibility impact
Across analysis of over 500 million LLM bot traffic events, the share of requests from the bots that drive actual AI search citations — GPTBot, ClaudeBot, PerplexityBot — that touched /llms.txt was statistically negligible.
Google's Gary Illyes confirmed that Google does not support llms.txt and has no plans to. No major LLM provider — OpenAI, Anthropic, Google, Meta, or Mistral — has publicly committed to using it as a signal in their production search or answer surfaces.
Where llms.txt does work
The real use case is IDE agents and developer tools, not consumer AI search. Cursor, Claude Code, GitHub Copilot, and Windsurf all fetch llms.txt when helping developers work with a codebase or API. This is why OpenAI, Anthropic, Stripe, Cloudflare, Mastercard, and Vercel have all adopted it — they have developer audiences for whom this routing layer is genuinely useful.
Who should implement it
Implement llms.txt now if:
- You have a developer audience or a public API
- Your product is represented in documentation that AI coding assistants might need to parse
- You want to be positioned for the emerging MCP/agent ecosystem
Lower priority if:
- Your audience is primarily consumers or non-technical buyers
- Your main goal is to appear more in ChatGPT or Perplexity answers
- You are choosing between llms.txt and fixing your schema markup — fix the schema first
Key takeaways
- llms.txt is a plain-text routing file for AI agents — not a signal for AI search citation ranking.
- Only 5–15% of sites have it; the bots driving AI search citations barely read it.
- Google doesn't support it; no major LLM provider has committed to using it for search surfaces.
- Where it works: IDE agents (Cursor, Claude Code, Copilot), MCP servers, developer tooling.
- High priority if you have a developer audience or public API; lower priority for consumer brands.
- If resources are limited, fix schema markup and content freshness first.