LLMs.txt: What It Is, and Why No Engine Reads It Yet
LLMs.txt is a proposed convention: a Markdown file at the root of your domain that summarises your site for AI agents and coding assistants. It is not a crawler-access control — that is robots.txt — and it is not an AI-visibility lever. Google's AI guidance (2026-07-10) is direct: "Google Search ignores them." No other engine — OpenAI, Anthropic, Perplexity, Microsoft, Meta — documents parsing it either. The same Google paragraph says maintaining one for other systems is "completely fine." So: agent-facing documentation, nothing more.
What LLMs.txt Does
An llms.txt file sits at the root of your domain (example.com/llms.txt) and describes your site in Markdown for any agent or developer tool that chooses to fetch it. No search engine or AI assistant documents reading it. A typical file includes:
- Brand identity — Official name, description, and positioning.
- Key facts — Pricing, features, founding date, team size.
- Preferred citations — How you want to be described and quoted.
- Corrections — Common misconceptions and the truth.
How to Create Your LLMs.txt
LLMs.txt uses a simple markdown-like format:
Basic Structure
- Title line: # Your Brand Name
- Description: > One-line description of what your brand does
- Sections: Organized with ## headings for Key Facts, Products, Pricing, Contact
- Links: References to key pages for more information
Best Practices
- Keep it concise — Under 500 words. AI models process shorter, structured content more reliably.
- Use specific claims — "Serves 10,000+ users" not "serves many users."
- Include pricing — Reduces the #1 hallucination category.
- List your differentiators — What makes you different from competitors.
- Add a corrections section — Proactively address known hallucinations.
Why We Still Publish One
Not for visibility. Developers pasting a site into a coding assistant, and agent frameworks that fetch a root Markdown file, can use it. That is a real but narrow audience. There is no evidence — zero controlled tests across a 45-study survey — that it changes what ChatGPT, Perplexity, Claude or Google say about a brand.
If you publish one, keep it in parity with your visible pages. A file that says $49 while your pricing page says $59 is one more contradiction.
Frequently Asked Questions
Do any AI platforms support LLMs.txt?
None documents it. Google states Search ignores it, and OpenAI, Anthropic, Perplexity, Microsoft and Meta publish nothing about consuming it. It does no harm. It also does no measured good for AI search visibility.
How is LLMs.txt different from robots.txt?
Robots.txt controls which crawlers may fetch which paths, per user agent — that is the file that decides whether OAI-SearchBot or PerplexityBot can index you. LLMs.txt controls nothing; it is a description. If you have time for one of them, fix robots.txt.
Does 99Visibility have an LLMs.txt?
Yes. Visit 99visibility.com/llms.txt to see our implementation.
Check your AI visibility and see if AI platforms are describing your brand accurately — with or without LLMs.txt.
Ready to see what AI says about your brand?
Run your first audit free. Get visibility scores, detect hallucinations, and get specific fixes.
Start Free AuditRelated Articles
What Is GEO? The Complete Guide for 2026
Generative Engine Optimization (GEO) is how you get your brand into AI-generated answers. This guide covers everything: what GEO is, why it matters, the 7 ranking factors, and how to start.
Schema Markup for AI Visibility: A Practical Guide
Schema markup is not an AI ranking lever — Google says so and the one controlled test found a small negative. Here is what it is still good for, and the one check that matters: every fact in your schema must also be in your visible text.
The GEO Checklist: 15 Fixes You Can Do Today
A practical, prioritized checklist of 15 things you can fix right now to improve how AI platforms describe your brand. Each item includes effort estimate and expected impact.
