What is llms.txt?

llms.txt is a proposed standard file — similar to robots.txt and sitemap.xml — that provides AI models with structured information about your website, brand, products, and services. When an LLM or AI agent visits your site, it reads llms.txt to understand who you are, what you offer, and what content is most important.

"If robots.txt tells crawlers what NOT to access, llms.txt tells AI what it SHOULD know about you." — GEOWorkbook

Why llms.txt Matters for GEO

The connection is direct:

AI engines increasingly visit websites during RAG (Retrieval-Augmented Generation)

Without llms.txt, AI must parse your entire site to understand your brand

With llms.txt, you provide a structured, authoritative summary of your brand identity

This directly influences how AI represents you in responses

Impact Data

• Sites with llms.txt report 23% higher citation rates in Perplexity
• 67% of enterprise sites are expected to have llms.txt by end of 2026
• ChatGPT's Browse mode and Perplexity both check for llms.txt during research

llms.txt Format & Structure

Basic llms.txt Structure

# Company Name

About
A concise description of your company, what you do, and your core value proposition. 2-3 sentences max.

Products & Services
Product A: Brief description with key differentiators
Product B: Brief description with key differentiators
Service C: Brief description of the service

Key Facts
Founded: Year
Headquarters: City, Country
Industry: Your industry
Specialization: Your niche expertise

Important URLs
Homepage: https://yoursite.com
Blog: https://yoursite.com/blog
Documentation: https://yoursite.com/docs
Contact: https://yoursite.com/contact

Social Proof
Trusted by X+ customers
Y industry awards
Featured in [Publications]

Advanced: llms-full.txt

While llms.txt is a summary (recommended: 500-1000 words), llms-full.txt is the comprehensive version (2000-10000 words) containing detailed information about:

• Complete product/service descriptions
• Detailed team credentials (E-E-A-T signals)
• Technical specifications
• Case studies and results
• Full FAQ content
• Pricing overview
• Integration details

Step-by-Step Implementation

Step 1: Audit Your Current AI Presence

Before writing llms.txt, understand how AI currently describes you:

Ask ChatGPT: "What is [Your Company]?"

Ask Perplexity: "Tell me about [Your Company] and their services"

Ask Gemini: "What does [Your Company] do?"

Document gaps and inaccuracies

Step 2: Write Your llms.txt

Key principles:

• Lead with your value proposition — the first paragraph matters most
• Use clear, factual language — no marketing fluff
• Include specific numbers — revenue, customers, years in business
• List your key differentiators — what makes you unique
• Reference authoritative third-party mentions — press, awards, certifications

Step 3: Write llms-full.txt

Expand on llms.txt with:

• Full product descriptions
• Team bios with credentials
• Detailed case studies
• Technical documentation
• Complete FAQ

Step 4: Deploy Both Files

Place both files at your site root:

• https://yoursite.com/llms.txt
• https://yoursite.com/llms-full.txt

Step 5: Reference in robots.txt

# robots.txt
User-agent: *
Allow: /

AI-specific files
Allow: /llms.txt
Allow: /llms-full.txt

Sitemap: https://yoursite.com/sitemap.xml

Also Consider: cats.txt

cats.txt (Content Attribution and Trust Signals) is a complementary file that provides:

• Content authorship information
• Trust signals and certifications
• Content licensing details
• Source attribution policies

Monitoring & Optimization

Track AI Agent Visits

Check your server logs for:

• ChatGPT's user agent: ChatGPT-User
• Perplexity's user agent: PerplexityBot
• Google's AI crawler: Google-Extended
• Anthropic's crawler: ClaudeBot

Iterate Based on Results

Re-test AI responses monthly

Update llms.txt when products/services change

Add new achievements, press mentions, case studies

Monitor competitor llms.txt for best practices

Common Mistakes to Avoid

Too long — llms.txt should be concise (under 1000 words)

Marketing language — AI prefers factual, clear language

Missing credentials — Always include E-E-A-T signals

Outdated information — Update quarterly at minimum

Forgetting llms-full.txt — The detailed version is where depth matters

Not blocking bad bots — Use robots.txt to block unwanted AI crawlers while allowing desired ones

GEO Audit Connection

When running a GEO Audit, llms.txt analysis is a critical component:

• Does the site have llms.txt? (Existence check)
• Is it properly formatted? (Structure check)
• Does it match the actual brand? (Accuracy check)
• Is llms-full.txt present? (Depth check)
• Are AI bot access controls configured? (Security check)

---

Related: Schema Markup for GEO

GEO Audit Guide

The Agentic Web