llms.txt is one of the simplest, most overlooked opportunities in AI visibility optimization. A small text file at your domain root, it gives AI language models a structured, efficient summary of your website — and it takes under an hour to implement correctly.
This guide covers everything: what llms.txt is, the full specification, what to include, how to implement it, and how to use it alongside other AI visibility signals.
What Is llms.txt?
llms.txt is a plain-text file placed at yourdomain.com/llms.txt that provides AI language models with a structured, human-readable summary of your website. It was originally proposed by Jeremy Howard (fast.ai) in 2024 and has since been adopted by thousands of websites.
Unlike robots.txt (which controls crawler access permissions) or sitemap.xml (which provides a machine-readable URL index), llms.txt is designed for LLM comprehension — helping AI systems understand what your website is about, who it serves, and where to find key content.
The analogy: robots.txt tells crawlers what to read, sitemap.xml tells them where to find things, and llms.txt tells AI what everything means.
Why llms.txt Matters
AI Crawlers Process Websites Inefficiently
Modern AI crawlers visit many pages but struggle to efficiently understand the architecture and purpose of complex websites. They may crawl your pricing page, your about page, and several blog posts — but without context, they can't efficiently determine which content is most important to understand your business.
llms.txt provides that context directly, allowing the AI to prioritize your most important content and understand your business quickly.
It Positions You as AI-Literate
A well-crafted llms.txt signals to AI platforms and users that you understand and actively optimize for AI search. This forward-looking positioning matters as the AI ecosystem evolves.
It's Increasingly Common
As of 2026, thousands of websites have implemented llms.txt, including major technology companies, developer tools, and forward-thinking businesses across industries. Early movers benefit from being part of the adopter cohort that shapes how the convention is used.
The llms.txt Format Specification
llms.txt uses markdown formatting and a simple structure:
# [Company/Product Name]
> [One-line description of what you do]
[2-3 paragraph summary of your product/business]
## [Section: Product/Services]
- [Key feature or service]: [URL]
- [Key feature or service]: [URL]
## [Section: Solutions]
- [Target audience/use case]: [URL]
## [Section: Learn/Content]
- [Article title]: [URL]
- [Article title]: [URL]
## [Section: Company]
- [Company page]: [URL]
The format is intentionally simple — plain text, markdown-ish, human-readable.
What to Include in Your llms.txt
1. Product/Business Summary (Required)
Open with a clear explanation of what your business does, who it serves, and why it matters. Write this for a reader who knows nothing about you:
# Scope — AI Visibility Platform
> AI visibility monitoring and optimization for local businesses, SaaS companies, and marketing agencies.
Scope monitors how your business appears across ChatGPT, Claude, Gemini, and Perplexity.
When someone asks an AI assistant about businesses in your category, Scope shows you
whether you appear — and exactly how to improve your visibility.
Millions of consumers now discover businesses through AI recommendations. Scope helps
businesses understand and improve their position in this new discovery channel.
2. What You Do (Required)
A structured bullet list of your core capabilities:
## What Scope does
- Scans ChatGPT, Claude, Gemini, and Perplexity using prompts relevant to your business
- Calculates an AI Visibility Score (0–100) showing how often AI recommends you
- Identifies which competitors AI recommends instead
- Provides a prioritized optimization action plan
- Monitors visibility daily and alerts when scores change
3. Who You Serve (Recommended)
Be specific about your target audience:
## Who Scope is for
- Local businesses (dentists, plumbers, restaurants, law firms, salons)
- SaaS companies and software products
- Marketing agencies managing multiple clients
- Enterprise marketing teams tracking AI share of voice
4. Pricing (Recommended for Commercial Sites)
AI platforms frequently mention pricing when recommending tools. Including it in llms.txt ensures accuracy:
## Pricing
- Free scan: One-time AI visibility check, no credit card required
- Starter ($29/month): 1 business, 25 prompts monitored, 4 AI platforms
- Growth ($49/month): Up to 3 businesses, 100 prompts per business, team seats
5. Key Content Links (Recommended)
Link to your most important content pages, organized by type:
## Product
- Free AI visibility scan: https://scope.online/scan
- Features overview: https://scope.online/features
- Pricing details: https://scope.online/pricing
## Learn
- What is AI visibility: https://scope.online/learn/what-is-ai-visibility
- What is GEO: https://scope.online/learn/what-is-geo-generative-engine-optimization
- AI search statistics 2026: https://scope.online/learn/ai-search-statistics-2026
- All guides: https://scope.online/learn
## Glossary
- AI Visibility: https://scope.online/glossary/ai-visibility
- GEO: https://scope.online/glossary/geo
- All terms: https://scope.online/glossary
6. Company Information (Recommended)
## Company
Scope Online Inc.
Website: https://scope.online
Founded: 2024
- About: https://scope.online/about
- Privacy: https://scope.online/privacy
llms.txt vs. llms-full.txt
The convention supports two files:
llms.txt — Concise summary (what we've described above). Ideal for quick AI comprehension. Keep it under 5,000 words.
llms-full.txt — Comprehensive index with full descriptions, article summaries, and detailed content listings. This is for AI systems that want to deeply understand your content architecture. It can be much longer.
For most businesses, start with llms.txt and add llms-full.txt once you have substantial content to index.
Implementation
Static File Implementation
Create llms.txt as a plain text file and place it in your website's root directory (the same level as your robots.txt and sitemap.xml):
/public
/llms.txt
/robots.txt
/sitemap.xml
Dynamic Implementation (Next.js Example)
For content-rich sites, generating llms.txt dynamically ensures it stays current as you add content:
// app/llms.txt/route.ts
import { NextResponse } from 'next/server';
import fs from 'fs';
import path from 'path';
export async function GET() {
// Read all learn articles
const learnDir = path.join(process.cwd(), 'content/learn');
const articles = fs.readdirSync(learnDir)
.filter(f => f.endsWith('.mdx'))
.map(f => f.replace('.mdx', ''));
const content = `# Your Business Name
> Your one-line description
[Your business summary]
## Learn (${articles.length} articles)
${articles.map(slug => `- Article: https://yourdomain.com/learn/${slug}`).join('\n')}
`;
return new NextResponse(content, {
headers: { 'Content-Type': 'text/plain; charset=utf-8' }
});
}
Linking From robots.txt
There's no formal standard for referencing llms.txt, but some implementations add a comment to robots.txt:
# LLM content index
# LLMs-txt: https://yourdomain.com/llms.txt
What llms.txt Cannot Do
Be clear about what llms.txt is not:
- It is not a sitemap — you still need
sitemap.xmlfor search engine indexing - It is not a guarantee — no AI platform is obligated to read or use your
llms.txt - It is not a substitute for schema markup, good content, or citation building
- It is not officially supported by all major AI providers (though adoption is growing)
Keeping It Current
Your llms.txt should be updated whenever:
- You add significant new content sections
- Your pricing changes
- You add major new product features
- You add new industry solutions or comparison pages
If you use dynamic generation (recommended), it stays current automatically.
Q: Does Google read llms.txt?
A: Google has not officially committed to using llms.txt in any specific way. However, Googlebot does crawl text files in website roots, and Gemini's retrieval systems may incorporate this content. It's a low-cost implementation regardless.
Q: Is there a validator for llms.txt?
A: There's no official validator, but you can test that your llms.txt is accessible by visiting yourdomain.com/llms.txt in a browser. It should display as readable plain text (not download as a file).
Q: Can llms.txt help with Perplexity citations?
A: Perplexity's real-time retrieval system can read any publicly accessible content, including llms.txt. When Perplexity crawls your site, it may incorporate llms.txt as a high-value summary document. Clear, structured content in llms.txt could contribute to more accurate Perplexity citations.