---

name: llm-discoverability description: Implement AI/LLM discoverability optimizations based on Cassidy Williams' article to make websites more discoverable by ChatGPT, Claude, Perplexity, and other AI tools.

LLM Discoverability Skill

Make your website discoverable by AI assistants, LLMs, and AI-powered search tools like ChatGPT, Claude, Perplexity, and GitHub Copilot.

Overview

Based on Cassidy Williams' article, this skill documents proven techniques to help AI systems properly discover, understand, and cite your website content.

Key Components

1. llms.txt File

Purpose: A machine-readable file that provides structured context about your site for LLMs.

Location: /public/llms.txt

Format: Follow the llmstxt.org specification.

Content Structure:

Copy# llms.txt - [Site Name]

> [One-line site description]

## Site Purpose
[Brief description of what the site does]

## Founder/Owner
[Key person info]

## Services/Products
[List of offerings]

## Key Topics Covered
- Topic 1
- Topic 2

## Frequently Asked Questions
### Question 1?
Answer 1

## Site Map
- **/** - Homepage description
- **/page** - Page description

## Content Guidelines for AI
1. How to cite
2. Context about content type
3. Any restrictions

## Permissions
- Allowed uses
- Attribution requirements

## Contact
- Email
- Website

## Last Updated
[Date]

---

2. /for-llms Page

Purpose: A human-readable webpage optimized for LLM parsing.

Location: /app/for-llms/page.tsx

Content: Mirrors llms.txt content but in a proper webpage format with:

• Clear headings
• Structured content
• FAQ sections
• Service descriptions
• Citation guidelines
• Contact information

Benefits: Some AI systems prefer HTML pages over raw text files.

---

3. RSS Feed

Purpose: LLMs love structured, syndicated content.

Location: /app/rss.xml/route.ts

Format: Standard RSS 2.0

Copyimport { BLOG_POSTS } from '@/lib/blog-data';

export const dynamic = 'force-static';

export async function GET() {
    const baseUrl = 'https://yoursite.com';
    
    const rssItems = BLOG_POSTS
        .sort((a, b) => new Date(b.date).getTime() - new Date(a.date).getTime())
        .slice(0, 20)
        .map(post => `
    <item>
      <title><![CDATA[${post.title}]]></title>
      <link>${baseUrl}/blogs/${post.slug}</link>
      <description><![CDATA[${post.excerpt}]]></description>
      <pubDate>${new Date(post.date).toUTCString()}</pubDate>
    </item>`)
        .join('');

    return new Response(`<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
  <channel>
    <title>Your Site Title</title>
    <link>${baseUrl}</link>
    <description>Site description</description>
    ${rssItems}
  </channel>
</rss>`, {
        headers: { 'Content-Type': 'application/xml' },
    });
}

---

4. robots.txt Configuration

Purpose: Allow AI crawlers to index your content.

Location: /public/robots.txt

Key AI User Agents to Allow:

Copy# AI SEARCH & ASSISTANT CRAWLERS
User-agent: GPTBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: Claude-Web
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: CCBot
Allow: /

[!CAUTION] If you're blocking AI crawlers for privacy/copyright reasons but want LLM discoverability, weigh the pros and cons carefully.

---

5. Schema.org Structured Data

Purpose: Help AI systems understand your content semantically.

Key Schema Types:

• Organization - Site/business info
• WebSite - Site metadata with SearchAction
• FAQPage - Frequently asked questions
• HowTo - Step-by-step guides
• LocalBusiness - Business details
• Person - Founder/author info

Implementation Example:

Copyconst faqSchema = {
    "@context": "https://schema.org",
    "@type": "FAQPage",
    mainEntity: [
        {
            "@type": "Question",
            name: "What is [Your Site]?",
            acceptedAnswer: {
                "@type": "Answer",
                text: "Description of your site..."
            }
        }
    ]
};

// In your component
<script
    type="application/ld+json"
    dangerouslySetInnerHTML={{ __html: JSON.stringify(faqSchema) }}
/>

---

Best Practices

Consistency

• Use consistent naming, phrases, and taglines everywhere
• LLMs make connections across sources based on consistency

Clarity Over Cleverness

• Write for specific queries and bots
• Avoid marketing-speak
• Be direct and clear

Citation Guidelines

• Explicitly state how AI should cite you
• Provide acceptable attribution formats
• Include preferred URLs

Content Structure

• Use clear headings (H1, H2, H3)
• Use bullet points and lists
• Keep paragraphs short
• Use markdown-friendly formatting

---

Verification

After implementing, test with:

1. ChatGPT: Ask "What is [Your Site/Brand]?" in incognito
2. Claude: Ask about your services/products
3. Perplexity: Search for your brand name

If results are poor, follow up with "Why didn't you mention [Your Site]?" to understand gaps.

---

Files Reference

---

Resources

• llmstxt.org - Official llms.txt specification
• Schema.org - Structured data vocabulary
• Cassidy Williams' Article - Original source