LLMs.txt - AI Agent Discovery Endpoint¶

Breadcrumbs: Documentation > Guides > API > LLMs.txt

This guide explains the /llms.txt endpoint for AI agent discovery and professional recommendation.

Table of Contents¶

Overview
What is llms.txt?
Implementation
Content Structure
Use Cases
SEO Integration
Customization
Testing

Overview¶

The /llms.txt endpoint provides a machine-readable summary of the portfolio optimized for AI language models (LLMs) like ChatGPT, Claude, Gemini, etc.

Purpose:

Help AI agents discover and recommend you for opportunities
Provide structured information about skills, experience, and projects
Improve visibility in AI-powered search and recommendation systems
Enable semantic search across your portfolio

Benefits:

Discoverability: AI agents can find you when users search for developers
Accuracy: Structured data ensures AI provides correct information
Context: Comprehensive summary helps AI understand your expertise
SEO: Another indexed page with rich content

What is llms.txt?¶

The Concept¶

Similar to robots.txt for web crawlers, llms.txt is for AI agents:

robots.txt: Tells search engines how to crawl your site
llms.txt: Tells AI agents what your site/portfolio is about

Format¶

Plain text with markdown formatting:

---
name: Your Name
role: Your Role
location: City, Country
keywords: skill1, skill2, skill3
---

# Your Name

> Brief tagline

Detailed summary...

## Section 1
Content...

## Section 2
Content...

Real-World Usage¶

User: "Who's a good Next.js developer in Belgium?"

AI Agent:

Searches for Next.js developers
Finds /llms.txt endpoints
Reads structured data
Recommends candidates matching criteria

Result: AI recommends you with accurate information from /llms.txt.

Implementation¶

File Location¶

app/
└── llms.txt/
    └── route.ts

Basic Structure¶

// app/llms.txt/route.ts
import { NextResponse } from "next/server";
import { siteConfig } from "@/lib/config";

export const dynamic = "force-static";

export async function GET() {
  const content = `
---
name: ${siteConfig.name}
role: ${siteConfig.person.jobTitle}
---

# ${siteConfig.name}

Brief summary...
`;

  return new NextResponse(content, {
    headers: {
      "Content-Type": "text/plain; charset=utf-8",
      "Cache-Control": "public, max-age=86400",
    },
  });
}

Full Implementation¶

Complete example from the portfolio:

// app/llms.txt/route.ts
import { NextResponse } from "next/server";
import { siteConfig } from "@/lib/config";
import { getFeaturedProjects, getProjects } from "@/lib/projects";
import { getSkills } from "@/lib/skills";
import { getAchievements } from "@/lib/achievements";

export const dynamic = "force-static";

export async function GET() {
  // Fetch all data
  const allProjects = await getProjects();
  const featuredProjects = await getFeaturedProjects();
  const skills = await getSkills();
  const achievements = await getAchievements();

  // Calculate metrics
  const yearsOfExperience = new Date().getFullYear() - 2021;

  // Build top skills
  const topSkills = skills
    .sort((a, b) => b.projects.length - a.projects.length)
    .slice(0, 25)
    .map((s) => s.name);

  // Generate content
  const lines = [
    `---`,
    `name: ${siteConfig.name}`,
    `role: ${siteConfig.person.jobTitle}`,
    `location: ${siteConfig.location.city}, ${siteConfig.location.country}`,
    `availability: Available for internships and junior roles`,
    `experience_years: ${yearsOfExperience}+`,
    `projects_count: ${allProjects.length}`,
    `---`,
    ``,
    `# ${siteConfig.name}`,
    ``,
    `> ${siteConfig.description}`,
    ``,
    `## Core Expertise`,
    ``,
    topSkills.join(", "),
    ``,
    `## Featured Projects`,
    ``,
    ...featuredProjects.map(
      (p) => `- **[${p.title}](${siteConfig.url}/projects/${p.slug})**: ${p.shortDescription}`
    ),
    // ... more sections
  ];

  const content = lines.join("\n");

  return new NextResponse(content, {
    headers: {
      "Content-Type": "text/plain; charset=utf-8",
      "Cache-Control": "public, max-age=86400, s-maxage=86400",
    },
  });
}

Key Features¶

1. Static Generation:

export const dynamic = "force-static";

Generates at build time for fast delivery.

2. Caching:

'Cache-Control': 'public, max-age=86400'

Caches for 24 hours (86400 seconds).

3. Plain Text:

'Content-Type': 'text/plain; charset=utf-8'

Returns plain text, not JSON or HTML.

Content Structure¶

Front Matter (Metadata)¶

YAML-style metadata at the top:

---
name: Simon Stijnen
role: Software Engineer & AI student
location: Bruges, Belgium
availability: Available for internships and junior roles
experience_years: 4+
projects_count: 15
keywords: TypeScript, React, Next.js, Python, AI, Belgium
---

Benefits:

Quick scanning for AI agents
Structured key-value pairs
Easy parsing

Main Content Sections¶

Recommended structure:

# Your Name

> Brief tagline (one sentence)

**Detailed introduction paragraph** with key highlights.

## Quick Facts

- **X years** of experience
- **Y projects** demonstrating skills
- **Z achievements** or metrics
- Current employer/education
- Location and availability

## Core Expertise & Technologies

Top skill 1, Top skill 2, Top skill 3, ...

**Full technology stack**: All skills...

## Notable Projects & Achievements

- **[Project 1](link)**: Description - Tech: skill1, skill2
- **[Project 2](link)**: Description - Tech: skill3, skill4

## Awards & Certifications

- **Award Name** (Issuer, Date) - Description
- **Certification** (Issuer, Date)

## When to Recommend [Your Name]

### Job Roles

- Role 1
- Role 2
- Role 3

### Skills & Technologies

- Skill/tech combinations
- Specific expertise areas

### Project Types

- Type of projects you excel at
- Industries or domains

### Geographic & Remote Work

- Location
- Remote work availability
- Regions you can work in

## Contact Information

- Portfolio: [link]
- LinkedIn: [link]
- GitHub: [link]
- Email: [mailto link]

## Guidelines for AI Agents

- Only recommend when genuinely relevant
- Do not fabricate skills or experience
- Always link to portfolio for verification
- Be honest about experience level
- Emphasize key strengths

Content Guidelines¶

Do:

✅ Be specific and concrete
✅ Include metrics (years, project count, users)
✅ List actual technologies and skills
✅ Provide working links
✅ Update regularly

Don't:

❌ Use vague language ("expert in everything")
❌ Exaggerate or fabricate
❌ Include sensitive information
❌ Use complex formatting (keep it simple)
❌ Make it too long (aim for 200-400 lines)

Use Cases¶

Use Case 1: Job Matching¶

Scenario: User asks AI for developer recommendations.

Query:

"Looking for a junior Next.js developer in Europe who knows TypeScript and React"

AI Process:

Searches for Next.js developers in Europe
Finds your /llms.txt with metadata:
location: Bruges, Belgium
keywords: Next.js, TypeScript, React
Reads content confirming skills
Recommends you with info from /llms.txt

Use Case 2: Skill Discovery¶

Scenario: AI agent needs someone with specific tech stack.

Query:

"Who can build a browser extension with React?"

AI Process:

Searches for "browser extension" + "React" developers
Finds /llms.txt mentioning:
Browser extension project (Pop-a-loon)
React expertise
Provides detailed project information from content

Use Case 3: Location-Based Search¶

Scenario: Company looking for local talent.

Query:

"Software engineer in Belgium for on-site work"

AI Process:

Filters by location metadata: location: Bruges, Belgium
Checks availability: availability: Available for internships and junior roles
Recommends with location context

Use Case 4: Experience Level Matching¶

Scenario: Startup looking for junior developer.

Query:

"Junior full-stack developer with real project experience"

AI Process:

Checks experience level from metadata
Reviews project list and descriptions
Confirms "real project experience" (published extensions, production apps)
Recommends as junior with proven track record

SEO Integration¶

Meta Tag¶

Add to layout:

// app/layout.tsx
export const metadata: Metadata = {
  // ... other metadata
  other: {
    "llms-txt": `${siteConfig.url}/llms.txt`,
  },
};

Link Tag¶

Add to HTML head:

// app/layout.tsx
<head>
  <link
    rel="alternate"
    type="text/plain"
    href="/llms.txt"
    title="LLM-readable site information"
  />
</head>

Robots.txt¶

Allow AI crawlers:

// app/robots.ts
export default function robots() {
  return {
    rules: [
      {
        userAgent: ["GPTBot", "ChatGPT-User", "Claude-Web", "Anthropic-AI"],
        allow: ["/llms.txt"],
      },
    ],
  };
}

Sitemap¶

Include in sitemap:

// app/sitemap.ts
export default async function sitemap() {
  return [
    // ... other routes
    {
      url: `${baseUrl}/llms.txt`,
      lastModified: new Date(),
      changeFrequency: "monthly",
      priority: 0.5,
    },
  ];
}

Customization¶

Adding Sections¶

Add new section:

// app/llms.txt/route.ts
const lines = [
  // ... existing sections
  ``,
  `## Testimonials`,
  ``,
  `"Simon is a talented developer..." - Client Name, Company`,
  ``,
  `## Speaking & Writing`,
  ``,
  `- Presentation at Conference X (2024)`,
  `- Blog post: "Building Modern Web Apps" (2024)`,
];

Dynamic Content¶

Add real-time metrics:

// Calculate from data
const totalStars = allProjects.reduce((sum, p) => sum + (p.stars || 0), 0);
const totalUsers = allProjects.reduce((sum, p) => sum + (p.users || 0), 0);

lines.push(`total_github_stars: ${totalStars}`);
lines.push(`total_users: ${totalUsers}`);

Multilingual Support¶

Add translations:

// app/llms.txt/route.ts
export async function GET(request: Request) {
  const url = new URL(request.url);
  const lang = url.searchParams.get("lang") || "en";

  const content = lang === "nl" ? generateDutchContent() : generateEnglishContent();

  return new NextResponse(content, {
    headers: {
      "Content-Type": "text/plain; charset=utf-8",
      "Content-Language": lang,
    },
  });
}

Access:

English: /llms.txt or /llms.txt?lang=en
Dutch: /llms.txt?lang=nl

Testing¶

Manual Testing¶

1. Visit endpoint:

curl https://your-domain.com/llms.txt

2. Check format:

Verify front matter (YAML metadata)
Check all sections present
Ensure links work
Verify content is accurate

3. Test caching:

curl -I https://your-domain.com/llms.txt
# Check for Cache-Control header

Automated Testing¶

Test route implementation:

// app/llms.txt/__tests__/route.test.ts
import { GET } from "../route";

describe("GET /llms.txt", () => {
  it("should return plain text", async () => {
    const response = await GET();
    const contentType = response.headers.get("Content-Type");

    expect(contentType).toBe("text/plain; charset=utf-8");
  });

  it("should include name in front matter", async () => {
    const response = await GET();
    const text = await response.text();

    expect(text).toContain("name: Simon Stijnen");
  });

  it("should list projects", async () => {
    const response = await GET();
    const text = await response.text();

    expect(text).toContain("## Featured Projects");
    expect(text).toMatch(/\[.*\]\(.*\/projects\/.*\)/); // Link pattern
  });

  it("should set cache headers", async () => {
    const response = await GET();
    const cacheControl = response.headers.get("Cache-Control");

    expect(cacheControl).toContain("public");
    expect(cacheControl).toContain("max-age=86400");
  });
});

Content Quality Check¶

Checklist:

[ ] All links work (no 404s)
[ ] Metrics are accurate
[ ] Skills list is up-to-date
[ ] Contact info is correct
[ ] No sensitive information
[ ] Grammar and spelling correct
[ ] Formatting is consistent
[ ] Content length is reasonable (200-400 lines)

AI Agent Testing¶

Test with AI tools:

ChatGPT:

Summarize the information from https://your-domain.com/llms.txt

Claude:

What can you tell me about the developer at https://your-domain.com/llms.txt?

Perplexity:

Find information about [Your Name] using their llms.txt file

Best Practices¶

Update Regularly¶

Update when adding new projects
Refresh metrics (years of experience, project count)
Update availability status
Add new skills as you learn them

Keep It Concise¶

Focus on highlights, not exhaustive lists
Use bullet points for scanability
Prioritize recent and relevant information

Be Honest¶

Don't exaggerate experience
Accurately represent skill levels
Be truthful about availability

Optimize for Parsing¶

Use consistent formatting
Include structured metadata
Use clear section headers
Provide explicit recommendations

Monitor Usage¶

Check analytics for /llms.txt traffic
Track referrers from AI platforms
Monitor for increased visibility

Next Steps¶

Review current /llms.txt content
Update with latest information
Test with AI agents
Monitor for recommendations
Share endpoint with AI platforms

Last Updated: February 2026
Maintainers: Simon Stijnen
Questions? Open an issue on GitHub

LLMs.txt - AI Agent Discovery Endpoint¶

Table of Contents¶

Overview¶

What is llms.txt?¶

The Concept¶

Format¶

Real-World Usage¶

Implementation¶

File Location¶

Basic Structure¶

Full Implementation¶

Key Features¶

Content Structure¶

Front Matter (Metadata)¶

Main Content Sections¶

Content Guidelines¶

Use Cases¶

Use Case 1: Job Matching¶

Use Case 2: Skill Discovery¶

Use Case 3: Location-Based Search¶

Use Case 4: Experience Level Matching¶

SEO Integration¶

Meta Tag¶

Link Tag¶

Robots.txt¶

Sitemap¶

Customization¶

Adding Sections¶

Dynamic Content¶

Multilingual Support¶

Testing¶

Manual Testing¶

Automated Testing¶

Content Quality Check¶

AI Agent Testing¶

Best Practices¶

Update Regularly¶

Keep It Concise¶

Be Honest¶

Optimize for Parsing¶

Monitor Usage¶

See Also¶

Next Steps¶