@dualnova/llms-txt

Parse, validate, and generate /llms.txt files — the emerging standard for guiding AI search crawlers (ChatGPT, Perplexity, Claude, Gemini).

A tiny TypeScript library and CLI for working with /llms.txt files. Zero runtime dependencies, ESM-only, Node 20+.

🇪🇸 ¿Español? Lee este README en español más abajo.

---

What is llms.txt?

llms.txt is an emerging convention (see llmstxt.org) that publishes a Markdown summary of a site at the well-known path /llms.txt. The goal is to give AI assistants and crawlers a curated, low-noise view of the site — the equivalent of robots.txt but for LLM-friendly content discovery.

A minimal example:

# Example Site

> A short summary of what this site is and who it is for.

## Key pages

- [Home](https://example.com/): main landing page
- [About](https://example.com/about): team and mission
- [Pricing](https://example.com/pricing)

Honest note on impact (May 2026)

As of this writing, primary sources from Google (John Mueller, Gary Illyes), the SE Ranking 300k-domain study and the OtterlyAI server-log audit report that no major AI search engine currently uses llms.txt as a citation-ranking signal. The file is still worth publishing as a low-cost, declarative statement of intent that may be adopted as the standard matures. This library treats llms.txt as a useful piece of site hygiene, not a silver bullet.

If you want measurable AI search visibility today, focus on these in order:

1. Unblock AI crawlers in robots.txt (GPTBot, ClaudeBot, Google-Extended, PerplexityBot).
2. Publish 134-167 word citable passages with the "X is [definition]" pattern on key pages.
3. Build brand mentions on Reddit, YouTube, and Wikipedia — these correlate 3× more strongly with AI citations than backlinks (Ahrefs Dec 2025 study, n=75,000 brands).
4. Publish llms.txt — this library.

Install

npm install @dualnova/llms-txt
# or
pnpm add @dualnova/llms-txt
# or run the CLI without installing
npx @dualnova/llms-txt validate https://example.com/llms.txt

CLI

# Validate the llms.txt in the current directory
llms-txt validate

# Validate a specific local file
llms-txt validate ./public/llms.txt

# Fetch and validate a remote one
llms-txt validate --url https://dualnova.org/llms.txt

# Show help
llms-txt --help

Sample output:

Validating /Users/me/site/public/llms.txt
  Title:        DualNova
  Description:  DualNova is a blockchain and AI software development company …
  Sections:     5
  Links:        18

WARN  Section "Recent client work" has no links and no content. (sections[3])

✓ 0 error(s), 1 warning(s)

Exit codes:

Code	Meaning
0	No errors. Warnings may still be printed.
1	Validation errors found, or invalid CLI arguments.
2	Could not read the file or fetch the URL.

Programmatic API

Validate

import { validateLlmsTxt } from '@dualnova/llms-txt';
import { readFileSync } from 'node:fs';

const source = readFileSync('./public/llms.txt', 'utf8');
const { valid, issues, parsed } = validateLlmsTxt(source);

if (!valid) {
  for (const issue of issues.filter(i => i.severity === 'error')) {
    console.error(`✗ ${issue.message}`);
  }
  process.exit(1);
}

console.log(`Parsed ${parsed.sections.length} sections.`);

Parse

import { parseLlmsTxt } from '@dualnova/llms-txt';

const { title, description, sections } = parseLlmsTxt(source);
for (const section of sections) {
  console.log(`## ${section.heading}`);
  for (const link of section.links) {
    console.log(`  - ${link.title}: ${link.url}`);
  }
}

Build

Generate llms.txt from typed input — useful in build pipelines that read from a CMS, sitemap, or MDX frontmatter.

import { buildLlmsTxt } from '@dualnova/llms-txt';
import { writeFileSync } from 'node:fs';

const md = buildLlmsTxt({
  title: 'Acme Robotics',
  description: 'Acme Robotics builds autonomous mobile robots for warehouse logistics.',
  intro: 'Founded in 2021. HQ in Pittsburgh, PA.',
  sections: [
    {
      heading: 'Key pages',
      links: [
        { title: 'Home', url: 'https://acme-robotics.example/', description: 'overview' },
        { title: 'Products', url: 'https://acme-robotics.example/products' },
      ],
    },
  ],
});

writeFileSync('./public/llms.txt', md);

Use in a Next.js build

Add a route at app/llms.txt/route.ts:

import { buildLlmsTxt } from '@dualnova/llms-txt';

export function GET() {
  const body = buildLlmsTxt({
    title: 'My Site',
    description: 'One-sentence summary.',
    sections: [
      { heading: 'Pages', links: [{ title: 'Home', url: 'https://my-site.example/' }] },
    ],
  });
  return new Response(body, {
    headers: { 'content-type': 'text/plain; charset=utf-8' },
  });
}

export const dynamic = 'force-static';

Use in CI to fail builds on broken llms.txt

# .github/workflows/llms-txt.yml
name: Validate llms.txt
on: [push, pull_request]
jobs:
  validate:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with: { node-version: '20' }
      - run: npx -y @dualnova/llms-txt validate ./public/llms.txt

Types

interface ParsedLlmsTxt {
  title: string;
  description?: string;
  intro: string;
  sections: LlmsSection[];
}

interface LlmsSection {
  heading: string;
  links: LlmsLink[];
  freeform: string;
}

interface LlmsLink {
  title: string;
  url: string;
  description?: string;
}

interface ValidationIssue {
  severity: 'error' | 'warning' | 'info';
  message: string;
  at?: string;
}

Examples in this repo

• examples/minimal.txt — bare minimum that passes validation
• examples/full.txt — a realistic small-company example

Contributing

PRs welcome — please open an issue first for non-trivial changes. The codebase is intentionally tiny (3 files in src/, no runtime deps). Tests use Node's built-in test runner.

License

MIT © DualNova LLC

---

🇪🇸 Léeme en español

¿Qué es llms.txt?

llms.txt es una convención emergente (ver llmstxt.org) que publica un resumen en Markdown del sitio en la ruta well-known /llms.txt. El objetivo es darle a los crawlers e IA una vista curada y de bajo ruido del sitio — el equivalente de robots.txt pero para descubrimiento de contenido amigable para LLMs.

Nota honesta sobre impacto (mayo 2026)

A la fecha, fuentes primarias de Google (John Mueller, Gary Illyes), el estudio de SE Ranking con 300k dominios y la auditoría de logs de OtterlyAI reportan que ningún motor de búsqueda AI mayor usa hoy llms.txt como señal de ranking de citaciones. Vale la pena publicarlo igual como declaración de bajo costo de intención que puede adoptarse a medida que madura el standard. Esta librería trata a llms.txt como higiene útil del sitio, no como bala de plata.

Si quieres visibilidad medible en AI search HOY, prioriza en este orden:

1. Desbloquea los AI crawlers en robots.txt (GPTBot, ClaudeBot, Google-Extended, PerplexityBot).
2. Publica pasajes citables de 134-167 palabras con el patrón "X es [definición]" en páginas clave.
3. Construye menciones de marca en Reddit, YouTube y Wikipedia — estas correlacionan 3× más con citaciones AI que los backlinks (estudio Ahrefs Dic 2025, n=75,000 marcas).
4. Publica llms.txt — esta librería.

Instalación y uso

Los ejemplos arriba aplican igual. La CLI tiene la misma sintaxis. Toda la documentación de la API es traducible directamente.

npm install @dualnova/llms-txt
npx @dualnova/llms-txt validate ./public/llms.txt

Licencia

MIT © DualNova LLC — equipo bilingüe (español + inglés) basado en Caracas, Bogotá y Miami. Construido como parte del esfuerzo abierto de Generative Engine Optimization (GEO) que mantenemos en dualnova.org.

---

Built by DualNova — blockchain and AI software development for LATAM and the US.