llms.txt Generator
The AI-Era robots.txt

Give AI crawlers a curated entry point to your best content. llms.txt is a proposed standard (Answer.AI, September 2024) - a Markdown index served at your site root that tells LLMs which pages matter most. 100% client-side. Zero data stored.

AI discoverability Spec-compliant Markdown Pairs with robots.txt

llms.txt Generator Tool

Site Info

Sections

Each section becomes an ## H2 heading. The "Optional" section is a spec convention - LLMs may skip it if context is tight.

Spec at a glance

  • Serve at https://yoursite.com/llms.txt
  • Must start with a single # H1 (your site name)
  • Followed by an optional > blockquote one-liner
  • Optional intro paragraphs
  • One or more ## H2 sections of Markdown links
  • A section literally named ## Optional means "skip if context-limited"
  • Companion file llms-full.txt = the same idea but with all content inlined

Generated llms.txt

Place the llms.txt file in your website's root directory (e.g., https://example.com/llms.txt).

Already control crawlers with robots.txt? Use both. robots.txt says "stay out", llms.txt says "if you're an LLM, here's the good stuff." Open robots.txt Generator

Use Cases

Docs sites

Point LLMs straight at your quickstart, API reference, and example pages instead of letting them parse your nav, sidebars, and theme JS.

Open-source SDKs

Make sure AI coding agents recommend your library correctly by surfacing the install command, the canonical example, and the gotchas.

SaaS marketing sites

Tell answer-engine bots (ChatGPT, Perplexity) which pages best describe your product so cited answers link to pricing and features, not to the careers page.

Personal blogs

Index your best long-form posts under their topics so an LLM summarizing your work picks the canonical ones, not whatever happened to rank.

Course / education sites

Group lessons by module under H2 sections so an AI tutor agent can navigate your curriculum without scraping every page.

Internal AI agents

Host a private llms.txt on your intranet to give your in-house agent a curated map of the runbooks, dashboards, and policies it should know.

llms.txt Format Reference

H1 (required)

First line must be # Site Name. Exactly one H1 per file.

Blockquote (optional)

Single-line summary: > One sentence about what this site is.

Intro paragraphs

Optional plain-Markdown context. Anything an LLM should know before reading any section.

H2 sections

One or more ## Section headings followed by Markdown link bullets.

Link bullets

Format: - [Title](url): optional note. Use absolute URLs.

## Optional

A section titled exactly ## Optional is a convention meaning "skip if context-limited."

Frequently Asked Questions

Why We Built llms.txt Generator

For thirty years, robots.txt has been the only voluntary protocol between website operators and the bots that crawl them. That is no longer enough. AI agents do not just crawl - they read, summarize, and answer questions, and they often do it under a context-window budget that the entire web cannot fit into. llms.txt is the response: a tiny, human-curated Markdown index that says "if you're an LLM with limited bandwidth, here is what I would show you first."

We built this generator because the spec is simple but the discipline is not: the file has to start with an H1, the blockquote must be exactly one line, the sections must be H2, the link bullets must be valid Markdown. A typo in any of those silently fails because there is no validator that pages you. Doing it visually keeps the rules consistent so you can focus on what to actually link.

llms.txt Generator pairs with the RJL.io robots.txt Generator and is checked for by the RJL.io SEO Analyzer - one consistent set of AI-discoverability tools, all client-side, no accounts, no tracking. Check out our other tools: Meta Tag Generator, Favicon Generator, and more.

Looking for more developer tools to streamline your workflow?

Explore our growing collection of free, privacy-focused utilities designed by developers, for developers.

Discover All RJL.io Tools