llms.txt Generator

llms.txt is the newest piece of the AI-visibility stack: one markdown file at your root that hands language models a clean, curated map of your best content. Enter your site, summary, and key links — and get a spec-compliant llms.txt to ship in minutes.

llms.txt is a markdown file at your site root, proposed by Jeremy Howard, that gives AI models a token-efficient map of your most important content — an H1 name, a blockquote summary, and sections of links. This generator builds a spec-compliant file from your inputs. It complements JSON-LD schema (meaning) and clean HTML/markdown (retrievable substance); layer all three for full AI readability. Runs in your browser.

The calculator

Site / project name iBecomes the H1.

One-line summary iBecomes the blockquote.

Key links iTitle | URL | note — use “## Heading” for sections.

Spec-compliant

Links included

0sections

—structure

0characters

Export

Generated llms.txt — save at the root of your site (/llms.txt)

Walkthrough

How to use this tool

Name the site and summarize it.The name becomes the H1; the summary becomes a one-line blockquote describing what the site is and who it serves.
List your most important links.Point AIs at the pages you most want them to read — docs, key guides, product, pricing. Format: Title | URL | optional note.
Group with sections.Start a line with “##” to create a section (e.g. ## Docs, ## Optional). An “Optional” section signals lower-priority links.
Copy and save at the root.Save the output as /llms.txt at your domain root, alongside robots.txt. Many AI tools and crawlers look for it there.
Pair with clean markdown.For best results, also offer markdown versions of key pages and keep your JSON-LD schema — llms.txt for navigation, schema for grounding.

From the desk

RGM Expert Says

Real Growth Matters — Paid social practiceHow we use this tool with clients

llms.txt is the newest piece of the AI-visibility stack, proposed by Jeremy Howard in late 2024: a single markdown file at your site root that gives language models a clean, token-efficient map of your most important content. Think of it as robots.txt’s helpful cousin — not “what you may crawl,” but “here is what matters and where to find it,” in the format LLMs parse most cheaply.

We treat it as complementary to schema, not a replacement. The honest framing from the research: JSON-LD gives superior factual grounding and entity recognition, while llms.txt offers higher token efficiency and better navigation for real-time AI crawlers. A hybrid wins — schema so engines understand what your content means, llms.txt (plus clean markdown page versions) so they can find and read it efficiently.

Two practical notes. First, adoption is still emerging — not every engine consumes llms.txt yet, so treat it as low-cost insurance, not a silver bullet. Second, point it at your genuinely best, most current pages and keep it maintained; a stale llms.txt that references dead links is worse than none. Generate it here, ship it at the root, and refresh it when your key content changes.

The math

How it works

The llms.txt format is a strict but simple markdown structure (Jeremy Howard’s proposed spec):

H1 name → > blockquote summary → optional prose → ## sections of [title](url): note bullets

H1 — a single # Name at the top: the site or project name.
Blockquote — a > summary line describing the site’s purpose.
Sections — ## Heading blocks containing markdown bullet links.
Links — - [Title](URL): note, pointing at your priority content.
Optional — a section literally named “Optional” marks lower-priority links a crawler can skip.

Follows the llms.txt proposal by Jeremy Howard. The generator builds the structure in your browser; save the output as /llms.txt at your domain root.

Why it matters

Markdown vs schema: why you want both

There’s a live debate about whether AI visibility comes from structured data (JSON-LD) or from clean, token-efficient markdown (llms.txt and markdown page versions). The honest answer is that they solve different problems. Schema is for meaning — it tells an engine this is an organization, this is an author, this is a question and its answer, with machine precision. llms.txt and markdown are for access — they let a model find and ingest your priority content cheaply, without wading through navigation, scripts, and markup.

Treating them as rivals is the mistake. The strongest AI-visibility setups layer all three: JSON-LD schema for grounding, an llms.txt for navigation, and clean markdown or server-rendered HTML so the substance is actually retrievable. Each closes a different gap an AI engine can fall into — misunderstanding you, failing to find your best pages, or choking on un-rendered content.

llms.txt is the cheapest of the three to ship and the easiest to forget. It’s one file, it takes minutes with this generator, and it costs nothing but the discipline to keep it current. As more engines adopt it, the sites that already maintain a clean, accurate llms.txt will have handed the machines a curated tour of exactly what they want quoted.

Benchmarks

The AI-readability stack

Layer all three — they cover different gaps.

Source: llms.txt proposal; comparative LLM-optimization analysis of llms.txt vs JSON-LD.
Layer	Solves	Format
JSON-LD schema	Meaning / entity grounding	Structured
llms.txt	Navigation to priority content	Markdown
Markdown / rendered HTML	Retrievable substance	Text

Voices worth trusting

What operators say

JSON-LD provides superior factual grounding for entity recognition, while llms.txt offers higher token efficiency and better navigation for real-time AI crawlers — a hybrid approach wins.

LLM-optimization analysis

Markdown vs JSON-LD (paraphrase)

llms.txt is a single markdown file at your site root that gives language models a clean, curated map of your most important content.

llms.txt proposal

Jeremy Howard (paraphrase)

Go deeper

Books worth reading

Content Strategy for the WebKristina HalvorsonCurate and structure what matters — the instinct behind a good llms.txt.The Art of SEOEnge, Spencer & StricchiolaCrawlability and access fundamentals AI retrieval depends on.Everybody WritesAnn HandleyClear, clean writing — what markdown surfaces best to machines.

Related on RGM

Keep learning

Course

Glossary

Concepts

Tools

FAQ

Common questions

What is llms.txt?

A single markdown file at a site’s root, proposed by Jeremy Howard in late 2024, that gives language models a clean, token-efficient map of the site’s most important content — an H1 name, a blockquote summary, and sections of markdown links.

How is llms.txt different from robots.txt?

robots.txt tells crawlers what they may access; llms.txt tells AI models what matters and where to find it, in markdown they parse efficiently. They serve different purposes and can coexist at the root.

Should I use llms.txt or JSON-LD schema?

Both. JSON-LD gives engines factual grounding and entity recognition (meaning); llms.txt gives token-efficient navigation to your priority content (access). The strongest setups layer schema, llms.txt, and clean retrievable HTML/markdown together.

Where do I put the llms.txt file?

At your domain root, e.g. https://yoursite.com/llms.txt, alongside robots.txt. Save the generated output as a plain-text file named llms.txt.

Do AI engines actually use llms.txt yet?

Adoption is still emerging — not every engine consumes it — so treat it as low-cost insurance rather than a guarantee. It’s one file, quick to ship, and positions you well as adoption grows. Keep it current and pointing at live, high-value pages.

What should I include in llms.txt?

Your genuinely most important, current pages: docs, key guides, product, pricing, and core hubs. Use a section literally named ‘Optional’ for lower-priority links a crawler can skip.

Related tools