GEO industry updates — straight to your inbox.
What is
llms.txt?
An open standard — proposed by Jeremy Howard of Answer.AI in September 2024 — that gives AI models a structured, authoritative summary of your website, so they represent your brand accurately at inference time.
The problem it solves
AI models increasingly rely on websites for information — but they can't read your entire site. Context windows are too small, HTML is noisy with ads and navigation, and automatically converting web pages into something LLM-friendly is imprecise.
Without a curated source of truth, models guess. They mix up your positioning, cite outdated pricing, or confuse you with competitors.
llms.txt lets site authors — who know their content best — tell AI systems exactly what to read. Think of it as a robots.txt for AI: instead of blocking crawlers, you're curating what matters most.
# Acme Corp > Cloud infrastructure tools for engineering teams. > Helping 1,200+ companies ship faster with zero-downtime > deployments, real-time observability, and secrets management. ## Products - [Acme Deploy](https://acmecorp.com/deploy/index.html.md): Zero-downtime deployment pipeline for Kubernetes - [Acme Monitor](https://acmecorp.com/monitor/index.html.md): Real-time observability and alerting - [Acme Vault](https://acmecorp.com/vault/index.html.md): Secrets management and access control ## Docs - [Getting started](https://docs.acmecorp.com/start.html.md) - [API reference](https://docs.acmecorp.com/api/index.html.md) - [Pricing](https://acmecorp.com/pricing.html.md) ## Optional - [Blog](https://acmecorp.com/blog/index.html.md) - [Changelog](https://acmecorp.com/changelog.html.md) - [Support](https://acmecorp.com/support.html.md)
Why it matters for your brand
AI is the new search
ChatGPT, Gemini, Perplexity, and Claude now answer questions that used to go to Google. If you're not in their context, you're invisible in AI-generated results.
Models hallucinate brands
Without a clear source of truth, AI models guess. They mix up your messaging, cite outdated pricing, or confuse you with competitors — and users trust the answer.
llms.txt gives you control
A well-structured file gives AI crawlers and inference systems a canonical, up-to-date snapshot they can trust — authored directly by you.
The format
The spec is intentionally minimal. A valid llms.txt is plain Markdown — human-readable and machine-parseable — served at the root of your site. Four elements, in order:
# H1Project or company nameOne line — the name of your site, product, or organisation.
> BlockquoteShort summaryA concise description of what you do. The most important part — AI models prioritise this.
## SectionsCurated link listsNamed sections (Products, Docs, Blog…) with links to clean Markdown versions of your pages.
## OptionalLower-priority linksContent useful for context but not essential — a special section AI tools may skip when context is tight.
The spec also recommends serving clean Markdown versions of individual pages at the same URL with .md appended — e.g. yoursite.com/about.html.md. This lets AI tools fetch page content without parsing HTML.
# Acme Corp > Cloud infrastructure tools for engineering teams. > Helping 1,200+ companies ship faster with zero-downtime > deployments, real-time observability, and secrets management. ## Products - [Acme Deploy](https://acmecorp.com/deploy/index.html.md): Zero-downtime deployment pipeline for Kubernetes - [Acme Monitor](https://acmecorp.com/monitor/index.html.md): Real-time observability and alerting - [Acme Vault](https://acmecorp.com/vault/index.html.md): Secrets management and access control ## Docs - [Getting started](https://docs.acmecorp.com/start.html.md) - [API reference](https://docs.acmecorp.com/api/index.html.md) - [Pricing](https://acmecorp.com/pricing.html.md) ## Optional - [Blog](https://acmecorp.com/blog/index.html.md) - [Changelog](https://acmecorp.com/changelog.html.md) - [Support](https://acmecorp.com/support.html.md)
Two variants
llms.txtStandard
Links to your key pages as curated Markdown files. Lightweight — AI tools fetch only what they need. Best for most sites.
llms-full.txtFull
All linked content inlined into one file. No extra fetches required — the entire context is self-contained. Better for documentation-heavy sites.
Who's already using it
Since Jeremy Howard proposed the standard in September 2024, adoption has grown rapidly. Over 784 websites have implemented llms.txt, with tech leaders paving the way:
Uses it to help AI navigate its vast documentation
Provides a curated index of its research and products
Covers its platform docs and deployment guides
Ensures accurate code and API info for developers
Source: llms-text.com/directory, July 2025 · llmstxt.org
llms.txt and Generative Engine Optimisation
Creating a llms.txt file is the first, most concrete step in a GEO (Generative Engine Optimisation) strategy. Just as SEO optimised pages for Google's algorithm, GEO optimises your brand presence for AI-generated answers.
But a file alone is not enough. AI models also draw on training data, live web citations, and retrieval-augmented generation sources. A full GEO audit looks at how your brand surfaces across all of these channels — and where the gaps are.
Frequently asked questions
What is llms.txt?
llms.txt is a Markdown file placed at your site's root (/llms.txt) that gives AI models a concise, structured overview of your site and links to its most important pages.
Where do I put the llms.txt file?
At the root of your domain, for example https://yourdomain.com/llms.txt, so AI crawlers can find it automatically.
Is llms.txt the same as robots.txt?
No. robots.txt tells crawlers which URLs they may access. llms.txt gives large language models curated, plain-text context about what your site is and what matters most.
Who created llms.txt?
The standard was proposed by Jeremy Howard of Answer.AI in September 2024.
Who has adopted llms.txt?
Companies including Anthropic, Cloudflare, and Vercel publish llms.txt files for their sites.
How do I create an llms.txt file?
Use this free generator: enter your website URL and it scrapes your pages, extracts structured entities, and produces a spec-compliant llms.txt and rich llms-full.txt.
Get started
Generate your llms.txt in seconds
Enter your website URL. We scrape up to 6 pages, extract structured entities, and generate a citation-ready file — QA-scored and ready to ship.