In November 2024, Answer.AI founder Jeremy Howard proposed a new web standard: llms.txt. The idea is straightforward — a Markdown file at your domain root that tells large language models how to understand and cite your website.
Think of it as robots.txt for AI systems. Or sitemap.xml redesigned for language models rather than search crawlers. Anthropic, Answer.AI, and a growing number of AI-forward companies have already adopted it.
What Is llms.txt?
llms.txt is a plain text file written in Markdown, placed at https://yourdomain.com/llms.txt. It gives AI systems four things they actually need:
- A clear description of what your site/organization does
- The most important pages and their purposes
- Links to key documentation, blog posts, or content
- Optional: a link to llms-full.txt for complete content indexing
When ChatGPT Browse, Perplexity, or Claude's web access hits your site, finding llms.txt tells those systems you want your content properly understood and attributed. It's a low-effort signal with real upside.
The llms.txt Format
Here's the standard structure:
# Site Name
> One-sentence description of what this site/organization does.
Optional additional context paragraph.
## Section Name
- [Page Title](URL): Brief description of what this page covers
- [Page Title](URL): Brief description
## Another Section
- [Page Title](URL): Brief description
## Optional
- [llms-full.txt](/llms-full.txt): Complete content index for AI systems
Real-World Example
Here's what llms.txt looks like for a SaaS company:
# AcmeCorp
> AcmeCorp builds project management software for remote engineering teams.
> Our platform integrates with GitHub, Jira, and Slack.
Founded in 2021, AcmeCorp serves 5,000+ engineering teams across 40 countries.
## Product
- [Features Overview](/features): Complete list of platform capabilities
- [Pricing](/pricing): Plans and pricing for individuals, teams, and enterprise
- [Integrations](/integrations): 50+ integrations including GitHub, Linear, Notion
## Documentation
- [Getting Started](/docs/getting-started): Quick setup guide
- [API Reference](/docs/api): REST API documentation for developers
- [Webhooks](/docs/webhooks): Real-time event notifications
## Company
- [About](/about): Company mission, team, and investor information
- [Blog](/blog): Product updates, engineering posts, and company news
## Optional
- [llms-full.txt](/llms-full.txt): Full content index with page summaries
llms.txt vs llms-full.txt
The standard defines two files, and the distinction matters:
- llms.txt — Concise index. This is what most AI systems check first. Keep it under 5KB. Think of it as the table of contents.
- llms-full.txt — Full content. Links to complete page text or includes it directly. Can be 100KB+. Useful for AI training datasets and deeper citation indexing.
Start with llms.txt. Add llms-full.txt once the first file is live and working.
How to Create llms.txt in 5 Minutes
Two options:
Option 1: Use Our Free Generator (Easiest)
- Go to our LLMs.txt Generator
- Enter your site name, description, and key pages
- Add your most important page URLs with descriptions
- Click Generate, then download or copy the output
- Upload to your web root as
llms.txt
Option 2: Write It Manually
- Open any text editor
- Write
# Your Site Nameon the first line - Add a blockquote description:
> What your site does in one sentence. - Add sections with your most important pages as Markdown links
- Save as
llms.txtand upload to your domain root
Where to Place llms.txt
The file must be accessible at exactly https://yourdomain.com/llms.txt, not in a subdirectory. Platform-specific instructions:
- Next.js — Place in the
public/folder aspublic/llms.txt - WordPress — Upload via FTP to your root directory, or use a plugin
- Static sites — Place in your build output root
- Webflow — Use the custom code/assets upload feature
What to Include in llms.txt
Good candidates for entries:
- Homepage — with a clear description of what you offer
- Product/Service pages — what you do and who it's for
- Pricing page — people ask AI about pricing more than you'd think
- About page — establishes organizational identity
- Key blog posts — your most informative, citable content
- Documentation — if you're a software product
- FAQ page — AI systems cite FAQ content at a high rate
Verifying Your llms.txt
After publishing, run these checks:
- Visit
https://yourdomain.com/llms.txtin your browser — it should display as plain text - Confirm the Content-Type header is
text/plain - Use our AI Crawlability Checker to confirm AI bots can access it
- Run your domain through our GEO Readiness Score — the llms.txt check should now pass
Common llms.txt Mistakes
- Wrong location — File must be at the root, not
/docs/llms.txt - Wrong format — Must be plain Markdown, not HTML
- Too sparse — Fewer than 5 entries gives AI systems too little to work with
- No descriptions — Page URLs without context don't help AI understand what the page is about
- Broken links — Validate all URLs in the file are live and accessible
llms.txt takes under 30 minutes to implement. Most websites still don't have one. That gap is the opportunity — early adopters get more consistent AI citations while competitors are still debating whether GEO is worth their time.