How to Create llms.txt
A useful llms.txt file is short, clear, and focused on your most important resources.
Create llms.txt by adding a plain text file at your site root with a short site summary and links to important pages, docs, and AI-readable resources.
llms.txt helps AI agents identify authoritative website resources. It should include concise descriptions, relevant links, and optional Markdown indexes.
Check your website's AI discoverability signals.
Run a free scan for robots.txt, sitemap discovery, Link headers, Markdown readiness, and AI bot access.
Main Explanation
Start with a short description of what your website does. Then list the pages an AI agent should understand first.
Avoid turning it into a full sitemap. The goal is guidance, not volume.
Why this matters for AI search
How to Create llms.txt matters because AI systems do not only look for keywords. They need accessible pages, clear explanations, stable source URLs, and passages that answer user intent directly.
When your content is easier to crawl and easier to summarize, it may become a better source candidate for answer engines and AI assistants.
Common mistakes to avoid
- Writing long introductions before answering the actual question.
- Hiding important content behind scripts, tabs, or gated UI.
- Publishing technical files once and never maintaining them.
- Using vague headings that do not match user questions.
- Forgetting internal links to related AI visibility topics.
Practical Steps
- Create /llms.txt at the root.
- Add a site summary.
- List key pages with short labels.
- Add Markdown or docs links.
- Expose it with a Link header if possible.
Practical example
A strong AI-ready page usually starts with a direct answer, then explains the context, then lists practical steps, examples, and related resources. This makes the page useful for humans while also giving AI systems cleaner passages to extract.
For example, if a page explains an optimization concept, it should define the concept, explain why it matters, show how to test it, describe common mistakes, and link to related implementation pages.
Recommended page structure
- Start with one clear H1 that matches the topic.
- Add a Quick Answer section near the top.
- Use an AI Summary section for concise machine-readable context.
- Break instructions into short steps and examples.
- Add FAQ questions that reflect real search and AI assistant prompts.
- Link to related pages so crawlers can understand the content cluster.
FAQ
What format should I use?
Plain text or Markdown-style lists are common and easy to parse.
Can I automate it?
Yes, but review it manually so it stays useful.
Should it include private URLs?
No. Only include public resources you want agents to find.