AI search fix
How to create an llms.txt file (with example)
Create llms.txt as a plain markdown file served at https://yourdomain.com/llms.txt. Follow the llmstxt.org layout: a level-1 title, a one-line blockquote summary, a short context paragraph, then sections such as Docs and Services with markdown list links to your priority URLs. Publish only after robots.txt and your CDN allow trusted AI crawlers to fetch pages — the file does not replace crawl access.
Keep the file small and factual. Link to HTML pages you want AI systems to prioritize, not PDFs behind login walls. Add an Updated line with the ISO date you last changed pricing or core routes.
Create llms.txt in five steps
- Confirm GPTBot, PerplexityBot, or other AI bots are not blocked in robots.txt or WAF.
- Draft H1 title, blockquote summary, and a two-sentence context block.
- Add sections with markdown links to docs, services, pricing, and policies.
- Upload to the site root as /llms.txt with text/plain or text/markdown.
- Request the URL in a browser and fix 404 or redirect chains.
Free · 2 minutes · no card
See what AI crawlers hit on your site
Technical blockers, missing context, weak AI-readiness signals — in one HTML report.
You'll get an HTML report showing whether /llms.txt is reachable and whether crawlers are blocked.
The llms.txt format, field by field
The llmstxt.org format is deliberately small. The first line is a level-1 markdown heading with your site or product name. Directly under it, a blockquote (a line starting with a greater-than sign) gives a one-sentence summary of what you do. A short paragraph of context can follow. After that come one or more sections, each a level-2 heading such as Docs, Guides, or Services, holding a markdown list of links to the URLs you most want AI systems to read. That is the whole grammar — no tables, no HTML, no tracking parameters.
If your documentation is large, the proposal also allows an llms-full.txt with expanded content, while llms.txt itself stays a lean index. Think of llms.txt as a table of contents a reader can scan in a few seconds, not a full export of your site.
A minimal example you can adapt
A workable file for a small SaaS might read: a heading line naming the company; a blockquote summary such as "Analytics for indie developers"; a one-line context sentence; then a Docs section linking to your getting-started guide and API reference, and a Services section linking to pricing and contact. Five to fifteen links is plenty for most sites.
Resist the urge to paste your entire sitemap. A short, curated file that points to your best pages is more useful to an assistant than hundreds of links it has to weigh. Every URL you list should return clean HTML that answers the question the link promises.
Where to put it and how to serve it
The file belongs at the domain root — https://yourdomain.com/llms.txt — the same place crawlers expect robots.txt. Serve it with a text/plain or text/markdown content type so browsers and tools display it rather than downloading it. On a static site, drop the file in your public or root output folder. On WordPress, an SEO plugin such as All in One SEO can publish and maintain it; on other stacks, a small route or a generator tool can output the same markdown.
If you can only host the file on a subdomain, note that the convention expects it on the primary domain crawlers use, so a subdomain-only copy may be overlooked unless you reference it elsewhere.
Keep it honest and current
Google documentation states you do not need llms.txt to appear in its generative AI features, and adoption across assistants is still early — so treat the file as a helpful convention, not a ranking lever. The value comes from accuracy: refresh it whenever pricing, routes, or core messaging change, and remove links that 404.
An out-of-date map is worse than none, because it can point AI systems at content that no longer matches your business. Publishing llms.txt will not, on its own, get you cited — crawl access and substantive pages still do the heavy lifting.
Related questions
Updated