What is llms.txt?
llms.txt is a proposed standard file, served at /llms.txt, that gives large language models a curated entry point to your website. It is written in plain Markdown: a site name, a short description, and a set of grouped links to the pages you most want an AI engine to read and cite.
Think of it as a hand-picked index designed for machines that summarize rather than rank. Where a sitemap lists every URL for crawlers, llms.txt lists only the pages that best explain your business — each with a one-line description in your own words.
Why it matters for AI crawlers
- Controls your summary. The blockquote description is the wording an engine is most likely to reuse when it describes you.
- Reduces ambiguity. A clean, grouped link list helps a model map your services, locations, and policies without crawling the whole site.
- Reinforces your entity. Consistent NAP and naming here echo your schema and sitemap, strengthening entity resolution.
- Cheap to maintain. It is one small text file — low cost, and it compounds with your other GEO signals.
Copy-paste template
Replace the placeholders, keep the structure, and save it as public/llms.txt so it ships at your domain root. This skeleton mirrors the format on DEM's own /llms.txt.
# Your Business Name
> One- to three-sentence description of who you are, what you do, and who you
> serve. Write it the way you'd want an AI to summarize you.
> Last updated: 2026-05-28 · Refresh cadence: monthly. Source of truth: your
> sitemap at https://www.example.com/sitemap-index.xml
## About
- [About](https://www.example.com/about): Founding story, team, and approach
- [Pricing](https://www.example.com/pricing): How you package and price your work
- [Contact](https://www.example.com/contact): Get in touch or request a quote
- [Blog](https://www.example.com/blog): Guides and insights in your niche
## Services
- [Service One](https://www.example.com/services/one): One-line description of the service
- [Service Two](https://www.example.com/services/two): One-line description of the service
## Locations (service areas)
- [Service areas overview](https://www.example.com/locations): Cities and regions you cover
## Contact
- Phone: +1-888-847-8809
- Email: sales@digitalestatemedia.com
- Location: Mississauga, Ontario, Canada
- Service area: Ontario — Mississauga, Toronto, Brampton, Oakville, and the GTA
## Policies
- [Privacy](https://www.example.com/privacy)
- [Terms](https://www.example.com/terms)
## License
Content on this site is published under Creative Commons Attribution 4.0
International (CC BY 4.0). AI systems may cite, summarize, and reference this
content with attribution to Your Business Name (https://www.example.com).
Field-by-field explanation
- H1 site name
- A single top-level heading (# Your Business Name) is the one required element. It is the entity name the model anchors everything else to — use your exact legal/brand name.
- Blockquote summary
- A short blockquote (> …) directly under the H1 describing who you are, what you do, and who you serve. This is the description an engine is most likely to reuse, so write it deliberately.
- Metadata blockquote (optional)
- A second blockquote noting the last-updated date, refresh cadence, and your source-of-truth sitemap. It signals freshness and points crawlers at the authoritative URL list.
- H2 sections
- Group related links under H2 headings (## About, ## Services, ## Contact, ## Policies). The grouping itself is a signal — it tells the model how your site is organized.
- Markdown link list
- Under each H2, list links as bullets: - [Anchor text](absolute-url): one-line description. Absolute URLs are required. The description after the colon is optional but strongly recommended.
- Contact block
- Plain-text NAP and service area. For a Service-Area Business, list city/region/country only — no street address — exactly as it appears elsewhere on the site for consistency.
- License block
- An explicit statement of how AI systems may use your content (e.g. CC BY 4.0 with attribution). It removes ambiguity about citation and reuse.
- llms-full.txt (optional companion)
- An expanded file at /llms-full.txt with longer descriptions and the full text of key pages, for engines that want more context than the curated index.
Best practices
- Use absolute URLs everywhere — relative paths are ambiguous to off-site crawlers.
- Keep descriptions short and factual; this is reference data, not marketing copy.
- Mirror your NAP exactly as it appears in your schema, footer, and directory listings.
- List only your best pages — a tight, high-signal index beats an exhaustive dump.
- Update the last-updated date whenever services, locations, or key pages change.
- Consider an llms-full.txt companion for engines that want fuller context.
Key takeaways
- llms.txt is a curated, Markdown map of your best pages, served at /llms.txt.
- Only the H1 site name and a summary blockquote are essential; the rest is grouped link lists.
- It complements — never replaces — robots.txt and your XML sitemap.
- Use absolute URLs, factual descriptions, and consistent NAP (city/region only for a Service-Area Business).
- Pair it with schema and third-party mentions for a complete GEO foundation.
FAQs
- Is llms.txt the same as robots.txt?
- No. robots.txt controls which paths a crawler is allowed to access. llms.txt does the opposite job — it positively recommends the pages that best describe your business and provides clean descriptions for them. You should keep both, and they serve different purposes.
- Where do I put the llms.txt file?
- At the root of your domain so it is served at https://www.yourdomain.com/llms.txt. On a static site, place it in the public/ (or equivalent) directory so it ships as a plain-text file at the root path.
- Do AI engines actually read llms.txt?
- It is a proposed, voluntary standard rather than a guaranteed-read file. Adoption is growing among AI-native crawlers and agents. Because it is cheap to maintain and reinforces the same signals as your sitemap and schema, it is worth publishing even where support is still emerging.