Free llms.txt Generator
Generate a spec-compliant llms.txt file for your website in seconds. We crawl your sitemap, group pages into sections, and hand you a file ready to drop at /llms.txt.
What is llms.txt?
llms.txt is a Markdown file at the root of your site that gives AI assistants a curated, human-authored map of your most important pages. Think of it as a sitemap for LLMs β optimized for answer generation, not ranking.
Why It Matters
- AI assistants have small context windows and benefit from a curated entry point
- You control which pages LLMs prioritize and how they're summarized
- It's the fastest-adopted GEO signal β and still missing on most sites
What We Generate
- An H1 site name and one-line summary from your homepage metadata
- Sections (Docs, Product, Blog, β¦) bucketed from your sitemap URLs
- An Optional section for legal and de-prioritized pages
Frequently Asked Questions
llms.txt is a plain-text Markdown file you host at https://yourdomain.com/llms.txt. It gives AI assistants like ChatGPT, Claude, and Perplexity a curated list of the pages on your site that matter most β with short summaries LLMs can use when answering questions about you. Unlike robots.txt or sitemap.xml, it's written for humans and machines together, and lets you prioritize content without relying on a crawler to guess.
We fetch your homepage to pull the site name and meta description, then discover your sitemap (via robots.txt or /sitemap.xml), group the URLs into standard sections like Docs, Product, and Blog using URL path heuristics, and render the whole thing as Markdown. The output is capped at reasonable sizes so the file stays scannable by LLMs with small context windows.
Usually you'll want a quick human edit β tightening the summary, renaming sections, and adding a short description on the most important links. The generator gives you a strong starting point that saves about 80% of the manual work; the last 20% is worth doing by hand so the file reflects your brand voice.
Upload the contents to https://yourdomain.com/llms.txt β at the site root, not in a subdirectory. The file must be served as plain text (content-type text/plain or text/markdown) and must be reachable without authentication. Most static hosts (Vercel, Netlify, Cloudflare Pages) let you drop it in the public folder; for CMS sites, use whatever mechanism you use to serve robots.txt.
Yes. No sign-up, no credit card, no rate limit on the anonymous path. The heuristic generator is completely free. GeoVector's paid plans add LLM-written summaries, automatic re-generation when your sitemap changes, and monitoring that alerts you when your llms.txt drifts from your site.
What the generated file looks like
llms.txt is Markdown with a fixed shape: one H1 for the site, an optional blockquote summary, then H2 sections of links.
# Example Analytics
> Product analytics for teams that ship weekly. Self-serve dashboards, an event API, and a warehouse sync.
## Docs
- [Quickstart](https://example.com/docs/quickstart)
- [Event API reference](https://example.com/docs/api/events)
- [Warehouse sync](https://example.com/docs/warehouse-sync)
## Product
- [Pricing](https://example.com/pricing)
- [Dashboards](https://example.com/features/dashboards)
## Blog
- [What we learned rebuilding ingestion](https://example.com/blog/rebuilding-ingestion)
## Optional
- [Changelog](https://example.com/changelog)
- [Status](https://example.com/status)Save it as /llms.txt at the root of your domain and serve it as text/plain or text/markdown. It is a proposal, not a standard, and no crawler is obliged to read it β but it costs one file, and it is the only place you get to say which pages matter.
How pages are grouped
Sections are assigned by URL path, first match wins. Anything under a path that reads as archival β tags, authors, paginated listings β goes to Optional.
| Section | Path patterns |
|---|---|
| Docs | /docs, /documentation, /guides, /reference, /api, /developers, /sdk |
| Product | /product, /pricing, /features, /solutions, /platform, /use-cases |
| Blog | /blog, /news, /updates, /changelog, /posts |
| Articles | /articles, /resources, /insights, /research, /library, /whitepapers, /case-studies |
| Support | /support, /help, /faq, /community, /forum |
| Company | /about, /team, /careers, /press, /contact, /partners, /investors |
| Legal | /legal, /privacy, /terms, /cookies, /security, /compliance |
| Other | Anything that matches none of the above |
The generator reads up to 500 URLs from your sitemap and derives link titles from the URL path. Rewriting those titles by hand is the highest-value edit you can make to the output: the title is what a model sees before deciding whether to fetch the page.
Sources and specifications
Everything on this page is checked against the primary documentation below. If a vendor changes a rule, that is where it shows up first.
- The /llms.txt file proposal β llmstxt.org
- Sitemaps XML format (protocol 0.9) β sitemaps.org
Last reviewed . We re-check this page against the source specifications whenever a vendor updates one.