llms.txt Builder.
Curate Your AI Map.

Yes — completely. No signup, no email gate, no account required. The entire file is generated inside your browser. Nothing you type is transmitted to any server.

The format is plain text with Markdown-style headings. It starts with a comment block (# lines), followed by a ## Purpose section and a ## Key URLs section where each entry is a labelled Markdown list item: "- Label: https://url — optional description". Comments and blank lines are fine. Keep it readable and avoid HTML.

At the root of your domain — always at https://yourdomain.com/llms.txt. Never inside a subfolder such as /files/llms.txt or /public/llms.txt (that is the folder structure, not the URL). On Astro, Next.js, Netlify and Cloudflare Pages the file belongs in the public/ folder of your repository.

Some AI systems that power ChatGPT browsing, Perplexity, and other retrieval-augmented tools check for llms.txt to find authoritative pages faster. Adoption is still growing, but early implementation costs nothing and positions you ahead of the curve.

llms-full.txt is an optional extended variant for documentation-heavy sites. While llms.txt lists your top-level entry points (10–20 URLs), llms-full.txt can contain hundreds of deeper references like API endpoints or individual guide pages. For most sites llms.txt alone is sufficient.

Start with 10–20 curated URLs. Adding 50+ links dilutes the signal and the file starts looking like a sitemap, which is not its purpose. If you run a large documentation site, keep llms.txt tight and point to llms-full.txt in the Notes section.

No. robots.txt controls crawler access (allow / disallow). llms.txt is purely advisory — it suggests which pages are most authoritative for AI context but has no enforcement mechanism. You still need both files for different purposes. Use our free robots.txt Generator to control AI training bots like GPTBot and Google-Extended.

Avoid ephemeral or low-signal pages: individual blog posts (unless they are landmark resources), campaign landing pages, login or checkout pages, pagination pages, and URL variants with tracking parameters. Focus on stable, evergreen content.

You can reference it in the Notes section, but that is not the primary purpose of llms.txt. A sitemap says "here is everything"; llms.txt says "start here". They serve different audiences — sitemaps are for search engine crawlers, llms.txt is for AI context systems.

Whenever your site structure changes significantly: new key sections added, major pages removed, or URLs changed. For most sites a quarterly review is more than sufficient. Do not update it for every blog post or minor content change.

Generative Engine Optimization (GEO) is the practice of making your brand and content discoverable in AI-generated answers from engines like ChatGPT, Perplexity, Gemini and Claude. llms.txt is one tactical element of a broader GEO strategy — it helps AI engines navigate your site, but the bigger picture includes content quality, structured data, brand authority, and consistent presence across the web.

No. It is a passive, advisory text file. The worst outcome is that AI systems ignore it. There is no known mechanism through which it could negatively affect search rankings or AI visibility. Adding it is low-risk and potentially useful.

Each domain and subdomain needs its own file at its own root. If you run docs.example.com and www.example.com separately, each needs its own llms.txt. You can cross-reference them in the Notes section if helpful.