Free tool · No signup · Built in Singapore

Generate a clean, handwritten-style llms.txt from any sitemap.

Paste your sitemap URL. We crawl it, group your pages by section, pull the real titles & meta descriptions, and hand you back a tidy llms.txt file ready to upload to your site root. No placeholder text, no junk, no signup. Most SG SMEs have nothing here yet — get the file in 60 seconds.

Spec: llmstxt.org Output: downloadable llms.txt Cost: $0 forever
Avg generation time
~60s
Most SMB sitemaps under 200 URLs
SG SMEs with llms.txt
<3%
First-mover discovery upside
Output size
2–8 KB
Within the spec's recommended budget
Beats RankMath auto-gen
Real titles, not placeholders
The Generator

Paste a sitemap. Hit generate. Download your file.

Works with sitemap.xml, sitemap-index.xml, and most CMS-generated sitemaps (WordPress, RankMath, Yoast, Astro, Next.js, Webflow). Up to 250 URLs per run.

Your site

Drop the sitemap link, optionally name your site, then generate.

llms.txt
Your generated llms.txt will appear here.
Why bother?

What an llms.txt actually does for you.

llms.txt is a public file at your site root that gives ChatGPT, Claude, Perplexity, Gemini and other AI engines a curated, structured map of your most important content. Think of it as a hand-written sitemap, written for LLMs, not for search robots.

1You decide what AI sees first

Editorial control over your AI footprint. Sitemap.xml dumps everything; llms.txt lets you put your top 20 pages first, group them by intent, and add a one-line description per URL. AI engines crawling your site read this before noisy nav junk.

2Cited content beats ranked content

Citations are the new clicks. When ChatGPT or Perplexity cites you in an answer, it pulls a real description — usually the closest piece of structured prose to the URL. A good llms.txt often becomes that prose.

3It's a 2-minute moat

Almost nobody in SG has one. We checked the top 100 SME sites in Singapore — fewer than 3% have a real llms.txt. The few that auto-generated theirs (via RankMath etc.) ship with placeholder titles that don't match real page content.

Why not just use RankMath?

Auto-generated llms.txt vs handwritten-style.

RankMath, Yoast and other plugins auto-generate llms.txt at the click of a button. The output looks busy — but the substance is mostly placeholder. Here's the difference, side by side.

RankMath / typical plugin auto-gen This generator
Page titles ×Slug-based or "Untitled" placeholders Real <title> from each page
Descriptions ×Site-wide tagline pasted under every URL Real meta description per page
Structure ×Flat dump of every URL in sitemap Grouped by section (Services, Blog, Resources, etc.)
Noise filter ×Includes pagination, tags, author archives, attachments Strips noise URLs (paginated, attachment, feed, etc.)
Spec compliance ×Often missing the H1 + blockquote header Follows the llmstxt.org spec exactly
Editable output ×Regenerated on every plugin update Download once, edit by hand, host yourself
Cost Bundled with paid SEO plugin Free, no signup
Methodology

How the generator actually works.

No black box. Here's exactly what happens between paste and download.

1 · Sitemap fetch

We fetch the sitemap URL via a public CORS proxy (browsers block direct cross-origin XML reads). If the file is a sitemap-index.xml, we recursively fetch up to 5 nested sitemaps and dedupe URLs.

2 · Noise filter

We strip URLs that almost never belong in llms.txt:

  • Pagination (/page/2/)
  • Tag, category & author archives
  • RSS & comment feeds
  • .pdf, .jpg, .zip attachments
  • WordPress wp-content media URLs

3 · Section grouping

URLs are grouped by their first path segment: /blog/*, /services/*, /about/*, etc. Top-level pages (homepage, contact) sit in a "Core pages" section. Sections are ordered by URL count.

4 · Optional enrichment

If you tick "Enrich with real titles & descriptions", we fetch each page through r.jina.ai (a free reader that returns clean markdown) and extract the H1 + first meaningful sentence. Capped at 250 pages, 6 parallel requests.

5 · llms.txt assembly

Output follows the llmstxt.org spec: # Site name as H1, italic blockquote summary, then ## Section headers each containing markdown bullets in [Title](URL): description format.

6 · Limits & privacy

Everything runs in your browser. We don't log URLs, store sitemaps, or track who used the tool. Hard caps: 250 URLs per run, 5 nested sitemap files, 90s total timeout. If you need more, drop us a line — we run a paid version for enterprise sites.

Common questions

About llms.txt & this tool.

Where do I upload the generated file?
Drop it at the root of your domain so it's reachable at https://yoursite.com/llms.txt. On WordPress, that's the same folder as robots.txt. On Astro / Next.js, it goes in your public/ folder.
Will AI engines actually read it?
ChatGPT, Claude, Perplexity and Anthropic's web fetchers all check for llms.txt when crawling a site. Google's AI Overview doesn't yet have a public commitment but is expected to follow. Either way, it's a no-cost upside — the spec is small and harmless if ignored.
Why does it ask permission to fetch my pages?
It doesn't — your browser fetches them via the public r.jina.ai reader so we never touch your server directly with credentials. If your pages are behind login, those URLs simply won't enrich and we'll fall back to the URL-only entry.
What if my sitemap has thousands of URLs?
We hard-cap at 250 URLs to keep the file under the spec's recommended size budget. If your sitemap is larger, we sample the most-likely-important URLs (shortest paths first, then alphabetical). Want full coverage? Get in touch.
What's an "llms-full.txt"?
A second file you can publish alongside llms.txt that includes the full text of your most important pages. We don't generate that yet — it's on the roadmap.
Can I edit the output?
Absolutely encouraged. Hand-editing the descriptions, reordering sections, deleting low-value pages — that's what makes the output "handwritten-style" in the first place. Treat the generator as a 90% draft, not a final.

Want a full AI visibility audit, not just a file?

Generating an llms.txt is step one. The full picture — which AI engines cite you, which competitors are pulling ahead, where the citation gaps are — takes deeper work. We do that for free for SG-based businesses.

Request my free audit → Read our analyses