LLMs.txt Generator.
Enter a domain. We read its sitemap, fetch the real title and meta description of every page, sort them into sections, and hand back a curated llms.txt you can publish at your site root. No sign-up, no page limit per account.
The generator
> blockquote under the title — the most-read line in the file.
Upload the finished file so it is served at https://yourdomain.com/llms.txt with a text/plain content type. Keep it curated — this file is meant to be a short reading list, not a second sitemap.
Scan a site to generate your llms.txt.
What llms.txt actually is.
A short, curated Markdown file at your domain root that points language models at the pages worth reading. It is a reading list, not a permission file.
It is not robots.txt
robots.txt answers "who is allowed to crawl what". llms.txt answers "of everything you are allowed to read, what should you read first". They solve different problems and you want both. Blocking GPTBot in robots.txt while publishing an llms.txt is a contradiction worth catching before you ship.
Shorter is better
The value is in the curation. A 600-line llms.txt that mirrors your sitemap tells a model nothing it could not already work out. Twenty well-chosen pages with honest one-line descriptions is a genuinely useful artefact — which is why this tool pre-sorts archives, carts and tag pages out of the way.
Nobody has confirmed it as a signal
No search engine or model provider has said llms.txt influences retrieval or ranking. Adoption is growing; proof is not in yet. Publish it because it takes ten minutes and documents your own site properly — and be sceptical of anyone selling it as a ranking factor.
The format
An H1 with the site name, a > blockquote summary, optional context paragraphs, then H2 sections of Markdown links with a colon-separated description each. A section called Optional marks links a model may skip when it is short on context.
Common questions.
Where exactly does the file go?
At the domain root, served as plain text: https://yourdomain.com/llms.txt. On most static hosts that means dropping it in your public/ or static/ directory. On WordPress, upload it to the web root rather than the theme folder.
What about llms-full.txt?
Some sites publish a second file containing the full Markdown content of their key pages, so a model can read everything in one request. It is optional and much heavier to maintain. Get llms.txt right first.
Why did the scan miss pages?
Usually one of three reasons: the sitemap is incomplete, the site paginates its sitemap beyond the twelve files we read, or there is no sitemap at all and we fell back to homepage links. The tool tells you which happened. It also caps at 400 URLs — well past the point where an llms.txt should have been trimmed anyway.
Some titles came back empty.
That page either blocked our fetch, timed out, or renders its title only in JavaScript. We fall back to a label derived from the URL slug, which you can edit in the downloaded file. A page whose title only exists after JavaScript runs is worth fixing regardless of this tool.
Do you keep the URLs I scan?
No. The scan runs in a stateless serverless function and nothing is written to a database. The generated file exists only in your browser until you download it.
A file is a start. Being cited is the goal.
Publishing llms.txt makes your pages easier to read. Getting quoted in AI answers takes brand mentions on sources those models already trust — which is what our AEO packages are built to produce.