llms.txt 生成器

What this file is, and what it is not

llms.txt is a plain Markdown file at the root of your site that gives an AI system a short, curated map of the pages worth reading, with one line explaining each. The idea is to remove the work of guessing your site's structure from markup.

It is worth being precise about its status: it is a proposal, not a standard. Compare it to robots.txt, which every crawler is required to obey, or sitemap.xml, which Google documents and consumes. No major answer engine has committed to reading llms.txt, and none documents doing so.

That does not make it useless — it costs almost nothing and some tools do fetch it. It does mean you should never let publishing one stand in for the two things that provably matter: being crawlable, and having content worth quoting.

• Worth doing: keeping the file accurate and short, with a note on each link.
• Not worth doing: listing every URL, or treating it as a ranking lever.

The format, in full

There are only four parts, and only the first is required: an H1 with your site name, a blockquote summary, H2 sections grouping your pages, and Markdown links with a description after each.

The description is the part people skip, and it is the part that carries the value. A bare list of URLs tells a model nothing it could not get from crawling you; one line saying what each page answers is the whole contribution.

# Acme CRM

> CRM software for teams under 20 people, priced per seat.

## Product
- [Features](https://acme.com/features): What the product does, screen by screen
- [Pricing](https://acme.com/pricing): Plans and what each includes

## Comparisons
- [Acme vs HubSpot](https://acme.com/vs/hubspot): Where each one fits

How to find out whether it is doing anything

This is the question most guides leave out, and it is the only one that matters after you publish. Check three things, in order of how much they tell you.

The first two are mechanical and take a minute. The third is the only one that shows real effect, because a file nobody fetches and a file fetched but never used look identical from outside.

• 1. It is served correctly. Visit yourdomain.com/llms.txt in a browser. You should see raw Markdown as text/plain — not an HTML page, and not a 404 because your build did not copy the file.
• 2. Every link resolves. Each URL in the file should return 200. A dead link is worse than an omission: the model gets an incomplete map and no indication that it is incomplete.
• 3. Your brand appears when asked. Ask an engine a real buyer question in your category and see whether you are named. That is the outcome the file is supposed to influence, and it is measurable — the free scan runs exactly this test across configured engines.

Questions people ask after using this

Does any AI engine actually read llms.txt?

Not reliably, and this is the part most guides skip. As of late 2026 no major answer engine has committed to consuming llms.txt as a ranking signal, and none documents it. It is a proposal, not a standard like robots.txt. Treat it as a low-cost signal that may help, never as a guarantee — and do not let publishing it stand in for the things that definitely matter, which are being crawlable and having content worth citing.

How can I tell whether my llms.txt is doing anything?

Check the three things in order. First, that the file is served at /llms.txt as text/plain and not wrapped in HTML. Second, that every URL inside it returns a 200. Third, and the only one that shows real effect, whether the engines name your brand when asked a buyer question — because a file that is never fetched and a file that is fetched but never used look identical from the outside. That last check is what the free scan measures.

Should I list every page or just the important ones?

Just the important ones. The point of the file is to remove noise, so a version that mirrors your sitemap adds nothing a crawler could not already find and makes the useful pages harder to spot. Ten to thirty curated pages with a note each beats three hundred bare URLs.

What is the difference between llms.txt and llms-full.txt?

llms.txt is a curated index of links with one-line descriptions. llms-full.txt contains the full text of your key pages, so a model can read the substance in one fetch rather than following links. The index is cheap to maintain; the full file is expensive and usually only worth it for documentation that changes often.

Can llms.txt get my site into AI answers on its own?

No. It can only affect what a model reads if that model already decides to fetch it. Appearing in an answer depends on being crawled, being ranked or retrieved for that question, and having content the model considers worth quoting. llms.txt influences the third at best.

Will publishing llms.txt hurt my Google ranking?

No. Google does not use it for search ranking, and a file at /llms.txt has no effect on Googlebot. It is also not a replacement for robots.txt or a sitemap — those do different jobs and are still required.