llms.txt generator · generated, validated, kept current

Tell assistants where your facts live.

llms.txt is a short file at your root that points AI crawlers at the canonical, readable version of everything that matters about you. It takes an afternoon, and most sites still do not have one.

Try the generator

Generate llms.txt from any site in a few clicks.

One-time browser setup. After that, open a site, click the bookmarklet and let RankBIT build the file.

no extension required

Install the RankBIT bookmarklet

It runs in your browser on the site you are viewing, discovers pages from the sitemap when available, analyzes the site and creates the llms.txt file for you.

Drag this button to your bookmarks bar
RankBIT · Generate llms.txt
Important: do not just click the button here. Drag it to the bookmarks bar first, then run it while you are on the website you want to scan.
01

Show your bookmarks bar

In Chrome or Edge use Ctrl + Shift + B on Windows or ⌘ + Shift + B on macOS.

02

Drag the purple button above

Drop RankBIT · Generate llms.txt onto the bookmarks bar. You only need to do this once.

03

Open the website you want to process

Go to the target site's homepage, then click the RankBIT bookmark. The tool searches common sitemap locations and can fall back to internal links.

04

Let it finish and save the file

A progress panel appears in the lower-right corner. When the scan is complete, your browser automatically downloads llms.txt.

How it works

Crawl, select, write, keep current.

The hard part is choosing what belongs in it.

01
Crawl
Your sitemap is read and every page scored on factual density — prices, limits, policies, specifications.
02
Select
The pages an assistant would actually need are chosen and grouped into sections. Marketing pages are left out.
03
Write
Each entry gets a one-line summary an assistant can use to decide whether to fetch it.
04
Keep current
When your site changes, the file is regenerated and any broken link is flagged before a crawler finds it.
What you get

Six things a hand-written file usually misses.

The format is simple. Getting it useful is not.

A valid file, first time

Correct structure, ordering and length limits — the three things most hand-written files get wrong.

The right pages

Selected on factual density, not on what your marketing team is proudest of.

Markdown companions

Plain-text versions of linked pages generated where they do not exist, so every link resolves to something readable.

Uptime monitoring

If the file starts returning 404 or the wrong content type, you know within the hour.

Auto-regeneration

New pricing page, new section. The file follows your site instead of drifting from it.

Paired with robots.txt

Access rules audited alongside it — a perfect llms.txt behind a blocked crawler is worthless.

Before / after

What a good file looks like.

Short, ordered, and every line earns its place.

Common attempt
# Acme
Welcome to Acme! We are a leading
provider of innovative solutions.

// no summary line
// links to marketing pages
// 400 lines long
Generated
# Acme

> Acme runs payroll for UK agencies.

## Product
- [Pricing](/pricing.md): plans and limits
- [Integrations](/docs/integrations.md)

## Company
- [About](/about.md): founded 2021, EU-based, GDPR-aligned
Questions

About the file.

Adoption is uneven and improving. It costs almost nothing, is already fetched by several crawlers, and we treat it as one signal among nine rather than a silver bullet.

No. robots.txt controls access; llms.txt guides understanding. You need both, and we audit them together.

We generate it where a full-text bundle helps — usually documentation-heavy sites. It is optional.

A file is one fix. See the whole picture.

llms.txt is just one signal. Run the free site check to see everything keeping you out of AI answers.