llms.txt Generator & Validator — Free, Spec-Aligned
Turn your existing website into a clean llms.txt, or check the one you already publish against llmstxt.org v2. Every finding names its rule and line, and every deduction is shown.
Enter a domain to generate. Switch tabs to validate a live file or pasted text. Results open on a shareable page with a downloadable report.
Checked against llmstxt.org v2 (August 10, 2026)
What an llms.txt file is
llms.txt is a small markdown file at the root of a website, proposed at llmstxt.org. It states what the site is in an H1 and a short blockquote, then lists the pages worth reading under H2 sections, one link per line with an optional description.
A web page is built for people: navigation, scripts, banners. An AI agent that reads it has to guess what matters. The file removes the guessing for any agent that chooses to look. It does not force any system to look, and it makes no promise about traffic, rankings or citations.
The only required element is the H1. The blockquote and the H2 sections are recommended. Version 2 of the proposal also allows a file at a subpath, such as /docs/llms.txt, covering the pages under it.
# Example Co > Example Co makes invoicing software for freelancers. > Docs, pricing and policies are linked below. ## Docs - [Getting started](https://example.com/docs/start): Install and first invoice - [API reference](https://example.com/docs/api): Endpoints and auth ## Product - [Pricing](https://example.com/pricing): Plans and what each includes ## Optional - [Privacy](https://example.com/privacy) - [Terms](https://example.com/terms)
The generator writes this shape from your own titles and descriptions. The validator checks it line by line.
How the llms.txt generator works
Built from what your site already publishes. No AI model writes a summary of your company, so nothing is invented and the same site always produces the same file.
robots.txt first, and we skip whatever it disallows for our crawler. Then your sitemap, or the links on your home page when there is none. Localised sites keep one language.
Home, docs, API, product, pricing, blog and about pages, up to 60 pages at 2 requests per second. Family caps stop one large blog from crowding out the documentation.
H1 site name, blockquote from your home page description, H2 sections of - [title](url): description lines, an Optional section for legal pages. The result is validated immediately and graded.
Path scope. Enter /docs in the scope field to keep only pages under that path; the file is then meant for /docs/llms.txt, the subpath form that llmstxt.org v2 allows. Where a page has a markdown version, the generator links that instead of the HTML page.
What to edit afterwards. The blockquote is your meta description; rewrite it for a reader who has never heard of you. Descriptions cut at 160 characters end with an ellipsis; finish the sentence. Then validate the published file.
How the llms.txt validator grades your file
Two verdicts, never blended into one number. The report lists every finding with its rule id and line, the deduction it caused, and the arithmetic to the final score. 16 rules in total.
| Verdict | Rules | Question it answers | Outcome |
|---|---|---|---|
| Spec validity | E1, E2, E3, E4, E5 | Is this a valid llms.txt at all? | Valid or invalid. Any hit makes the file invalid; the quality score is untouched. |
| Quality | W1, W2, W3, W5, W6 | How useful is the file to an agent that reads it? | 0–100, starting at 100, per-hit deductions with caps. A ≥ 90 · B ≥ 75 · C ≥ 60 · D ≥ 40. |
| Information | I1, I2, I3, I4, I5, I6 | What else is worth knowing? | Reported only. Changes neither verdict. |
A missing file, or an HTML page served in its place, fails spec validity and leaves quality not assessable. Blocking an AI crawler in robots.txt is recorded as information, never scored, because it may be a deliberate choice. None of the quality rules is required by the spec, and the report says so on each one.
See all 16 rules and deductions. Checked against llmstxt.org v2 (August 10, 2026).
What llms.txt is, and what Studio does.
Transcript
When someone asks ChatGPT, Claude or Perplexity about a business, the assistant often goes and reads the website. But a web page is built for people — navigation, menus, scripts, ads. The useful part is buried, and the agent has to guess what matters.
llms.txt is a small text file at the root of your site. It's a proposal from llmstxt.org — a plain Markdown map: what your site is, and links to the pages that explain it best. Documentation, product, pricing, policies. Not every page — the ones that matter.
Let's be clear about what it doesn't do. It doesn't guarantee rankings, citations or traffic. Some AI systems read it, some don't. What it does is remove the guessing for any agent that chooses to look.
llms.txt Studio builds the file from what your site already publishes — your sitemap, your titles, your descriptions. No AI model invents a summary of your company. It's free, and you don't need an account.
Already have one? Validate it. You get two answers: is it a valid llms.txt under the current spec, and how useful is it — a grade from 0 to 100, with every deduction shown.
Then there's the part that matters after launch. Sites change. Pages move. Links break. And the file just sits there, quietly going stale. Monitor checks it every week. If the grade drops, a link dies, or the file drifts from your site, you get an email with the exact lines that changed — and one click to regenerate. It also shows you when a documented AI crawler fetched your file, with the timestamp.
Generate one now, free, at llmstxt.echorank360.com. Keep it correct with Monitor. That's the whole product: a correct, current file, and evidence when something touches it.
What is free, and what Monitor adds
Generating and validating stay free. Monitor is for domains you own and want to keep correct without checking by hand: websites change, and the file goes stale quietly.
| Feature | Free | Monitor |
|---|---|---|
| Generate llms.txt from your own pages | Yes, once a day | Yes, plus one-click regenerate per verified domain |
| Validate a live file or pasted text | Yes, once a day | Yes |
| Shareable result page and PDF report | Yes | Yes |
| Weekly presence, validity and grade check | — | Yes |
| Change alert with a line diff | — | Yes |
| Drift: dead links, sections your site has that the file lacks | — | Yes |
| Canary fetch log with crawler user-agent labels | — | Yes |
| Account required | No | Yes, plus domain ownership (DNS TXT or meta tag) |
| Price | $0 | $12/month or $99/year per domain |
llms.txt Studio makes your file correct, current and easy to verify. Whether any AI system reads, uses or cites it is that system's decision — we do not guarantee AI inclusion, citations, rankings or traffic.
Where the file lives on your platform
Where to put the file, how it comes out as text rather than an HTML page, and the one thing that usually goes wrong on each platform.
A file in public/ is served at the site root; nothing else is needed unless middleware or i18n rewrites get in the way.
A plain file uploaded to the web root is served before WordPress routing runs; do not create it as a Page.
Webflow does not host arbitrary files at the site root, so /llms.txt needs a redirect or a proxy in front of the site.
Anything in static/ is copied to the published root unchanged; that is where llms.txt goes.
Files in static/ are copied to the build root and respect baseUrl, which is the thing to watch.
All platform guides · The format itself, in depth, in the knowledge base.
Questions about llms.txt, answered plainly
Will an llms.txt file improve my rankings or get my site cited by AI assistants?
Nobody can promise that, and we do not. llms.txt is a proposal from llmstxt.org that some AI systems read and others ignore. What the file does is give any agent that chooses to look a clean, plain-text map of the pages that matter. llms.txt Studio makes your file correct, current and easy to verify. Whether any AI system reads, uses or cites it is that system's decision — we do not guarantee AI inclusion, citations, rankings or traffic.
What does the generator use to build the file?
Your robots.txt, your sitemap (or the links on your home page when there is no sitemap), and up to 60 of your own pages: their titles, meta descriptions and headings. No AI model writes or rewrites anything. The same site produces the same file every time.
Where does llms.txt go, and how must it be served?
At the root of your site, as https://example.com/llms.txt, served as plain text or markdown rather than an HTML page. llmstxt.org v2 also allows a file at a subpath such as /docs/llms.txt that covers only the pages under it; the generator's path scope builds exactly that. Platform specifics are in the guides for Next.js, WordPress, Webflow, Hugo and Docusaurus.
What is the difference between spec validity and the quality grade?
Two separate verdicts, never blended. Spec validity asks whether the file is a valid llms.txt at all: it exists, is served as text, has an H1, its file-list lines are well formed, and it is not an HTML page. Quality is Studio's own set of recommendations, scored from 100 down with each deduction and its cap printed on the report. 16 rules in total; none of the quality rules is required by the spec.
Do I also need an llms-full.txt?
No. llms-full.txt is not part of the llmstxt.org proposal; it is a convention some sites use to inline page content into one large file. The generator can build one for you (tick the option), but the validator only ever judges llms.txt.
How much can I use the free tools?
1 generate and 1 validate per day per network address, resetting at 00:00 UTC, with no account. Monitor removes the limit for regeneration on domains you own and have verified.
What does Monitor's crawler fetch log actually prove?
That a request carrying a given User-Agent string followed a canary link that exists only in your llms.txt, with the timestamp. User-Agent strings can be spoofed, so it does not prove the company behind that name made the request, and a log with no fetches does not prove that any AI system ignored your file.
Is anything I paste or generate published?
No. Each result lives at an unlisted link (/r/<id>) that search engines are told not to index, and the free tools need no account. Nothing you paste into the validator is stored anywhere public.