Free tool
llms.txt generator, plus the part everyone skips.
Most generators crawl your site and hand back a list of your pages. That tells an AI what exists, not who you are. This one looks for the facts an answer engine actually needs — who owns the site, what you sell, where you work, what it costs — and tells you which ones are missing.
Free. No signup, no email, nothing stored.
What this actually checks
No black box. The analyser fetches your homepage and up to twelve linked pages, reads them once in memory, and looks for the specific facts an answer engine needs before it will describe a business. Here is the whole list.
- Organization identity
- Looks for an Organization (or subtype) node in your homepage JSON-LD and reads its name. This is weighted most heavily of anything checked: if a machine cannot resolve who owns the site, nothing else it finds is attributable to you.
- Business description
- Takes the organization's description from structured data, falling back to your meta description. A site whose only self-description is a tagline gives an engine nothing concrete to repeat.
- Pricing signal
- Scans the fetched pages for currency-shaped values ($, €, £, or an amount followed by CAD/USD/EUR/GBP). Deliberately conservative — a false positive here would tell you your pricing is discoverable when it is not.
- Service area
- Reads areaServed from your organization markup. Without it, an assistant asked for someone “near me” has no basis to include you.
- Contact route
- Looks for a mailto: link or a recognisable contact path, including the French /contactez and /nous-joindre forms.
- Founder or author entity
- Finds a founder or Person entity. It follows @id references to nodes declared on other pages rather than marking them absent — referencing a shared entity by @id is the correct way to model this, and penalising it would mark down the sites doing it properly.
- External corroboration and declared languages
- Collects sameAs profile links, plus the languages your site declares through its html lang attribute and any hreflang annotations.
- Existing llms.txt
- Checks whether you already publish one, so you find out before writing a second.
What it cannot tell you
- It reads your homepage and up to twelve linked pages. A large site will have relevant facts it never sees, so treat a low score as a prompt to look, not a verdict.
- It reads what your site states, not what is true. A confident, well-marked-up page containing wrong facts scores well.
- It cannot tell you whether any AI system will actually cite you. Nothing can — that is the honest state of this field.
- Structured data is read from the homepage only. Organization markup that appears exclusively on an inner page will not be found.
How it handles your site
Requests are rate-limited and the fetcher validates the addresses a hostname resolves to before it connects, so the endpoint cannot be pointed at internal or private network ranges. Pages are analysed in memory and discarded when the response is sent — nothing is written to a database, and no email is required.
Putting the file on your platform
Generating the file and serving it are different problems, and the second one is platform-specific in ways most guides skip. These are checked against each platform's own documentation and carry the date they were last verified.
Questions
What is llms.txt?
A plain-text file at the root of your website that tells AI systems what your site is and which pages matter. It is to answer engines roughly what robots.txt is to crawlers: a short, machine-readable statement you write yourself, instead of leaving the summary to be inferred.
Where do I put the file?
At the root of your domain, so it resolves at https://yoursite.com/llms.txt. It must be served as plain text.
Does llms.txt actually do anything yet?
It is a proposed convention, not a standard any engine is obliged to honour, and adoption is early. We say that plainly because most tools do not. The reason to publish one is that it costs almost nothing and it makes you decide what your own summary should say — and that exercise is worth more than the file.
Why does this tool show warnings instead of just generating a file?
Because the file is the easy part. If your site has no organization markup, no stated pricing and no declared service area, an answer engine still cannot describe you — with or without llms.txt. The warnings are the part that changes whether you get recommended.
Do you store my site's content?
No. Pages are fetched, analysed in memory, and discarded when the response is sent. Nothing is written to a database and no email is required.
Built and maintained by Jaeden Doody at StillAwake Media, a Montréal studio that does answer engine optimization and technical SEO. The same checks run against this site — see our own llms.txt.
More free tools and platform guides: StillAwake tools.