llms.txt Generator
A free browser-based generator and validator for llms.txt — the proposed convention that tells AI systems what a site contains.
listed free · dofollow link · not sponsored
The field no other directory carries
What these three numbers are, exactly. One query, six runs, one model, page text shown to it. The denominator is the number of runs we could actually read for this query — not the total number of runs anywhere. Two of the thirty runs in the study failed to parse and are excluded from both sides of the fraction, here and on every other listing. A page that is cited 6 of 6 is not promised to stay there.
The same eight checks you can run yourself
Read the two amber rows again. Our own check tells this page it is too short and that it has nothing question-shaped to lift — and it was still cited every single time, beating eight sources that ranked above it. That is the honest state of the evidence: these checks are hygiene, not a score, and we do not sell you one.
What this listing ships to machines
{
"@context": "https://schema.org",
"@type": "SoftwareApplication",
"name": "llms.txt Generator",
"url": "https://llmstxtgenerator.dev/",
"additionalProperty": [
{ "@type": "PropertyValue",
"propertyID": "ai-citation-runs",
"name": "AI citation runs",
"value": 6,
"unitText": "of 6 runs" },
{ "@type": "PropertyValue",
"propertyID": "ai-citation-rate",
"name": "AI citation rate",
"value": 1.0,
"unitText": "ratio" }
]
}
The numbers are named, typed and dated, so a machine is told what they are rather than left to guess. This is the part no other directory does — and the two fields above are a valid snippet you can lift and adapt for your own page today.
What this listing does not claim
One query, not a market. Everything above was measured on a single query, llms.txt generator. It says nothing about how this site performs on any other query, in any other language, in any other country.
Six runs is a small sample. 6 of 6 means six citations out of six attempts. It is a direction, not a guarantee. The same pool of ten produced two sources cited 4 of 6 and six sources cited 0 of 6.
The model is not ChatGPT. We measure with gpt-oss-20b, chosen because it is independent of us and reproducible. Its citation preferences cannot be assumed to match any commercial assistant.
The page features are correlations. We report what the page carried. We cannot prove that any single feature is why it was cited — the pages that carry JSON-LD may simply be built by people who know what they are doing.
Nothing on this listing was paid for. It is free and unsponsored. If a placement is ever paid for, it will say so on the listing itself and carry rel=sponsored. Payment never changes a citation count.
Your listing would carry the same block
Check your own page right now
- The same eight checks, run on your page
- Every fix as copy-ready code
- Runs in your browser, nothing uploaded
What you get when it does open
- A listing page carrying your citation status
- Re-measured every month, kept over time
- We do not promise the number will rise