What you get
Five files, written from your site rather than from a template: llms.txt (the index engines read first), llms-full.txt (your key pages' text in one file), robots.txt (every AI crawler named and allowed, private paths kept out), sitemap.xml (the pages found, dated where your server said so) and a JSON-LD graph (Organization, WebSite, WebPage) that keeps whatever you already declare and adds what was missing.
Where each file goes
- llms.txt and llms-full.txt: the site root, served as plain text.
- robots.txt: the site root. If you already have one, merge the named blocks in and keep your own Disallow lines.
- sitemap.xml: the site root, then submit it in Google Search Console and Bing Webmaster Tools.
- schema.json: inside a
<script type="application/ld+json">tag in the home page's head.
What the files say to an engine
An answer engine deciding whether to name you does three things: fetches your pages, works out what you are, and decides whether to trust it. robots.txt and the sitemap are the fetch. llms.txt and the schema are the what: the same facts, once in prose and once in a graph, with the same name, the same summary and the same profiles, so nothing contradicts. The trust comes from the rest of the web, which is what the full check measures.
After you upload
Run the llms.txt checker, the AI crawler checker and the schema checker against your domain; each should now pass. Then run the full check to see whether being readable has become being named.
Questions about the files
Is this the same as the report's technical checks?
It is the fix for four of them. The report checks whether llms.txt, robots.txt, a sitemap and Organization schema are there; this writes the four files from your own pages so you can upload them in an afternoon.
Will the files be right as generated?
Right enough to upload, and honest where the crawl found nothing: a summary line or a legal name we could not read is marked REPLACE rather than invented. Read the summary line and the schema before you publish them; the rest is mechanical.
Does llms-full.txt matter?
Less than the others. It is the text of your key pages in one file for models that fetch once. Large documentation sites benefit; a ten-page site can skip it.
Why name every AI crawler in robots.txt when a wildcard allows them?
Because a wildcard is silent and a named block is a statement. Anyone reading the file, including our checker and the engines' own documentation, sees the welcome. It also stops a later edit from blocking one by accident.
Do you keep anything?
No. The pages are read, the files are written into the page you are looking at, and nothing is stored.
Readable is not the same as named
These files let the engines read you. The full check tells you whether they name you, and who they name instead.