The new AI SEO rituals: markdown mirrors, llms-author.txt, and what Google actually said about them
Every few months the AI SEO world produces a new ritual. Right now there are two making the rounds: publishing a parallel set of markdown pages that mirror your HTML content, and adding an llms-author.txt file to claim authorship for AI systems. Google was asked about both this week, and the answers are worth hearing straight.
On the markdown trend, Search Engine Journal reports that Google's John Mueller responded to the practice of creating a set of markdown pages to mirror HTML content for AI SEO (searchenginejournal.com). On llms-author.txt, Mueller answered an SEO who was using the file to distinguish themself from other people with the same name (searchenginejournal.com). And in a separate Reddit reply covered days earlier, Mueller said Google does not use llms.txt or llms-author.txt, and that he does not know of any crawler or LLM confirming they use those files, other than SEO tools (seroundtable.com).
We recommend reading the coverage yourself at the links above rather than taking our paraphrase. But the pattern across all three answers is consistent, and it is the pattern that matters.
Why rituals spread faster than fundamentals
A ritual is attractive because it is finishable. You add the file, you generate the markdown copies, you check the box, you feel ahead of the curve. Fundamentals are unattractive because they are never finished. Fixing heading structure, cleaning up contradictory facts across pages, making content render without JavaScript tricks, keeping structured data honest, none of that gives you the satisfying click of a box being checked.
But the machines doing the reading are blunt. A crawler uses the directives it supports and ignores the rest. A file that no engine reads is not a signal, it is a souvenir. And a second copy of your site in a different format is a second thing to keep accurate, which means over time it becomes a second place to be wrong.
That last point is the one we care most about. Duplicated surfaces drift. When your markdown mirror says one price and your HTML page says another, you have not optimized for AI, you have manufactured a contradiction for it to find.
The boring, load-bearing version
If you want AI engines to read you correctly, give them one canonical, readable version of the truth. Make it render, make it parse, make the facts agree with each other, and mark it up with structured data that engines have actually documented support for. None of that is novel. All of it is checkable.
Checkable is our favorite word. The free scan reads your site the way Google and the AI answer engines actually read it, and tells you in plain language what they see today, including whether your pages are readable without tricks and whether your structured data holds up. No email required to get your diagnosis.
Sources: Search Engine Journal, "Google On Using Markdown For AI SEO," July 7, 2026 (searchenginejournal.com). Search Engine Journal, "Google Answers Question About LLMs-Author.txt For SEO," July 6, 2026 (searchenginejournal.com). Search Engine Roundtable, "Google: Cloudflare Content Signals Robots.txt Directive Has No Effects," July 6, 2026 (seroundtable.com).