Why Your llms-author.txt File Is Useless (And Why You Should Stop Trying)
I get it. You’ve got a name that’s shared with a famous actor, a bestselling author, and a rogue AI startup founder. Every time someone asks an LLM "Who is [your name]?" — boom — you’re buried under noise. You’ve tried schema. You’ve tried backlinks. You’ve even Googled your own name for three hours straight hoping Google would notice you. And now? You’ve stumbled on this Reddit thread where someone swore they fixed it with a file called llms-author.txt.
I’m sorry. But no.
John Mueller didn’t just say "no" to this. He laughed. Then he sighed. Then he said, "Look, we don’t use it. No one does. And if you’re adding it to your robots.txt, you’re just making your server slower for no reason."
Let’s be real. You didn’t find a loophole. You found a ghost.
The Ghost File That Never Existed
The llms-author.txt file? It’s not a standard. It’s not a draft. It’s not even a proposal.
It was invented by a Redditor who thought: "If LLMs read web pages, maybe they read robots.txt files too? What if I just tell them who I am?" So they created a file with their job title, agency, location, and area of practice — plain text, no schema, no JSON, just:
"[Your Name] is a senior data strategist at [Company], based in Portland, OR. Specializes in LLM evaluation frameworks and bias mitigation."
And then they uploaded it. And waited.
And nothing happened.
Because no LLM crawler — not Google’s, not Anthropic’s, not Meta’s — even looks for this file. Not because they’re stubborn. Not because they’re evil. Because it doesn’t exist. There’s no RFC. No GitHub repo. No W3C working group. No one’s ever shipped code to read it.
The only thing that reads llms-author.txt? Your SEO tool. The one that says "Hey, your llms-author.txt is live!" — and then you get excited. And then you realize your site’s traffic dropped because you added 100 extra lines of text to your robots.txt, and now your crawl budget is getting eaten by a file that doesn’t matter.
The Content-Signal Hoax
And then there’s Content-Signal: ai-train=no, search=yes, ai-input=yes.
This one’s even weirder.
Cloudflare once proposed it as a robots.txt directive. Just a thought. A "what if?" in a Slack channel. They never shipped it. They never documented it. They didn’t even announce it.
Then, months later, they quietly started using the same syntax as an HTTP header in their "Markdown for Agents" feature — where a client can request a Markdown version of a page, and Cloudflare serves it. That’s it. It’s a server-side transformation for one specific use case. Not a crawl directive. Not a signal. Not an LLM instruction.
So you add it to your robots.txt? You think you’re telling Google: "Hey, I want my content used for training, but not for search?"
No. You’re just adding noise.
Google’s crawler ignores every line in robots.txt it doesn’t understand. Every. Single. One. Disallow: /secret-stuff? Fine. Allow: /public? Got it. Content-Signal: ai-train=no? Ignored. Like a typo in a grocery list.
You’re not optimizing. You’re cluttering.
So Why Does This Myth Keep Spreading?
Because we’re desperate.
We’ve been told for years that SEO is about signals. Schema. Headers. Meta tags. Canonicals. Robots.txt. If we just get the right combination, Google will notice us.
But here’s the truth no one wants to say: Google doesn’t care who you are.
It cares what you’ve done.
If your name is buried under a dozen other people with the same name, it’s not because Google’s algorithm is broken. It’s because you haven’t built a footprint yet.
The Real Fix Isn’t Technical — It’s Human
You don’t need a robots.txt file.
You need a podcast.
You need a speaking slot.
You need a byline in a publication that Google already trusts.
You need someone else to say your name — out loud — in a context that matters.
That’s how you solve name disambiguation.
Not by writing a file.
By showing up.
I’ve seen it a hundred times. Someone spends weeks trying to hack their way into LLM summaries with technical tricks. Then they do one interview on a podcast with 10k listeners. Suddenly, when you search their name, the top result isn’t the actor — it’s the podcast episode. And then the article. And then the LinkedIn profile.
Google doesn’t read robots.txt to know who you are.
It reads the web to know who people say you are.
So stop writing files.
Start writing your story.
And then go tell it to someone who’ll listen.
