Google confirmed llms.txt does nothing for Search rankings. Stop hand-crafting them for Google.

Every other site audit I look at lately has a freshly minted llms.txt sitting in the root, and the owner is convinced it's helping them show up in AI. Google just put that one to bed.

They updated the AI optimization guide. The line that matters:


So what that actually means in practice:

• For Google Search and Google's own AI Overviews / AI Mode -> the file does nothing. Not a penalty, not a boost, just ignored. Google's line hasn't changed: normal SEO is what gets you pulled into AIO, not a special text file.
• For other AI systems -> some tools and crawlers outside Google do read llms.txt. If one of those is a real traffic source for you, keep the file. Just don't kid yourself it's a Google play.

Reason I'm posting this: llms.txt has turned into the meta keywords tag of 2026. A thing people spend an afternoon hand-crafting because a LinkedIn post said it's the future, while the stuff that actually moves rankings sits untouched.

If you've got a spare hour, the boring answer still wins -> internal links that genuinely point at your money pages, content that answers the query instead of dancing around it, and real links from sites in your niche. AI answers are reading the same signals, so that work pays off on both fronts.

Anyone here actually tracked traffic from llms.txt on the non-Google surfaces? Curious whether it's moved anything real, or if it's hype end to end.

Hetneo said:

llms.txt has turned into the meta keywords tag of 2026

Click to expand...

That's a good comparison.

Has this been properly tested though?

Because what they say is one thing...

Two Rules .


1. Do opposite of what Google tells you .
2. Read rule 1 again .

Hetneo said:

Every other site audit I look at lately has a freshly minted llms.txt sitting in the root, and the owner is convinced it's helping them show up in AI. Google just put that one to bed.

They updated the AI optimization guide. The line that matters:

"neither harm nor help your site's visibility or rankings in Google Search, as Google Search ignores them."​

So what that actually means in practice:

• For Google Search and Google's own AI Overviews / AI Mode -> the file does nothing. Not a penalty, not a boost, just ignored. Google's line hasn't changed: normal SEO is what gets you pulled into AIO, not a special text file.
• For other AI systems -> some tools and crawlers outside Google do read llms.txt. If one of those is a real traffic source for you, keep the file. Just don't kid yourself it's a Google play.

Reason I'm posting this: llms.txt has turned into the meta keywords tag of 2026. A thing people spend an afternoon hand-crafting because a LinkedIn post said it's the future, while the stuff that actually moves rankings sits untouched.

If you've got a spare hour, the boring answer still wins -> internal links that genuinely point at your money pages, content that answers the query instead of dancing around it, and real links from sites in your niche. AI answers are reading the same signals, so that work pays off on both fronts.

Anyone here actually tracked traffic from llms.txt on the non-Google surfaces? Curious whether it's moved anything real, or if it's hype end to end.

Click to expand...

Google is a search engine first before an AI company. They already have access to your site's content and navigation. The robots.txt is sufficient for them. It makes for an extra file like the llms.txt to not make any meaningful difference.

That being said, remember that Google is known to say something doesn't work, even though it does. I mean, take this with a grain of salt.

I am in two minds about how to comment on llms.txt files. On one hand they're mostly just 'busy work' and you can spend your time more efficiently on other things. On the other hand, an llms.txt file that someone has put effort into is essentially a concise structured map of the site's content hierarchy, which for scrapers like me is a potential goldmine. So I'm gonna say "You should totally create detailed llms.txt files detailing all your juicy money pages and keywords and make them freely accessible to me"

tiiberius said:

Has this been properly tested though?

Click to expand...

Honestly? Not cleanly, and I'm not sure it can be. llms.txt isn't an isolated variable - the moment you write a good one you've also mapped out the site's content, and the AI surfaces read that same content directly. So if traffic moves, was it the file or the fact you just did a content audit while making it? On Google's surfaces you can't even start, because they say they ignore the file outright.

The one place you can test is non-Google: pick a crawler that actually fetches llms.txt, watch your server logs for that user-agent, and see if it lines up with referral traffic from that tool. That's the only clean signal I've seen. If anyone's sitting on that log data I'd genuinely love to see it - beats arguing from Google's PDF.

mosesjohn001 said:

overlooking fundamentals such as site structure, content quality, and topical authority

Click to expand...

This is the whole thing right here. The afternoon someone burns hand-crafting an llms.txt is almost always the afternoon their internal linking or a thin money page actually needed. The shiny file feels productive because it's new; the boring work feels like chores because you've heard it a hundred times. That part of the game hasn't changed since 2015.

HenryObi said:

Google is known to say something doesn't work, even though it does

Click to expand...

Fair, and I keep that same grain of salt on the shelf. But there's a tell for when to use it: Google fudges when the truth would cost them - "links aren't that important," "there's no crawl budget" - the stuff where you playing along keeps you off their case. llms.txt being ignored costs them nothing to admit; they gain nothing from you skipping a text file. That lands it in the "probably straight" column for me. Not gospel, the incentives just happen to line up this time.

Steptoe said:

a concise structured map of the site's content hierarchy

Click to expand...

Best point in the thread, and it cuts clean against the hype. A granular llms.txt is a gift-wrapped index of exactly which pages you make money on - handed to every competitor running a gap analysis and every scraper like yourself. Folks are publishing their whole content strategy in plaintext at /llms.txt because a LinkedIn post told them it's the future. So yes, everyone, please list your juiciest money pages in root. Steptoe says thanks.

Topiano said:

Do opposite of what Google tells you

Click to expand...

The one time reverse psychology costs you an afternoon instead of saving one. Although - going by Steptoe above - maybe do the opposite here too and don't write the file.

Hetneo said:

The one time reverse psychology costs you an afternoon instead of saving one. Although - going by Steptoe above - maybe do the opposite here too and don't write the file.

Click to expand...

Google Search probably has nothing to do with llm.txt in the first ground.

No measurable gains from the `llms.txt` file either. It is not really necesary for non-Google AI bots, but you still need to spend that hour on working on your internal links and content depth as they make a difference