Does llms.txt Improve AI Search Visibility?
Google says llms.txt does not help its Search or generative AI features, while OpenAI's publisher guidance focuses on OAI-SearchBot access.
The short answer is provider-specific. Google explicitly says you do not need an llms.txt file for its Search or generative AI features and that Google Search ignores the file. OpenAI's current publisher guidance focuses on allowing OAI-SearchBot and managing robots directives; it does not present llms.txt as a requirement for ChatGPT search inclusion.
That does not prove every present or future AI system ignores llms.txt. It does mean a small business should not treat the file as a universal ranking switch or prioritize it ahead of accessible, accurate, useful web pages.
What llms.txt is intended to provide
llms.txt is a proposed convention for placing a machine-readable guide at the root of a website. Supporters intend it to point language models or AI tools toward selected documentation and context, often in a concise Markdown format.
A proposal can be useful without being supported by every provider. Publishing a file, seeing a bot request it, or finding a product that reads it does not establish a ranking benefit across unrelated AI search systems.
Google's explicit position
Google's current documentation for AI features and websites says no new machine-readable files or AI text files are needed. It specifically states that you do not need an llms.txt file because Google Search does not use it.
Google instead points website owners to established fundamentals: allow crawling, make important content available as text, use internal links, provide a good page experience, and keep structured data consistent with visible content. It also says there is no special schema markup required to appear in its AI features.
What OpenAI currently tells publishers
OpenAI tells publishers who want content surfaced in ChatGPT search to allow OAI-SearchBot and its published IP addresses. It distinguishes that search crawler from GPTBot, which relates to potential model improvement, and says noindex prevents a page from appearing in ChatGPT search.
The guidance does not say that an llms.txt file is required. Follow the provider's current crawler documentation rather than assuming a community convention is an official prerequisite.
Why file discovery is not proof of ranking value
A server log may show that a crawler fetched llms.txt. That proves a request occurred, not that the contents affected retrieval, citation, ranking, or recommendations. Likewise, a favorable answer after publishing the file does not isolate the file as the cause.
Discovery: did a system request or locate the file?
Parsing: did it process the contents successfully?
Use: did the contents influence a retrieval or answer process?
Outcome: did measured mentions, citations, qualified visits, or conversions change?
Causation: can the change reasonably be attributed to the file rather than other factors?
When maintaining the file may still be reasonable
A low-maintenance llms.txt file may be reasonable if a tool you rely on explicitly documents support, your technical team wants to experiment, and the file can be kept accurate without displacing higher-priority work. Treat it as an experiment tied to named consumers, not as invisible insurance for all AI search.
Avoid putting confidential information, private URLs, unpublished claims, or content you cannot maintain into the file. Remember that a public root file is publicly accessible.
What to prioritize before llms.txt
Make important service, location, evidence, and contact pages publicly crawlable.
Remove accidental noindex directives and infrastructure blocks.
State what the business offers, who it serves, where it operates, and important limitations.
Support claims with current first-party evidence and responsible sourcing.
Use descriptive internal links so important pages are discoverable.
Keep business profiles and structured data aligned with visible page facts.
Measure repeatable buyer questions and referral outcomes with limitations recorded.
How to evaluate future provider support
Look for an official provider document that names llms.txt, explains what reads it, states the supported syntax, and describes the effect or limitations. Record the date because guidance can change. A third-party blog claiming “AI platforms use llms.txt” is not a substitute for provider-specific documentation.
If a provider adds support, test the exact behavior it documents. Check logs, validate the file, and measure the relevant outcome over time. Do not generalize one provider's support to every AI answer engine.
A no-hype founder decision
For most small businesses today, llms.txt should not outrank basic website clarity, crawler access, accurate profiles, evidence, and measurement. If those foundations are sound and maintaining the file is inexpensive, an explicitly labeled experiment can be sensible.
The honest position is narrower than either “llms.txt is essential” or “llms.txt is useless.” Google says it ignores the file. OpenAI's current publisher FAQ does not require it. Other systems may make different choices, so evaluate named providers using current primary documentation.
Primary sources used
Google Search Central, AI features and your website: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
OpenAI, Publishers and developers FAQ: https://help.openai.com/en/articles/12627856-publishers-and-developers-faq