What Is llms.txt Actually For?

The llms.txt conversation has gotten ahead of the actual use case. It is often described as if it is the new sitemap for AI, or a ranking lever, or a shortcut around the harder work of making pages useful. That is the wrong frame.

For travel teams, llms.txt can be useful. It can point AI systems, agents, or tools toward the pages you think are most important. It can summarize what a site contains. It can make a source layer easier for a machine to scan. But it does not replace the public pages, the crawl rules, the internal links, or the facts those pages need to contain.

I would treat llms.txt as a packaging layer, not the product. The product is still the source system: the pages that explain the destination, hotel, attraction, venue, event, or itinerary clearly enough to be used in an answer.

What this dispatch covers

• What llms.txt can reasonably do for a travel site.

• What Google and OpenAI say about AI visibility basics.

• Why crawler access still matters more than a new text file.

• What a useful llms.txt should include for a destination or hotel.

• How to avoid turning it into another AI gimmick.

The Useful Definition

A useful llms.txt file is a plain-text guide to the source material on a site. It can say, in effect: these are the pages that explain who we are, what this place offers, which topics matter, and where an AI system should look first if it wants a clean overview.

That can help when an AI tool, agent, or crawler chooses to read it. The file can reduce ambiguity. It can make important pages more discoverable. It can express editorial intent in a format that is easy to parse.

The mistake is assuming that a machine-readable guide creates trust by itself. A guide to weak pages is still weak. A guide to pages that contradict partner sources is still confusing. A guide that says the destination is good for families does not help much if the linked pages do not explain the actual family trip: ages, timing, neighborhoods, transportation, weather, prices, and things to do.

What It Does Not Do

Google is direct on this. In its AI optimization guidance, Google says it ignores llms.txt for Search visibility and rankings. The same guidance says site owners do not need new machine-readable files, AI text files, AI-specific markup, special chunking, or AI-specific rewriting to appear in Google Search generative AI experiences.

Google's AI features documentation points back to familiar fundamentals: allow crawling, make sure important pages can be indexed, use internal links, put meaningful text on pages, and make structured data match visible content when used. That is not a glamorous answer, but it is useful because it keeps teams from mistaking a file for a strategy.

In other words, llms.txt is not a ranking switch for Google. It may be useful for other systems. It may become part of how some tools navigate a site. It should not be used as an excuse to postpone the source work.

Crawler Access Comes First

If a travel team wants to be present in AI search, access has to be checked first. OpenAI says publishers who want their content included in ChatGPT Search should allow OAI-SearchBot in robots.txt and avoid blocking OpenAI's published IPs. The ChatGPT Search help page also says ranking uses multiple factors and top placement cannot be guaranteed.

That is a good operating principle across AI search. Make sure the pages that matter can be crawled. Make sure they are not hidden behind scripts, blocked by robots rules, trapped in PDFs, or written as thin landing pages with little reusable information. A well-written llms.txt file pointing at blocked or shallow pages will not solve the problem.

The order matters: crawler access, indexable source pages, internal links, factual clarity, then packaging. llms.txt belongs near the end of that sequence.

What A Travel llms.txt Should Include

For a destination, hotel, venue, or attraction, a useful llms.txt should be short enough to maintain and specific enough to be worth reading. It should not try to list the entire website. It should point to the source pages that explain the entity, the trip intents it serves, and the stable facts that are most likely to support answers.

• A concise description of the destination, hotel, venue, or attraction.

• Links to official planning pages, such as getting here, neighborhoods, seasons, accessibility, meetings, events, and itineraries.

• Links to high-value evergreen pages that answer real traveler questions.

• A note on what each linked page is useful for, written in plain language.

• Contact or media information if journalists, agents, or partners need a source of record.

• A date or version marker so the file can be audited over time.

The best version reads less like SEO copy and more like a field guide to the site's source material. It should help a machine find the right page, not try to persuade the machine inside the file itself.

What Still Matters More

The visible page still matters more because it is where trust is earned. If the page says a destination is easy for car-free travel, it should explain transit access, walkable areas, distances, airport or rail connections, and realistic itinerary shapes. If a hotel wants to be cited for a gallery weekend, it should explain the neighborhood, nearby venues, dinner options, transit, and why staying there reduces friction.

AI systems need sources they can quote, summarize, compare, and cross-check. That means clear pages, stable URLs, coherent entity names, strong internal links, current information, and alignment between official and partner sources. The llms.txt file can point toward that work. It cannot create it after the fact.

For Drifter, this is the reason we start with the source layer. Before asking whether a destination needs a new machine-readable file, we ask which traveler questions the place should answer, which pages support those answers, and whether AI systems are already finding a better source somewhere else.

FAQ

Should a travel website have llms.txt?

It can be worth adding if it is maintained and points to genuinely useful pages. It is low-risk when written plainly, but it should not be treated as the main AI visibility program.

Will llms.txt improve Google AI Overviews visibility?

Google says it ignores llms.txt for Search visibility and rankings. For Google, the practical work is still crawlability, indexability, helpful visible content, internal links, and accurate structured data where appropriate.

Is robots.txt more important than llms.txt?

For access, yes. If important crawlers are blocked, they may not reach the pages that explain the place. llms.txt can guide a reader, but robots rules control access in a more direct way.

What should not go in llms.txt?

Do not put claims there that are not supported on public pages. Do not make it a keyword dump. Do not list stale pages because they look important internally. The file should be a clean map to useful source material.

The Practical Role

llms.txt is useful when it organizes a source system that already deserves to be read. It is not useful as a shortcut around clarity. For travel teams, the work is to make the place understandable first, then package the best sources in a way machines and people can navigate.