What percentage of hotels have a llms.txt file?

Only 6.3% of 105,002 reachable hotel websites serve a valid llms.txt file (6,590 hotels). The companion llms-full.txt is even rarer at 0.3% (265 hotels). This means 93.7% of hotels have not yet adopted the emerging standard for communicating with AI models.

Which countries lead in hotel llms.txt adoption?

The USA leads major markets at 12.4%, followed by Spain (8.6%), Netherlands (8.2%), and the UK (6.9%). France trails at just 3.8% — consistent with French hotels being the most resistant to AI integration, as also seen in their higher robots.txt AI blocking rates.

What tools generate hotel llms.txt files?

WordPress SEO plugins account for 33.4% of hotel llms.txt files: AIOSEO (19.6% combined), Yoast SEO (10.2%), and Rank Math (3.6%). Hotel-specific platforms like Backhotelite (4.2%, Spain) and ComboCMS (2.1%, Italy) also contribute. The majority (57.4%) appear to be custom implementations.

Do higher-star hotels adopt llms.txt more?

Yes. 5-star hotels adopt at 10.9% — 2.5 times the rate of 1-star hotels (4.4%). This mirrors the pattern seen with schema.org adoption, where investment in technical SEO correlates with hotel class. Hotels with llms.txt also score 62% higher on schema.org quality overall.

What content do hotel llms.txt files contain?

42.9% follow the intended format — a site index listing pages with URLs and descriptions. However, 14.2% contain only a sitemap reference, 7.3% misuse the file for access control rules (like robots.txt), and 0.7% are near-empty. Only 2.9% contain rich hotel-specific descriptions with amenities, policies, and room details.

How was this study conducted?

We fetched /llms.txt and /llms-full.txt from 105,002 reachable hotel websites sourced from Google Maps. Files were validated to exclude HTML error pages and empty responses. Content was analyzed for generator identification, content classification, language detection, and structural analysis. Data was collected in March 2026.

ResearchResearch / llms.txt StudyMarch 2026

Hotel llms.txt Adoption Study

Name: Hotel llms.txt Adoption Study 2026
Creator: Nicolas Sitter
Published: 2026-03-21
License: https://creativecommons.org/licenses/by/4.0/

We scanned 105,002 hotel websites for llms.txt files. Only 6.3% have one — and 7.3% of those misuse it as a robots.txt clone.

105K

Hotels Scanned

6.3%

Adoption Rate

12.4%

US Lead

0.3%

llms-full.txt

Quick answers

How many hotels have an llms.txt file?

Only 6.3% of hotels serve a valid llms.txt file. Nicolas Sitter's Hotel llms.txt Adoption Study (2026) checked 105,002 reachable hotel websites across 7 countries and found just 6,590 with an llms.txt, meaning 93.7% have not adopted the emerging standard. The companion llms-full.txt is far rarer, present on only 0.3% (265 hotels). Adoption rises with hotel class: 5-star properties reach 10.9%, about 2.5 times the 4.4% rate of 1-star hotels. The USA leads major markets at 12.4%, followed by Spain at 8.6%, while France trails at 3.8%. Most files are plugin-generated, with WordPress SEO plugins accounting for 33.4%. — Nicolas Sitter, Hotel llms.txt Adoption Study 2026

Does llms.txt help hotels get cited by AI search?

There is no evidence that llms.txt causes better AI visibility; the link is correlation, not causation. Nicolas Sitter's Hotel llms.txt Adoption Study (2026) found that hotels with an llms.txt file score 62% higher on schema.org quality (22.4 versus 13.8) and have 49% higher JSON-LD adoption (80.6% versus 54.2%) than hotels without one. The study attributes this to llms.txt being a proxy for technical SEO maturity rather than a direct ranking lever, since both signals stem from the same investment in optimization. Quality is also uneven: 14.2% of files are just a sitemap reference, 7.3% misuse the file for access control, and only 2.9% contain rich hotel descriptions. — Nicolas Sitter, Hotel llms.txt Adoption Study 2026

Updated · June 2026

The adoption numbers below are the original measurement and still stand. They pair with the hotel robots.txt & AI-blocking study — inviting crawlers via llms.txt and blocking them via robots.txt are two sides of the same dial. Checking your own site? The free tools cover schema and crawl access.

Summary Adoption Generators Content By Country By Stars Schema Link File Sizes FAQ Methodology

Update — May 2026

Shopify now ships llms.txt, llms-full.txt, and agents.md by default on every store, and exposes them through a discovery sitemap at /sitemap_agentic_discovery.xml. Live example: respire.co/sitemap_agentic_discovery.xml — the sitemap points to all three files plus a .well-known/ucp Universal Commerce Protocol endpoint.

This isn’t reflected in the adoption numbers below — our crawl predates the rollout, and Shopify-hosted hotels are a minority of the sample. But if Shopify is doing it, the trajectory is clear: llms.txt adoption will stop being a marker of technical SEO maturity and start being whatever your platform ships by default. Worth tracking on the next crawl.

Update — June 2026

Google can’t quite decide whether llms.txt matters. Its search side has been openly dismissive — the line from Search Central has been that Google doesn’t use llms.txt for ranking, in roughly the same breath it buried the old keywords meta tag. Fair enough.

And yet Google’s Chrome team now ships a Lighthouse audit that recommends publishing an llms.txt file for agentic browsing, noting that “without this file, agents may spend more time crawling the site to understand its high-level structure and primary content” (developer.chrome.com). So the company telling you it’s useless is also the company telling its own agent developers to look for it. The resolution isn’t hypocrisy — it’s that search and agents are different surfaces. llms.txt was never really about ranking; it was always aimed at the agent doing the reading.

TL;DR

We fetched /llms.txt and /llms-full.txt from 105,002 hotel websites across 7 countries. Only 6.3% have a llms.txt file (6,590 hotels) and just 0.3% serve a llms-full.txt (265 hotels). The US leads adoption at 12.4%, while France trails at 3.8%. WordPress SEO plugins (AIOSEO, Yoast, Rank Math) drive 33.4% of all files, but 7.3% of llms.txt files misuse the format as robots.txt-style access control rules. Hotels with llms.txt score 62% higher on schema.org quality — suggesting it's a marker of technical SEO maturity.

Executive Summary

The llms.txt file is an emerging standard for websites to communicate their content structure to AI models. Unlike robots.txt (which tells bots what not to crawl), llms.txt tells AI models what a site is about — a curated index of pages with descriptions that helps LLMs understand and accurately represent a property.

Our analysis of 105,002 hotel websites reveals extremely early adoption. At 6.3%, llms.txt is where robots.txt was in the early 2000s — a technical signal adopted by forward-thinking properties and platforms, but unknown to the vast majority. The companion file llms-full.txt (designed for detailed content) is even rarer at 0.3%, and every hotel with llms-full.txt also has llms.txt — the two-file approach has not gained traction.

Most adoption is plugin-driven, not strategic. WordPress SEO plugins (AIOSEO, Yoast, Rank Math) account for 33.4% of all llms.txt files, often auto-generated with little curation. The best files — like rich hotel descriptions with room types, amenities, and policies — represent just 2.9% of the total. Meanwhile, 7.3% of files are outright misconfigurations: robots.txt-style access control rules served as llms.txt, providing zero value to AI models.

93.7%

No llms.txt file

42.9%

Follow intended format

7.3%

Misuse as access control

The key finding: llms.txt adoption is a proxy for technical SEO maturity, not an isolated decision. Hotels with llms.txt have 62% higher schema.org scores (22.4 vs 13.8) and 49% higher JSON-LD adoption (80.6% vs 54.2%). Both stem from the same investment in search optimization — whether driven by a web agency, CMS platform, or technically aware hotelier.

Adoption Overview

How does llms.txt adoption compare to robots.txt? (n=105,002 hotels)

6.3%

Have llms.txt

6,590 hotels

0.3%

Have llms-full.txt

265 hotels

82.2%

Have robots.txt

For comparison

Only llms-full.txt

Always paired with llms.txt

llms.txt vs robots.txt adoption rates

llms.txt adoption overview

Metric	Count	% of Reachable
Has llms.txt	6,590	6.3%
Has llms-full.txt	265	0.3%
Has either	6,590	6.3%
Has both	265	0.3%

llms-full.txt adoption is negligible. Every hotel with llms-full.txt also has llms.txt — there are zero hotels with only the full variant. This suggests the standard's two-file approach hasn't gained traction; hotels treat llms.txt as a single deliverable.

Who Generates These Files?

WordPress SEO plugins drive 33.4% of hotel llms.txt files. The majority (57.4%) are custom.

57.4%

Custom / Unknown

3,784 files

33.4%

WordPress Plugins

AIOSEO + Yoast + Rank Math

4.2%

Backhotelite

Spain-specific platform

2.1%

ComboCMS

Italy-specific platform

Generator breakdown of hotel llms.txt files

Generator breakdown

Generator	Count	% of llms.txt Files
Custom	3,784	57.4%
AIOSEO	1,197	18.2%
Yoast SEO	675	10.2%
Backhotelite	280	4.2%
Rank Math	238	3.6%
robots-style	179	2.7%

WordPress SEO plugins are the primary driver of identifiable adoption. AIOSEO (19.6% combined free + pro), Yoast SEO (10.2%), and Rank Math (3.6%) together account for 33.4% of all llms.txt files. These plugins added llms.txt generation as a feature in late 2025/early 2026, making adoption as simple as toggling a setting. But auto-generated files need curation — many are generic site indexes, not curated hotel descriptions.

Industry-specific platforms

Spain

Backhotelite (4.2%) — A hotel-specific CMS platform primarily used by Spanish properties. Accounts for 18.1% of Spanish llms.txt files.

Italy

ComboCMS (2.1%) — An Italian hotel CMS that auto-generates llms.txt. Accounts for 10.5% of Italian llms.txt files.

Misuse

robots-style (2.7%) — 179 files contain User-agent/Allow/Disallow directives instead of the intended llms.txt format. These provide zero value to LLMs.

What's Inside These Files?

Only 42.9% follow the intended spec. 7.3% misunderstand the format entirely.

Content type distribution of hotel llms.txt files

Content type breakdown

Content Type	Count	%	Description
Site Index	2,826	42.9%	Page listings with URLs and descriptions
Other	2,013	30.5%	Mixed or unstructured content
Sitemap Only	933	14.2%	Just a sitemap reference, no page details
Access Control	484	7.3%	User-agent allow/disallow rules (misuse)
Hotel Description	194	2.9%	Detailed property info, amenities, policies
Summary	91	1.4%	Brief business summary with contact info

7.3% of llms.txt files are robots.txt clones. 484 hotels serve User-agent/Allow/Disallow rules as their llms.txt — a fundamental misunderstanding of the standard. These are primarily generated by Backhotelite, a Spanish hotel platform. While well-intentioned (explicitly allowing AI crawlers), this format doesn't help LLMs understand the site's content. The llms.txt spec is about describing your content, not controlling access to it.

The 2.9% doing it right: rich hotel descriptions

194 hotels serve llms.txt files with detailed property information — room types, cancellation policies, amenities, and contact details. These are exactly what an AI concierge needs to recommend a property.

The sweet spot for content is 11-50 pages listed (39.1% of files), covering a typical hotel site's key pages: rooms, amenities, location, dining, events, contact, and blog posts. The median file lists 15 pages.

Adoption by Country

The US leads at 12.4%. France trails at 3.8% — consistent with its resistance to AI integration.

llms.txt adoption rate by country

Full country-level adoption data

Country	Reachable Hotels	Has llms.txt	% Adoption	llms-full.txt %	Top Generator
USA	7,445	924	12.4%	0.2%	Custom
Spain	16,411	1,418	8.6%	0.3%	Custom
Netherlands	2,891	237	8.2%	0.4%	Custom
United Kingdom	10,547	729	6.9%	0.3%	Custom
Germany	22,268	1,256	5.6%	0.2%	Custom
Italy	27,319	1,309	4.8%	0.3%	Custom

The US leads at 12.4% — 2x the adoption rate of France. American hotels are more likely to embrace AI visibility, and the US also has the highest "custom" rate (77.1%), suggesting tech-savvy hoteliers writing their own files rather than relying on CMS defaults.

The France Paradox: Blocks Most, Adopts Least

France has the highest AI blocking rate (7.5% in our robots.txt study) AND the lowest llms.txt adoption (3.8%) among major markets. This is a consistent signal: the French hospitality industry is the most resistant to AI integration.

7.5%

France AI blocking rate

Highest of 7 countries

3.8%

France llms.txt adoption

Lowest of 7 countries

3.3x

Block-to-adopt ratio

vs US at 0.2x

For comparison, the US has the opposite pattern: lowest blocking rate (2.1%) and highest llms.txt adoption (12.4%). The divergence suggests fundamentally different attitudes toward AI in the hospitality industry — with France leaning toward restriction and the US toward visibility.

Spain is surprisingly strong at 8.6% — driven partly by Backhotelite (18.1% of Spanish llms.txt files), a hotel CMS platform that auto-generates the file. Platform decisions matter: a single CMS provider shipping llms.txt support can meaningfully move a country's adoption rate.

Adoption by Star Classification

5-star hotels adopt at 2.5x the rate of 1-star properties.

llms.txt adoption rate by star classification

Adoption rates by star classification

Stars	Hotels	Has llms.txt	% Adoption	llms-full.txt %	Avg File Size
5-star	2,062	225	10.9%	0.3%	11.0 KB
4-star	16,548	1,459	8.8%	0.2%	17.9 KB
3-star	30,199	1,812	6%	0.3%	19.6 KB
2-star	10,222	539	5.3%	0.3%	10.9 KB
1-star	2,699	118	4.4%	0.3%	12.9 KB
Unclassified	43,272	2,437	5.6%	0.3%	10.9 KB

Clear correlation with hotel class. 5-star hotels adopt llms.txt at 10.9% — 2.5x the rate of 1-star hotels (4.4%). This parallels our schema.org findings where higher-rated properties invest more in technical SEO. Interesting size pattern: 3-star hotels have the largest average file size (20 KB), driven by WordPress sites with many blog posts and pages. 5-star hotels have smaller but more focused files (11 KB), suggesting curated content rather than full site dumps.

Schema.org Correlation

Hotels with llms.txt have 62% higher schema.org scores — it's a proxy for technical SEO maturity.

22.4

Schema Score (with)

6,590 hotels

13.8

Schema Score (without)

98,412 hotels

80.6%

JSON-LD (with)

vs 54.2% without

+62%

Score Difference

Strong correlation

Average schema.org score: with vs without llms.txt

With llms.txt

Hotels6,590

Avg Schema Score22.4

JSON-LD Adoption80.6%

Without llms.txt

Hotels98,412

Avg Schema Score13.8

JSON-LD Adoption54.2%

llms.txt adoption is a proxy for technical SEO maturity. Hotels that care enough to implement llms.txt also tend to have better structured data. It's not that llms.txt causes better schema — it's that both are symptoms of a hotel (or its web agency) that takes search optimization seriously. Use llms.txt presence as a quick diagnostic: if a hotel client doesn't have one, they likely have schema.org gaps too.

File Size Distribution

Median: 3.3 KB (~15 pages). The long tail reaches 1 MB.

3.3 KB

Median Size

Typical 10-20 page listing

15 KB

Average Size

Skewed by large files

1 MB

Maximum Size

Large resort site dumps

Median Pages

Avg 38.8 pages

llms.txt file size distribution

File size distribution

Size Range	Count	% of Files
1-100 B	171	2.6%
101-500 B	282	4.3%
501 B-1 KB	820	12.4%
1-5 KB	2,620	39.8%
5-10 KB	1,208	18.3%
10-50 KB	1,110	16.8%

Pages listed per llms.txt file

Pages Listed	Count	% of Files
0 pages	1,398	21.2%
1-5 pages	376	5.7%
6-10 pages	994	15.1%
11-20 pages	1,100	16.7%
21-50 pages	1,473	22.4%
51-100 pages	737	11.2%

The median of 3.3 KB represents a typical 10-20 page listing with descriptions. The long tail above 100 KB (188 files) consists of large resort/chain sites that list hundreds or thousands of pages with full descriptions. 21.2% of llms.txt files list zero pages — these are the access-control and minimal files that don't serve the intended purpose.

Frequently Asked Questions

Methodology

Data Collection

Source: Global hotel index — 121,425 hotels from Google Maps
Reachable websites: 105,002 (86.5% of total)
Files checked: /llms.txt and /llms-full.txt per hotel
7 countries: US (7.4K), ES (16.4K), NL (2.9K), UK (10.5K), DE (22.3K), IT (27.3K), FR (17.6K)
Files fetched during March 2026 crawl window

Content Analysis

Generator detection: Heuristic identification from file header comments (AIOSEO, Yoast, Rank Math, etc.)
Content classification: Rule-based categorization into site-index, access-control, hotel-description, summary, sitemap-only, minimal
Language detection: Keyword-frequency heuristics for English, German, Spanish, Italian, French
Structural analysis: Section header extraction, URL counting, page listing enumeration

105,002

Hotels Scanned

86.5% of index

6,590

Valid llms.txt Found

6.3% adoption

Countries Covered

US, ES, NL, UK, DE, IT, FR

Continue Reading

Explore more research on AI hotel search.

AI Hotel Landscape 2026

robots.txt AI Blocking Study Schema Adoption Study Anatomy of ChatGPT Search Google AI Mode Study All Research