The problem llms.txt solves
When an AI search engine crawls your hotel website, it encounters a mix of content: navigation, booking widgets, photo galleries, marketing copy, guest review widgets, cookie banners, and somewhere in there, actual information about your property.
AI systems are good at extracting meaning from messy content, but they perform better when they have a clean, authoritative source to start from. The llms.txt file is that source: a simple text file that tells the AI, upfront, exactly what it needs to know about your property.
What to put in a hotel’s llms.txt file
A good hotel llms.txt file contains:
1. Property name and category. One line: The Harbour House Hotel — a 24-room boutique hotel in Porto, Portugal.
2. A 2-4 sentence description. Plain prose, no marketing superlatives. Describe the property type, location (neighbourhood and nearest landmark), what makes it distinctive, and who it’s right for. This is what the AI may use word-for-word in a recommendation.
Example:
The Harbour House Hotel is a 24-room boutique hotel in Porto's Ribeira district,
facing the Douro river. Originally an 18th-century merchant's house, it has been
a hotel since 1987. It is particularly suitable for couples and solo travellers
who want a historic property with walkable access to Porto's main sights.
The nearest metro station (São Bento) is a 6-minute walk.3. Key attributes, listed. Things that match specific traveller queries:
Amenities: restaurant, bar, rooftop terrace, free WiFi, no pool, no parking
Room types: double rooms (18), suites (6)
Price range: €120-280 per night
Languages: Portuguese, English, Spanish, French
Pet policy: no pets
Check-in: 15:00, Check-out: 11:004. Nearby landmarks and transport. This powers location-based queries:
Location details: Ribeira district, Porto.
Walking distance: São Bento station (6 min), Porto Cathedral (8 min),
Luis I Bridge (4 min), Clérigos Tower (12 min).
Airport: Francisco Sá Carneiro Airport, approximately 20 minutes by taxi.5. Links to your most useful pages for AI systems that can follow links:
Full property description: /about/
Room types and rates: /rooms/
Guest FAQ: /faq/
Location and directions: /location/How to create and deploy the file
Create a plain text file (.txt extension). Put everything in it using clear headings and simple formatting: # Heading, - list item. No HTML.
Upload it to the root of your website so it’s accessible at yourdomain.com/llms.txt. If you use WordPress, upload via your hosting file manager or FTP. If you use a site builder (Squarespace, Wix, Webflow), check their settings for adding files to the site root, or ask your website developer to add it.
The whole process takes 30-60 minutes.
Does it work immediately?
AI systems that browse the web in real time (Perplexity, ChatGPT with browsing) will read your llms.txt file the next time they crawl your site. For Perplexity, this can happen within days of the file going live. For ChatGPT’s training data, the effect takes longer, as it depends on when the next training data collection run happens.
The earliest measurable effect is improved accuracy in how AI describes your property: fewer vague or incorrect descriptions, more specific attribute mentions.
See also: What schema markup does a hotel need | Get found: the technical playbook