llms.txt: the structured summary AI systems read directly.
llms.txt is a plain-text file at the root of a website that gives AI language models a structured, authoritative summary of the business. Without it, AI systems reconstruct an understanding from scattered HTML - less reliable, less complete, and more likely to contain errors. VERIS is an AI search infrastructure specialist for independent service businesses.
The audit identifies whether this service is needed for your site.
AI systems are piecing together your business from scattered clues.
When a user asks ChatGPT or Perplexity about a specific hotel, restaurant, or law firm, the AI system either draws from its training data or performs a live web search. In both cases, the AI is reconstructing an understanding of the business from whatever it can find - HTML pages, directory listings, review sites, and news mentions. The business has no direct input into that reconstruction.
llms.txt changes this. It is a structured markdown file at yourdomain.com/llms.txt that explicitly tells AI systems what the business is, what it does, where it is, and which pages contain the most relevant information. It follows an emerging specification that some AI products and crawlers are beginning to support directly.
The companion file, llms-full.txt, goes further - it contains the detailed operational information that AI systems need to make accurate, specific recommendations: room types and rates for hotels, cuisine and hours for restaurants, specialties and new patient status for medical practices, practice areas and jurisdiction for law firms. This is the data that determines how accurately AI systems describe the business to potential customers.
LLM referral traffic conversion rate. A clearer AI-readable business summary matters because this traffic tends to be high intent when it does arrive.
Two files. Both structured. Both deployed.
Markdown file following the llms.txt specification. H1 business name, blockquote description, and sections for services, pricing/booking, location, and about with annotated links.
Expanded operational detail specific to the business category: room types, cuisine, specialties, policies, or practice areas.
Confirm that robots.txt does not disallow /llms.txt or /llms-full.txt so AI crawlers can access both files.
Both files confirmed accessible at their URLs. Screenshots taken of live files.
# The Grove Inn > A boutique hotel in Asheville, North Carolina with 18 rooms, > specializing in couples and wellness retreats in the Blue Ridge > Mountain area. Independently owned since 2009. ## Accommodation - https://thegroveinn.com/rooms: Room types, amenities, and rates - https://thegroveinn.com/packages: Seasonal packages and special offers ## Reservations & Policies - https://thegroveinn.com/book: Direct booking (best rate guaranteed) - https://thegroveinn.com/policies: Check-in 15: 00, check-out 11: 00, 48-hour cancellation, pet-friendly with prior approval ## Location - https://thegroveinn.com/location: 142 River Street, Asheville NC 28801. 10-minute walk to downtown, free parking on-site. ## About - https://thegroveinn.com/about: Founded 2009 by [owner name]. Member of the Independent Lodging Alliance.
What this service does not include.
- Content writing for the pages linked from llms.txt
- Creating pages that do not currently exist on the site
- Updating the business information after delivery (covered under maintenance)
- SEO optimization of any page
- Translation into non-English languages
Check if llms.txt exists in 5 seconds.
Open your browser. Go to yourdomain.com/llms.txt. If you see a 404 Not Found page, the file does not exist. If you see a structured markdown document, it is live. After VERIS deploys the file, you can visit the URL and read exactly what AI crawlers are receiving about your business.
Step 1: Visit the URL in your browser. Step 2: 404 = file absent. Structured text = file live. Step 3: Verify the content is accurate and complete.
Common questions about llms.txt.
VERIS treats llms.txt as a practical structured summary file that sits alongside schema, crawler permissions, and discovery setup. The operational question is whether the file is live, readable, and maintained accurately.
llms.txt is a short summary with annotated links to important pages. llms-full.txt is the detailed operational document - full policy text, complete service descriptions, and business context. AI crawlers can read both, but llms-full.txt is where the specific operational data lives.
No. They serve different functions. Schema markup is machine-readable structured data embedded in HTML - it tells search engines and AI systems what entities exist on the page. llms.txt is a human-readable summary that AI language models can directly parse. Both are part of Layer 1.
The monthly maintenance retainer handles updates. Any time the business changes - new hours, services, policies, prices - the llms.txt and llms-full.txt files should be updated to reflect the change.
Not automatically. llms.txt only helps when a crawler or product is designed to look for it, the file is public, and the relevant robots directives allow access. Support is still uneven, so llms.txt should complement schema and crawl access rather than replace them.
Find out if your llms.txt exists - and what it should say.
The audit checks both /llms.txt and /llms-full.txt and assesses whether the content accurately represents your business.