llms.txt for Restaurants: How Restaurants Directly Instruct ChatGPT and Gemini
In 1994, robots.txt was introduced — a text file that tells search engine crawlers which parts of a website they may index. Since then, robots.txt has been a standard tool for every website operator.
In 2024, its successor was proposed: llms.txt. A text file that explains to AI crawlers — the systems feeding ChatGPT, Gemini, Perplexity and other LLMs with training data — what a website is, which content is trustworthy, and how the AI should handle this data.
For restaurants, llms.txt is one of the most direct routes to informing AI systems about their existence, specialties and trustworthiness.
→ AI Visibility Hub
→ What Is AI Visibility?
What llms.txt Is
llms.txt is a simple text file (Markdown format) stored in the root directory of a website — accessible at `https://yourdomain.com/llms.txt`. It provides Large Language Models (LLMs) with a structured summary of what a website is and what it contains.
The file has no official standards body — it was developed as a community proposal by Answer.ai (Jeremy Howard) and has since gained widespread adoption. Major websites including Anthropic, Cloudflare and Stripe have implemented llms.txt files.
For restaurants, an ideal llms.txt contains:
# China Restaurant YUNG Frankfurt
> Family restaurant since 1989. Chinese cuisine from Canton and Sichuan.
> Complete EU-14 allergen labeling. DE/EN/ZH/ES/TR/AR and more languages on request.
## Restaurant Information
- Address: Gutleutstr. 41, 60329 Frankfurt am Main
- Open: Tue-Sun 12pm-10pm, Mon closed
- Reservations: +49 69 12345678
- Cuisine: Cantonese, Sichuan, Dim Sum
## Digital Menu
- Complete menu: https://chinayung.chiwai.eu
- Allergens per dish: complete, EU-14 standard
- Languages: DE, EN, ZH, ES, TR, AR
- GBP push: active (Google Food Menus API)
## Trust Signals
- chiwai AI-visible Certificate 2026
- Wikidata entry: Q139379023
- Schema.org: Restaurant + MenuItem + NutritionInformation
## Contact
- Website: https://chinayung.de
- Email: [email protected]This structure instructs AI crawlers with all relevant facts — without interpretation, without guessing.
Why llms.txt Matters for Restaurants
AI systems like ChatGPT and Gemini need training data. They crawl the web to learn which restaurants exist, what they offer and how trustworthy their information is. Websites without llms.txt are treated by these crawlers like any other website — the crawler must figure out on its own what the site is about.
Websites with llms.txt offer the crawler a shortcut: „Here is exactly what I am, what I offer, and why my data is trustworthy.“ This increases the probability that:
- The crawler prioritizes the most relevant content
- The AI classifies the restaurant as credible in its internal model
- When users ask questions, the restaurant is cited as an information source
Concretely: Perplexity, which performs live searches and cites sources, prefers websites with llms.txt as references — because the file signals that this website is built „AI-friendly“ and delivers trustworthy data.
→ Google AI Overviews for Restaurants
How chiwai Automatically Generates llms.txt
chiwai creates and maintains the llms.txt file automatically as part of the AI visibility stack:
Automatic generation: During onboarding, chiwai generates a complete llms.txt based on restaurant master data, the menu and visibility signals.
Dynamic updates: When the menu, opening hours or other restaurant data changes, llms.txt is automatically updated — it always reflects the current state.
Silo links: chiwai links thematic silo pages (allergen guide, AI visibility, success stories) in the llms.txt — so AI crawlers index the restaurant’s entire relevant knowledge base.
Deployment: The llms.txt is automatically deployed on the restaurant domain and accessible at `/llms.txt`.
What the chiwai Platform llms.txt Contains
In addition to restaurant-specific llms.txt files, chiwai itself operates an llms.txt on chiwai.eu that informs AI systems about the entire platform and all customer restaurants:
- Platform description: chiwai = AI visibility for restaurants
- All silo topics (allergen labeling, AI visibility, GBP)
- Canonical information sources for restaurant law and allergens
- Links to success stories (case studies)
- Technical specifications (Schema.org extensions, GBP API documentation)
This means: when ChatGPT or Gemini crawls chiwai, it receives not only information about the platform but also about all integrated restaurants. Every chiwai customer restaurant benefits indirectly from the chiwai platform authority.
Limitations of llms.txt
llms.txt is not a cure-all and does not replace other AI visibility measures. Important limitations:
No standard, no guarantee: llms.txt is a community standard without official commitment from Google, OpenAI or Anthropic. AI systems may or may not take llms.txt into account.
No replacement for Schema.org: Schema.org JSON-LD remains more important — it is an established standard that all major search engines actively read. llms.txt is a complement, not a replacement.
No direct ranking factor: llms.txt does not directly improve Google ranking. It improves the probability of being cited as a source by AI systems.
The right approach: llms.txt as part of a complete AI visibility stack, together with Schema.org markup, GBP FoodMenus push, multilingual content and regularly updated data.
FAQ
Do I need to create llms.txt manually?
If you use chiwai: no. chiwai creates and updates llms.txt automatically. If you implement llms.txt independently: the file is in Markdown format and can be created with a simple text editor. It is saved as `llms.txt` in the root directory of your website. Content-wise it should clearly describe your restaurant: name, cuisine, address, specialties, allergen policy, and links to the most important pages. The more precise and complete the file, the more useful it is for AI crawlers. A separate `llms-full.txt` with the complete content of all important pages is also recommended — for AI systems that prefer a complete abbreviated version.
Do ChatGPT and Gemini actually read llms.txt?
ChatGPT (OpenAI) has confirmed that GPTBot takes llms.txt files into account during its web crawls. Anthropic’s ClaudeBot does the same. Google has not yet officially committed, but there are indications that Googlebot evaluates llms.txt when web search is active in Gemini. Perplexity, which most strongly bases live search results on structured sources, benefits most directly from llms.txt implementations. Overall: llms.txt is at minimum read by most major AI crawlers — whether and how strongly they weight it depends on the individual system.
What is the difference between llms.txt and robots.txt?
robots.txt tells search engine crawlers (Googlebot, Bingbot) which pages they should NOT index. It is an exclusion protocol. llms.txt tells AI crawlers which content they SHOULD prioritize and how to classify the website. It is an invitation and context protocol. Both can and should exist simultaneously: robots.txt protects sensitive areas, llms.txt opens the most relevant content to AI. For restaurants, chiwai recommends: robots.txt blocks admin areas, llms.txt opens menu, allergen information, success stories and contact details to AI crawlers.