Proposed by Jeremy Howard in 2024, llms.txt is a plain-text file designed for large language models (ChatGPT, Claude, Gemini, etc.). Published at the root directory (/llms.txt), it provides AI models with concise, structured information about a business or website.
It may look like a simple text file, but when filled out correctly, it can fundamentally change how AI models make sense of your business.
Why Is llms.txt Needed?
When large language models crawl web content, they face several challenges:
- Web pages contain dozens of irrelevant elements — navigation menus, ad zones, footer links — making it harder for the model to find the actual content.
- HTML markup increases the noise the model has to process.
- The most important information about a business is often spread across multiple pages.
llms.txt solves this: it gathers the most critical information about your business in one place, in a clean text format the model can easily digest.
What Goes Inside llms.txt?
An llms.txt file typically includes these sections:
- Business description: Who you are, what you do, who you serve — in two or three sentences.
- Target audience: Which customer profiles you serve.
- How it works: The steps of your service process.
- FAQ: The questions customers ask most often, with answers.
- Keywords and topics: Which queries you want to be associated with.
- Contact and links: Your main page, service pages, contact info.
All of this content is organized in Markdown format, with headings and short paragraphs. No complex HTML, no JavaScript — just clean text.
How Do ChatGPT and Claude Use llms.txt?
When ChatGPT performs web searches, and in the context of Claude Projects or Artifacts, the llms.txt file is read with priority when present. Instead of processing the entire page, the model uses this summary file to develop a fast, accurate understanding of the business.
The result: the model "knows" your business rather than "guessing" it. The risk of hallucination drops; the probability of accurate attribution rises.
The Difference Between llms.txt and robots.txt
robots.txt tells crawlers about restrictions: "Don't crawl this directory." It's a prohibition file.
llms.txt extends an invitation to AI: "Here's who I am. Here's what I do. Here's when you should mention me." It's an introduction file.
What Happens Without llms.txt?
When AI must crawl a business without an llms.txt file:
- It processes the entire page — including nav, footer, and cookie banners.
- It may misinterpret the business's actual purpose.
- It tries to reconcile conflicting information from different pages.
- Result: vague, incomplete, or inaccurate representation.
How Does mytrend.ai Create llms.txt?
mytrend.ai automatically generates the llms.txt file when you fill in your business profile, and publishes it at mytrend.ai/[business-slug]/llms.txt. The file is updated every time the profile is updated.
The file includes: business description, target audience, service steps, FAQ, keywords, and contact information — all in clean Markdown format, the way AI prefers to read it.
To see your own llms.txt file: llms.txt feature page →
Is llms.txt Enough on Its Own?
llms.txt is a powerful signal, but not sufficient alone. For best results, it needs to work alongside this trio:
- Schema.org JSON-LD — structured business data in machine-readable format
- IndexNow — instant update notification to Bing
- llms.txt — summary introduction file for large language models
When all three work together, your business is represented as accurately as possible across Bing, Google, and direct model context.
Further reading: What Is GEO? → | What Is Schema.org? → | What Is IndexNow? →