What is the llms.txt file?
The llms.txt file is a text file designed to provide large language models (LLMs) with a concise, structured, and easily readable overview of the most relevant content available on a website.
The concept was proposed by Jeremy Howard with the goal of helping artificial intelligence better interpret publicly available online information, thereby improving the accuracy of the responses and summaries it generates.
Although the file is named llms.txt, this approach does not follow an official standard and may include formatting similar to Markdown. This is because the file is not intended for traditional crawlers like Googlebot, but rather to support the semantic understanding of LLMs, which can interpret semi-structured natural language text.
Like the more well-known robots.txt, it should be placed in the root directory of the domain. However, while robots.txt is used to control access for search engine crawlers, the llms.txt file is designed to guide language models directly toward selected content, typically in .md (Markdown) format - that is more readable and relevant than the entire HTML website.
Advantages and Disadvantages of Implementing the llms.txt File
Advantages
- Control and Clarity: You can specify exactly which content you want the AI to take into account.
- Simplification: Allows you to provide Markdown versions of your content, free from visual or technical noise (banners, scripts, layouts).
- Faster Interpretation: Helps models conserve tokens and better contextualize information.
- AI-First Optimization: Enhances the experience for users who rely on AI tools to search for or synthesize information.
Disadvantages
- It is not an official standard (at the moment): not all LLMs actively support the llms.txt file.
- Requires maintenance: like a sitemap, it needs to be updated periodically to remain effective.
- Does not replace robots.txt or SEO: it is a complementary tool, not a shortcut.
