llms.txt: the file that helps AI read your site

The llms.txt file is a text file placed at the root of a website that provides language models (ChatGPT, Claude, Perplexity) with a structured summary of the site and its main pages. It is easy to set up and costs nothing. But it is not a priority if your site is not already crawlable by AI bots and properly structured (heading tags, schema.org, quality content). llms.txt is an optional layer that comes after the fundamentals, not before.

AI answer engines (ChatGPT, Claude, Perplexity, Google AI Overviews) are playing an increasing role in how people find information. For businesses, a new question is emerging: how do you make sure AI understands your site and cites it in its responses? This is the subject of AEO (Answer Engine Optimization), a discipline that is gradually taking shape alongside traditional SEO.

In this context, a new standard is emerging: llms.txt. It is a text file, comparable to robots.txt, but designed specifically for language models. It provides AI with a structured summary of your site: who you are, what you offer, and what your main pages are. The concept is simple. But does it actually work? This article takes stock, without overselling: what it is, how to set it up on Webflow, and most importantly what it is actually worth in 2026.

‍

What is llms.txt?

The llms.txt file is a text file placed at the root of a website (accessible at mysite.com/llms.txt) that provides LLMs (Large Language Models) with a structured summary of the site. The principle is similar to robots.txt, which tells search engine crawlers which pages to explore and which to ignore. But where robots.txt addresses crawlers, llms.txt addresses language models.

The file typically contains the site name, a short description of its activity, the main pages with their URLs and content descriptions, and optionally additional resources (guides, reference articles). The format is simple Markdown, readable by both humans and machines. The standard was proposed by Jeremy Howard, founder of fast.ai, and is the subject of an open specification available at llmstxt.org.

In short, llms.txt is a Markdown text file placed at the root of a website that provides language models with a structured summary of the site and its main pages. It is the equivalent of robots.txt, but designed for AI instead of traditional search engines.

‍

How to configure llms.txt on Webflow

The llms.txt file follows a simple Markdown format. Here is an example structure for a web agency site:‍

# BeBranded> Webflow agency specializing in branding, design and development of high-performing websites.
  ## Main pages
  - [Webflow Agency] (https://www.bebranded.xyz/agence-webflow): Webflow site design and development services using the Client-First methodology.
  - [Portfolio](https://www.bebranded.xyz/projects): Portfolio of projects completed by the agency.
  - [Blog](https://www.bebranded.xyz/blog): Articles on Webflow, SEO, design and no-code.
  - [Contact](https://www.bebranded.xyz/contact): Contact form for an initial conversation.
  
  ## Resources
  - [Complete Webflow guide](https://www.bebranded.xyz/blog/what-is-webflow): Understanding what Webflow is and what it enables.
  - [Webflow SEO checklist](https://www.bebranded.xyz/blog/optimize-your-webflow-site-with-our-seo-checklist): All SEO optimizations to configure on Webflow.

Webflow has announced support for llms.txt. Configuration goes through the project settings (Project Settings > SEO) where the file can be added directly. Alternatively, the file can be created manually and hosted as a static page accessible at the /llms.txt URL. The content should be updated when the site's main pages change (new service added, new blog section, or updated descriptions).

A llms-full.txt variant can contain more detailed information about each page: longer summaries, subpages, content categories. This extended version is useful for sites with a large volume of content (blog with dozens of articles, product catalog) but remains optional.

‍

llms.txt is not a silver bullet

It is important to be honest about the current state of this standard. In 2026, there is no concrete proof that LLMs systematically use the llms.txt file to decide which sites to cite in their responses. The impact is real but difficult to quantify.

None of the major LLM providers (OpenAI, Anthropic, Google) have officially confirmed that llms.txt influences the way their models select and cite sources. The standard is emerging: it is adopted by a growing number of sites (especially in tech and SaaS), but it is far from being a universal standard like robots.txt or sitemap.xml.

That said, setting it up costs nothing (a few minutes of work) and presents no risk. It is a reasonable "why not," not a strategic priority. If you already have a well-structured, well-indexed site, adding llms.txt is a useful step. If your site has fundamental SEO, content, or performance issues, solving those problems will have an incomparably greater impact than any llms.txt file.

‍

The real prerequisite: a site crawlable by AI

This is the most important message of this article. Before dealing with llms.txt, you need to make sure AI bots can access your content. Without this foundation, llms.txt is completely pointless.

The first point to check is the robots.txt file. If your robots.txt blocks AI crawlers (GPTBot for ChatGPT, ClaudeBot for Claude, PerplexityBot for Perplexity), your content is invisible to these models. Adding an llms.txt file to a site that blocks AI bots is like putting a welcome sign in front of a locked door. The robots.txt must allow access for AI bots on the pages you want to make visible. On Webflow, the robots.txt is configurable in Project Settings > SEO > Robots.txt.

The second point is content structure. If your site has no coherent heading hierarchy (H1, H2, H3), no alt attributes on images, no schema.org structured data, LLMs understand your content poorly. Semantic markup is what allows AI to distinguish a heading from a paragraph, a product description from a testimonial, a blog post from a service page. Without this structure, the LLM sees raw text without context.

The third point is content quality. Google's E-E-A-T criteria (expertise, experience, authoritativeness, trustworthiness) apply to language models as well. Generic, superficial, or outdated content has little chance of being cited, with or without llms.txt. Useful, expert, up-to-date, and clearly structured content has the best chance of being understood and recommended by AI.

The real priority for AEO, in order: a site technically accessible to AI bots (open robots.txt), content structured with complete semantic markup (heading tags, schema.org, alt attributes), quality content that demonstrates expertise (E-E-A-T), and solid performance (Core Web Vitals). llms.txt comes after all of that, as an optional layer that complements an already solid foundation. For comprehensive technical optimization coverage, our Webflow SEO checklist details every point to verify.

‍

Conclusion

Setting up an llms.txt file takes only a few minutes and costs nothing. It is a simple step that can help language models better understand your site and services. For a site that is already well-structured and well-indexed, it is a logical complement to integrate.

But the real priority for being visible in AI responses is what comes before llms.txt: a site crawlable by AI bots (open robots.txt), content structured with complete semantic markup (heading tags, schema.org), quality content that demonstrates expertise, and solid performance. Without this foundation, llms.txt is just another file on a server, with no real impact.

BeBranded integrates AEO best practices into every Webflow site, from schema.org markup to the llms.txt file, including semantic structure and robots.txt configuration. If you want your site optimized for both traditional search engines and AI answer engines, you can get in touch with us for an initial conversation.