Imagen estilo futurist punk en la que se muestran versiones robotizadas de robots.txt y llms.txt en el artículo "implementar llms.txt".
SEO EN

Thinking about implementing llms.txt? ? Read this first: You might not get the results you’re expecting

23 August, 2025

Over the past few months, the idea of implementing llms.txt has caught the attention of SEO professionals—and with them, a wave of social platforms.

The concept is straightforward: provide a file that tells large language models (LLMs) what content they can or can’t use for training, generating responses, or citing sources.

At first glance, it sounds like the natural evolution of the well-known robots.txt—a tool we’ve long relied on to guide search engines.

But in reality, llms.txt adoption is still in its infancy. Major players in AI development—OpenAI, Google, Microsoft, and Perplexity among them—haven’t officially integrated it into their workflows. Which raises a key question: Is it worth investing time and resources into implementing it right now?

In this article, I’ll take a closer look (or we will, if you’re reading along) at what advocates of llms.txt are proposing, its potential benefits, and why, from my perspective, it might not be a top priority just yet.

Table of contents: Considerations about implementing llms.txt

Table of contents

What is llms.txt?

The llms.txt file is an unofficial proposal aimed at creating a communication channel between content owners (typically websites) and developers of language models (like OpenAI).

In its original form, llms.txt is a Markdown text file placed at the root of a domain. It presents, in a human-readable format, a list of “preferred” URLs for LLMs to consult during inference (not training), along with contextual notes and potential exclusions.

The proposal was introduced on September 3, 2024, accompanied by formatting examples and conventions (Howard, 2024a; 2024b). Search Engine Land described it as a “treasure map” for AI—useful for highlighting documentation, FAQs, and key guides.

Its emergence reflects growing concerns around intellectual property and transparency in how data is used by AI. As more companies adopt chatbots, assistants, and AI-powered search engines, the idea of having control over what content is shared with them has sparked interest—as well as lawsuits over IP violations. Case in point:

However, the key difference is that robots.txt became a standard because giants like Google and Bing adopted it quickly. With llms.txt, that’s not the case—at least not yet. Neither OpenAI, Anthropic, Google, nor other major players recognize it as part of their protocols.

That leaves us in a sort of limbo: a tool with potential, but without the backing needed to ensure its implementation has any real impact on how AI interacts with our content.

Origins and Context of the Debate

The conversation around implementing llms.txt emerged from the exponential growth of language models and the way they’re reshaping how users consume information. Platforms like ChatGPT, Claude, and Gemini can synthesize massive volumes of content and deliver answers in seconds—raising concerns among creators and media companies about copyright violations and declining website traffic.

The question is: if LLMs are using our content to train or respond to queries, shouldn’t we have a say in what gets used and what doesn’t? That’s where llms.txt comes in.

The idea was first proposed by independent developers and a few AI ethics experts, who saw in this file a chance to offer a simple, non-invasive, and technically easy-to-implement mechanism.

But while some see it as a step toward transparency, others question its viability. For instance, since it’s neither mandatory nor backed by international standards, there’s no guarantee AI companies will actually respect its directives. Let’s not forget, some LLMs don’t even honor robots.txt instructions. Looking at you, Anthropic and OpenAI.

Even if it were adopted, there would still be technical limitations: LLMs may have already been trained on content before the file existed, and removing that knowledge isn’t exactly straightforward.

This context is crucial to understanding why, despite the initial enthusiasm, adoption of llms.txt remains minimal.

implementing llms.txt

Advantages of llms.txt (According to Its Advocates)

Supporters of implementing llms.txt argue that this file could offer several key benefits:

  1. Content control: It would allow site owners to decide which parts of their website can be used to train LLMs.
  2. Transparency: Since it’s publicly accessible, any user could view a site’s policy regarding AI usage.
  3. Simple implementation: Technically, it would be as easy to configure as a robots.txt file, as noted by Mintlify and ahrefs.
  4. Potential to become a standard: If adopted by major companies, it could be integrated into SEO and content best practices.

For websites hosting highly sensitive information—such as academic research, proprietary data, or commercially valuable content—the ability to restrict its use by AI models is undeniably appealing.

Advocates also highlight that llms.txt could be useful for AI-oriented SEO experiments, allowing companies to analyze which types of content gain more visibility in conversational responses.

However, all these benefits remain theoretical. Without formal backing from key players in the ecosystem, any impact will be marginal. That’s why, in the next section, I’ll explain why—at least from my perspective—implementing llms.txt today may not be the best investment of our most valuable resource: time.

Why I Believe Using llms.txt Isn’t Worth It Right Now

Despite the advantages mentioned earlier, my stance is, at best, skeptical. So if someone asked me whether I recommend implementing llms.txt, my answer would be NO—for five main reasons:

  1. Lack of official adoption: None of the major AI players have integrated it into their processes. Implementing it now is like putting up traffic signs no one follows.
  2. No SEO benefit: Unlike robots.txt or sitemaps, it has no impact on indexing or organic ranking.
  3. Technical limitations: LLMs can’t “unlearn” content they’ve already been trained on. Blocking access now (or in the future) doesn’t change the past.
  4. More effective alternatives: Structured data, schema markup, and sitemaps already serve key functions for guiding bots and AI, as noted here.
  5. SEO priorities: Technical optimization, site speed, and content quality remain far more critical factors.

In short, llms.txt is an interesting idea—but premature. Allocating resources to configure it may create a false sense of control without delivering any tangible benefits.

Comparison with Other SEO Tools

When comparing llms.txt with well-established tools, it becomes clear that there are already more effective options than adopting a standard that has yet to gain recognition:

  • Sitemaps: Tell search engines which pages exist and how frequently they should be crawled.
  • Structured data: Help search engines understand content and display rich results.

All of these tools already influence how AI and search engines interpret content. llms.txt, by contrast, has no proven impact on visibility or ranking.

That’s why, before investing time in implementing llms.txt, I’d recommend reinforcing these technical elements and optimizing content quality to better address user queries.

My Recommendations for the Present and Near Future Regarding llms.txt

In an ideal scenario, llms.txt could evolve into an official standard, backed by leading companies and tech governance bodies.

However, until that happens, my recommendation is to prioritize:

  • Technical optimization: Reduce unnecessary JavaScript and CSS; aim for faster, cleaner websites.
  • Question-oriented content: LLMs and search engines value clear, direct answers.
  • Structured data and sitemaps: These remain the most effective bridge between your site and search/AI platforms.

If llms.txt becomes a recognized norm in the future, it will be worth evaluating. For now, time and resources should be focused on what delivers proven impact.

Final Thoughts

For now, my conclusion is simple: there’s no need to rush into implementing llms.txt. It’s better to focus on optimizing your site, creating high-quality content, and improving user experience.

If this topic interests you and you’d like to continue the conversation, feel free to connect with me on LinkedIn, Bluesky, and X.com. Of course, you are also invited to keep exploring SEO topics on my blog.

FAQ about implementing llms.txt

1. What is llms.txt?

A proposed file designed to tell language models which content they’re allowed to use for training or generating responses.

Will it help me rank better on Google?

No. As of now, it has no impact on SEO.

Do AI companies respect it?

Currently, none have officially adopted it.

Are there better alternatives?

Yes: sitemaps, structured data, and schema markup.

Could it be useful in the future?

Potentially, if major companies adopt it as a standard.

References

You May Also Like

impact of Atlas on PPC campaigns
The Impact of Atlas on PPC Campaigns: How It Will Affect Your Costs and Digital Security
26 October, 2025
ai summary link generator
AI Summary Link Generator | Boost Your Organic Visibility in LLMs!
16 August, 2025
Google Search Central Live MX 2025
My takeaways from the Google Search Central Live MX 2025
30 September, 2025

Leave a Reply