Do You Need an llms.txt File? A Straight Answer for SaaS

llms.txt is being sold hard as a 2026 AI-visibility must-have — a file you add to your site so AI engines understand and cite your content. The short, honest answer for most SaaS sites: no, you don't need one for AI search visibility, because no major AI engine reads it for that. Here's the evidence, and the one situation where it's genuinely worth adding.

What llms.txt is supposed to do

llms.txt is a proposed markdown file at your site root that gives large language models a curated, clean guide to your most important content, the idea being that AI engines would read it and represent your site more accurately in answers. It sounds reasonable, and it's easy to add, which is why it's marketed everywhere.

The problem isn't the idea. It's that the engines it's meant for don't use it.

What the data actually shows

As of early 2026, the evidence against llms.txt as an AI-search tactic is blunt:

• No major AI company reads it in production. OpenAI, Google, Anthropic, Meta, and Mistral have none of them publicly committed to reading or acting on llms.txt in their live systems (Codersera).
• Google explicitly declined it. In mid-2025, Google's Gary Illyes said Google doesn't support llms.txt and isn't planning to, and John Mueller compared it to the long-discredited keywords meta tag (Codersera).
• The crawlers ignore it. Across one monitored window of over 500 million AI bot visits, only 408 requests targeted llms.txt directly — statistically negligible (Codersera).
• Almost every file sits unused. In a large adoption study, roughly 97% of llms.txt files received zero AI requests (PPC Land).

Adoption has grown — a study of 300,000 domains found about 10% now have the file, but adoption isn't usage. Ten percent of sites added a file that the engines it targets don't fetch.

Why the engines don't need it

The reason is simple: the AI engines already have a way to understand your site, and it's the same one search has always used. Google's AI Overviews draw from your organic rankings. Perplexity runs a live web crawl and reads your actual pages. ChatGPT leans on established sources and its training data. None of them needs a special hand-curated file, they read the web the way crawlers already do, governed by robots.txt, which is the file that actually controls AI crawler access. Both OpenAI and Anthropic direct site owners to robots.txt for crawler management, not llms.txt.

The one case where llms.txt IS useful

There's a real use case, and it's just not "AI search visibility." The things that actually fetch llms.txt today are developer tools: IDE agents like Cursor, Claude Code, and Copilot, MCP servers, and in-product AI assistants that help users work with a tool's documentation.

So if your SaaS product has developer documentation — an API, an SDK, a technical reference that developers use AI assistants to navigate — an llms.txt (or the fuller llms-full.txt) pointing to your clean docs can genuinely help those coding agents surface your documentation accurately. That's a documentation-experience win, not a search-ranking one.

If you don't have developer docs, there's no meaningful case for it right now.

What to do instead

If your goal is AI search visibility, spend the effort where the engines actually look:

1. Keep ranking. AI Overviews pull almost entirely from top organic results — your SEO is the real lever.
2. Get your robots.txt right. That's the file that controls AI crawler access. Make sure you're not accidentally blocking the crawlers you want.
3. Write extractable, current content that engines can quote, the actual tactics that earn citations.
4. Build topical authority with clusters, which correlates strongly with getting cited.

None of that is llms.txt.

The bottom line

Add an llms.txt if you ship developer documentation and want coding assistants to navigate it well. Skip it if you're adding it because a blog post said it boosts AI search visibility, the engines don't read it, the data is unambiguous, and the time is better spent on the fundamentals that actually feed AI answers. For the fuller picture on what does and doesn't move the needle, see our guide to GEO vs SEO.