llms.txt Generator & Validator

What is llms.txt?

This llms.txt generator writes the file that's meant to give AI models the same kind of guidance robots.txt gives crawlers. robots.txt controls who can crawl your site. llms.txt is a proposal for telling a model how to understand it. The spec (llmstxt.org) is short: an H1 with your site name, an optional one-line summary, and H2 sections listing the pages that matter. The citation block this generator adds is a Lumina convention on top of that, not part of the spec. No blocking, just context.

This llms.txt generator creates a valid file from your inputs. Fill in the fields, hit Generate, and drop it at your site root (example.com/llms.txt). Be clear-eyed about what it buys you. No search engine has documented reading the file, and Google said in June 2026 that Search ignores it. The readers today are documentation platforms and coding agents. Writing one takes five minutes, so being early costs you almost nothing.

How to create an llms.txt file

Four fields, one file. Fill in your site name, a one-line description, the 5-10 most important URLs you want AI models to know about, and a few lines on what your site does. Contact is optional. Hit Generate and save the output as llms.txt at your site root. The whole process takes under five minutes, and you can update the file whenever you launch a new landing page.

Validating an llms.txt file

Switch the tool to Validate, enter a domain, and it fetches the live file. The spec makes the H1 the only required element, so almost every file clears that bar. What trips them up sits around it: a server that answers unknown paths with index.html, so an agent asking for llms.txt gets your app shell. A Content-Type of text/html. Relative URLs in a file that agents fetch with no page context. Links that 404 because the page moved six months ago.

The validator keeps spec deviations apart from broken plumbing, and it only flags a bullet without a link when that bullet sits in a section which is otherwise a list of links. Free-form detail lists are part of the format, so they never count as errors. Hit Check every link and it requests each listed URL and reports the status code, which is the one check no spec document covers.

llms.txt vs. robots.txt

Different jobs, same location. robots.txt says who is allowed to crawl which paths. llms.txt offers a model a short description of your site and the pages you'd point it at. robots.txt is about access control. llms.txt is a suggestion an agent is free to ignore. The difference that matters in practice: every serious crawler fetches robots.txt, while llms.txt is only fetched by the handful of tools that chose to support it.

What to include in your llms.txt file

Keep it short. The spec sets no word limit, it just says the file should stay small enough to fit in an agent's context — the validator here starts warning past 1,200 words. Include your site name, a one-paragraph description of what you do, and a list of 5-10 important URLs with a short note after each. Add an "Optional" section for links an agent can skip when it needs a shorter context; that section name is the one convention the spec spells out. Skip press releases and anything that won't matter in six months.

Explore more tools

FAQ