Loading...
The GEO playbook Moonrank's GEO Agent uses to publish and maintain an llms.txt file — the one place AI crawlers and assistants read to understand what your site is.
Download the methodology, see how the agent drafts your file, or run it live inside the GEO Agent.
The llms.txt skill is the process Moonrank's GEO Agent follows to give AI assistants one canonical, machine-readable description of your business — what you sell, who it's for, and which pages are worth reading.
It's written as a SKILL.md, a plain-language instruction format that people and AI agents can both follow. Read it to understand what belongs in the file and what doesn't, adapt it for your own stack, or let the GEO Agent generate and maintain it for you.
Downloading the playbook is free with a Moonrank account. Keeping the file accurate as your pricing and positioning change is what the GEO Agent handles on an ongoing basis.
“ChatGPT keeps saying we're a social scheduling tool. We haven't been that for two years. Why?”
Your homepage copy leads with a tagline no crawler can parse, your pricing page is client-rendered, and the strongest plain-text description of you anywhere is a 2022 directory listing. Fix: publish an llms.txt that opens with the current positioning, mirror pricing and product as markdown, and link the three pages that state what you do now.
Paste your site and a one-paragraph description of what the business actually does today.
The GEO Agent finds the pages that explain the business and flags the ones crawlers read as empty.
A brand summary plus a prioritized, one-line-per-page index — written for a model, not a browser.
When pricing, positioning, or the page inventory changes, the agent regenerates the file.
Finds the pages that actually explain your product, pricing, and positioning.
Flags JS-heavy pages where AI crawlers get thin or empty text.
The one sentence assistants repeat when asked what you do.
Puts pricing and positioning first, where models weight them heaviest.
Clean .md versions of key pages where the HTML is hard to parse.
Catches the drift between your live site and what the file still claims.