Schema and llms.txt for AI

The technical layer of GEO: access for AI crawlers, llms.txt and structured data for models.

Schema and llms.txt for AI

Before an AI model can cite a page, it has to read and understand it. This GEO layer handles the technical accessibility of content for AI crawlers and the signals that help models recognize what a page is about.

Access for AI crawlers

/seo geo your-domain.com

The first step is checking whether the model bots are allowed in. The console verifies robots.txt rules and Content-Signal headers for the most important agents:

llms.txt - a content index for models

An llms.txt file in the root is a content map written for models: a list of the most important pages with a short description, plus an optional llms-full.txt with the full Markdown content. The console generates both and keeps them consistent with the sitemap.

• 0 - a concise index: titles, URLs and one-sentence descriptions of the key pages.
• 0 - the full Markdown content for models that read the whole document.
• Mirror 0 - every subpage has a Markdown version at a .md URL, ready to be cited.

Structured data for models

Schema.org is not only for Google rich results - structured data also helps models link entities (brand, author, product) and recognize authority. The types with the most weight for GEO:

• 0 and 1 - the identity of the brand and authors (an E-E-A-T signal).
• 0 and 1 - question-answer pairs that models carry straight into answers.
• 0 with author, datePublished, citation fields - context and freshness.