How to Run GEO/AEO Audits Using Content Hub MCP
Run a GEO/AEO audit from a single chat message on your own domain, a single page, or any company's documentation over Content Hub MCP.
A GEO/AEO audit measures how citeable a domain or page is to AI models. It checks structured data, answer-first formatting, freshness signals, and whether AI crawlers can read the page.
Connect Claude Desktop to Infrasity's Content Hub over MCP (see the Content Hub MCP setup guide). You can then run an audit from a single chat message. The input determines the audit type:
- A bare root domain triggers a full crawl.
- A specific page path triggers a single-page audit.
- A domain that is not an Infrasity client triggers a standalone live audit.
The following sections show one example of each.
Run a domain-wide audit
Request a domain-wide audit of infrasity.com using Content Hub MCP.
Claude states the scope before it runs the audit: a root domain triggers a full crawl across roughly a dozen pages and dozens of checks. It then returns an overall score and a category breakdown covering content structure, GEO and AEO readiness, technical SEO, structured data, authority, and content quality.
Claude then reports strengths, such as a well-structured llms.txt, all AI crawlers allowed, and solid internal linking. It follows with the largest problems ranked by impact. These typically include weak content quality and gaps in authority signals, such as missing author attribution or freshness dates.
The report ends with pages that have structural issues (missing H1s, duplicate H1s, thin content) and a quick-wins list of the highest-impact fixes. Claude offers to pull citation data to relate the audit to actual AI-citation performance.
Run a single-page audit
Request a single-page audit of https://www.infrasity.com/services/developer-marketing-agency.
Claude notes that a specific path triggers a single-page check instead of a crawl. It returns a score by section: semantic answer coverage, crawlability, schema signals, internal linking, and content structure.
Claude lists strengths: fully covered semantic answers, flawless crawlability, and solid internal linking. It then lists the key failures to fix, which usually fall into three groups:
- Freshness, such as no visible last-updated date.
- Opening copy that does not state the direct answer.
- Headings that are not phrased as questions.
The report ends with trust and authority gaps (no measurable outcomes, no client logos), missing schema or breadcrumb structure, and a quick-wins list. Each quick win identifies the exact heading or opening line to rewrite. Claude also offers to compare the page with a competitor's.
Run a docs audit for external domains
You can audit domains that are not Infrasity clients. Provide a documentation URL, such as https://openrouter.ai/docs/quickstart, and request a docs audit.
Claude states that the audit runs as a standalone live audit with no linked company profile, because the domain is not an Infrasity client. It scores the documentation by section: AI/LLM discoverability, technical crawlability, content completeness, and agent tooling.
Strengths include a well-configured llms.txt and robots.txt, plus documented code examples and error codes. Key problems typically include:
- Weak content quality, such as a quickstart page too thin to guide a user to a working state.
- Pages that return errors to AI crawlers although the documentation is otherwise indexable.
The report ends with structural gaps, such as no internal cross-linking between documentation pages and no MCP server endpoint listed. The quick-wins list ranges from expanding the quickstart to splitting an oversized llms-full.txt file. Claude offers to run a full domain audit or compare against a competitor's documentation.
One MCP connection supports three audit modes: a full domain crawl, a single-page check, and a standalone docs audit for any company.
To audit your own domain or a page before publication, open app.infrasity.com and run an audit.
