aeo site audit

The problem

Websites are often invisible to AI agents not because of bad content, but because of accidental crawler blocks, JavaScript-only rendering, or lack of structured data. Developers struggle to know if their site is actually "AI-ready" or why assistants like Claude and Perplexity aren't citing their pages.

What it does

• Scans robots.txt for AI-specific crawler blocks and user-agent access.
• Validates JSON-LD and Schema.org markup for machine-readable claim extraction.
• Identifies actionable surfaces and WebMCP tool annotations for agent-driven task completion.
• Audits server-side rendering to ensure content is fetchable by non-JS extractors.
• Generates a severity-weighted AEO score based on reachability, understanding, and actionability.

Frameworks & tools

WebMCP, Schema.org, llms.txt, JSON-LD, and robots.txt protocols.

Why this beats prompting it yourself

Standard LLM prompts rely on stale training data regarding crawler behaviors and emerging specs like WebMCP. This skill uses a fixed, severity-weighted rubric that prevents "hallucinated" SEO wins and focuses on rigorous technical fetchability that a simple chat interface cannot verify.

Use cases

• Diagnose why a site is missing from Perplexity or ChatGPT search results.
• Audit a pre-launch staging site for AI discoverability and agent-actionability.
• Validate that Schema.org types correctly match the content on product pages.
• Identify "half-adopted" patterns like llms.txt or incomplete WebMCP annotations.

Known limitations

Static scans cannot see tools registered in external JS bundles or content behind heavy client-side rendering. Negative findings are reported as not found in fetched HTML, not as absolute absence.