Think Lighthouse for AI citations. This audits web pages across seven research-backed categories to measure how well ChatGPT, Claude, Perplexity, and Gemini can extract and cite your content. It runs 30+ checks covering entity extraction, readability, answer capsules, structured data, and authority signals. The scoring methodology comes from Princeton's GEO research on what generative engines actually quote. Use it in CI/CD with fail-under thresholds, track score changes over time with diff mode, or get instant feedback via the MCP integration. No API keys, no external calls beyond fetching the target URL. Output as terminal pretty-print, JSON, Markdown, or standalone HTML reports.
claude mcp add --transport stdio agencyenterprise-aiseo-audit uvx aiseo-audit