ai-citation-trackerlisted

# AI Citation Tracker Defines a fixed query set per topic cluster, sends each query to ChatGPT/Perplexity/Gemini (and approximates AI Overview via SerpAPI), captures the citation links, and stores daily/weekly snapshots so brand visibility in AI search can be measured longitudinally — not just inferred from referral traffic. ## Why this matters in 2026 Referral traffic from AI engines (`chatgpt.com`, `perplexity.ai`, `copilot.com`) is a lagging indicator — by the time it shows up in GA4, the citation was made days ago. This skill is a leading indicator: - **Netsujo** sees 6-9 sessions/week from AI engines (May 2026 data) — small but trending up - **Miyakodeit** sees 12 sessions/90d from AI engines - The same brand mentions could equal 10x more downstream traffic if AI surfaces increase queries - Without tracking we cannot tell whether content changes (Phase 1 retitles, intro rewrites) actually improve AI citation share ## What this skill tracks For each topic cluster (defined in `config/clusters.yml`): 1. **Query set** — 10-30 representative search queries that real users would issue (e.g., "Web3 京都 開発会社", "DID とは", "もくもく会 京都") 2. **Engine matrix** — ChatGPT (via Playwright), Perplexity (API), Gemini (via Google AI Studio), AI Overview (via SerpAPI Google search) 3. **Citation extraction** — parse the response HTML/markdown for source links, normalize hosts, classify as own/competitor/3rd-party 4. **Snapshot storage** — JSON per (date, engine, query) in `reports/ai-cit