► Analytics · Dashboard
Every component below is the same React code the production dashboard renders. Hover the chart, switch line/bar, watch the sparklines tick — this is the actual product surface, just running on synthetic data.
Every per-site dashboard includes these surfaces.
// the dashboard layer
You can't — GA4 filters out bot traffic by design and even when it doesn't, GPTBot doesn't fire JavaScript so the GA tag never loads. The way to track AI bot traffic is server-side: log User-Agent at the edge (Cloudflare Worker / Vercel middleware / nginx log shipper) and classify in a dedicated tool. Crawlytics handles that pipeline end-to-end with sub-millisecond classification overhead.
Depends what you need: (1) raw logs + grep — free, zero abstraction, hard to share with stakeholders; (2) Cloudflare's bot dashboard — free if you're on CF, limited to CF's bot list; (3) a dedicated tool like Crawlytics — purpose-built for AI bot identification (25+ signatures across 19 providers), AI referral attribution, and llms.txt/markdown serving from the same surface. We're biased, obviously — try the live demo to judge.
Ingest events are written to the aggregate tables within milliseconds of arrival. The dashboard re-queries on every load — refresh equals up-to-the-second.
Not yet. The data API is being made public once the JSON/CSV export endpoint ships. After that, any visualization layer can read the same data the UI renders.
Recharts for the main traffic and provider charts; a small custom SVG sparkline component inside each metric card.
It is generated deterministically per site so it reloads with the same shape — a mild upward trend, weekly seasonality, eight AI providers distributed by share. Indistinguishable from a real low-traffic site once you connect a real installer.
This page is part of Crawlytics.app. View all pages: llms.txt · llms-full.txt