insighttrack

InsightsTrack provides a full‑featured analytics platform you can run on your own infrastructure, eliminating the need for third‑party services like Google Analytics. The system records events through a single lightweight script tag and stores raw events in PostgreSQL. A background worker syncs these records into DuckDB, which powers all analytical queries, delivering sub‑second response times for typical dashboards. Key features include: - Real‑time visitor monitoring, traffic source breakdowns, and funnel analysis. - Visual heatmaps, scroll‑depth tracking, rage‑click detection, and Core Web Vitals reporting. - An AI analyst called **Pulse** that answers natural‑language questions with live charts, tables and CSV exports, using Anthropic, OpenAI or Google models. - Multi‑user team management with custom roles, permission controls and invitation links. - Exportable reports, a drag‑and‑drop reporting studio, and a read‑only DuckDB SQL editor. - No cookies or fingerprinting; visitor IDs are pseudonymous and DNT/GPC is respected, making the platform GDPR‑friendly. Deployment options are flexible: - **Docker Compose** – one‑command setup that launches PostgreSQL, DuckDB, the Express API, the React dashboard, a demo site and pgAdmin. - **Manual Node.js** – run the Express backend and React frontend separately, connecting to an existing PostgreSQL instance. - **Cloud services** – backend on Railway/Render with a PostgreSQL add‑on, frontend on Vercel, Cloudflare Pages or Netlify. The tech stack consists of React 18 + Vite + Tailwind for the UI, Node.js 20 + Express for the API, PostgreSQL 16 for write persistence, and DuckDB for columnar analytics reads. Authentication uses JWTs and bcrypt, while caching is handled in‑memory with TTLs. Comprehensive documentation, API reference, and test suites are included. InsightsTrack is released under the MIT license and is free to self‑host, modify and scale.