Show HN: DataNav – A personal AI data analyst (github.com)

DataNav is an open-source, MIT-licensed personal AI data analyst that lets users aggregate data from services like Gmail, Google Calendar and YouTube Activity into a dedicated personal data lake and query it with natural language. It runs locally or in a hosted multi-user mode (Supabase for auth) and generates interactive visual reports on demand—examples include meeting counts, top email senders, or spending trends—while emphasizing user control: each user gets their own PostgreSQL database and data is not shared unless explicitly configured. Under the hood DataNav uses AI agents (codeAgent, reportingAgent, gEval and a default agent) configured to call large models (config references gpt-4.1 and gpt-5) and requires an OpenAI API key. Connectors use Google OAuth so you must create Google credentials and enable the appropriate APIs (Gmail, Calendar, YouTube) in Google Cloud. The stack is developer-friendly: Docker Compose boots PostgreSQL, a Next.js app and services; configuration is via datanav.config.ts plus environment variables (DB, Google, OpenAI, optional Supabase). The project is extensible—new connectors, UI components and agents can be added—but operators should weigh privacy and security (OAuth secrets, DB credentials, and third-party model usage/costs) when deploying.