Querypex — AI Data Analyst with Full Transparency

Querypex is an AI-powered data analyst built for students, small business owners, and lean teams who have large datasets and no dedicated analyst to make sense of them. Instead of requiring SQL or BI expertise, Querypex automatically profiles a connected dataset, surfaces the most important insights, and lets users ask follow-up questions in plain English — with every answer grounded in a real, executed SQL query, not a guess. The app connects to data four ways: CSV/Excel upload, a live Supabase database, a generic API connector, and Bright Data for live real-world pricing. Its AI chat, powered by Claude, always shows the exact query behind every answer, along with a confidence indicator, so nothing is a black box. Querypex also runs real machine learning — linear regression, logistic regression, and PCA — computed server-side rather than approximated by an LLM, with each model honestly reporting its own accuracy and limitations. The demo dataset is real: 1.5 million genuine, anonymized grocery transactions from the public Instacart Market Basket dataset, independently verified before loading. Users can also upload their own CSV, and the Dashboard and AI Chat adapt to it dynamically. Additional features include a searchable data explorer, CSV/PDF/PowerPoint export, and a downloadable Jupyter notebook that reproduces the full analysis pipeline in Python. Querypex was built end-to-end using native.builder's agent workflow: a Product Architect scoped the project into a formal PRD, a Task Planner broke it into ordered build tasks, and a Builder agent implemented each one — including real infrastructure problem-solving, like resolving a Supabase quota limit and fixing repeated query timeouts with materialized views. The result is a genuine application, not a thin wrapper around a chatbot — with a real data layer, a scale strategy, and complete transparency at every step.