An AI-powered personal finance SaaS that gives users a dedicated financial copilot, streaming dashboard insights, and real-time anomaly detection. Built as a full production-grade application, not a demo.

Personal finance apps show you charts but never tell you what to do. Users want guidance and context-aware advice but cannot afford a financial advisor. Existing tools have no intelligence - just static dashboards that present data without any interpretation.

A full-stack SaaS with a dedicated AI Financial Chatbot that remembers session context and suggests follow-up questions. Five AI-generated dashboard cards stream independently: Health Score, Cashflow Forecast, Weekly Report, Spending Insights, and real-time Anomaly Detection with alerts.

The auth system uses production-grade JWT with access and refresh token rotation - the same pattern used in enterprise applications. The app supports 14 account types across 12 months of realistic transaction data, with multi-currency support and full budget tracking. Fully mobile-first and responsive.

AI insights are most valuable when contextual and data-grounded - generic advice adds no value.

Multi-agent AI cards that stream independently improve perceived performance significantly compared to a single blocking request.

Production-grade auth with JWT rotation is essential to demonstrate real-world readiness, even in a showcase project.

Next.js, PostgreSQL, Prisma, OpenAI API, JWT (jose), Framer Motion, Tailwind CSS