cd ~/projects
CASE STUDY2025

WealthWise

AI-Powered Financial Assistant

Voice-first personal finance PWA with AI budget recommendations and 92% expense categorization.

Next.jsMongoDBGemini APIPWALIVE
04 / 05

WealthWise is an installable PWA that makes expense tracking frictionless: speak an expense, and it is logged, categorized and reflected in AI-generated budget recommendations. Gemini powers the insight layer; offline-first design keeps the app usable without connectivity.

01THE PROBLEM

Expense tracking fails because it is friction — manual entry gets abandoned within weeks. And generic budgets ignore how people actually spend. The app has to capture spending at the speed of speech and turn it into personalized guidance.

02THE APPROACH
  • Voice-enabled expense capture: natural-language entries are parsed and auto-categorized across 12 spending categories.
  • Gemini API generates budget recommendations from actual spending patterns rather than static templates.
  • PWA architecture (service worker, IndexedDB) for offline logging with background sync.
  • Next.js + MongoDB backend for cross-device history and analytics.
03THE OUTCOME
  • 92% expense-category detection accuracy across 12 spending categories.
  • Live in production at mywealthwise.tech as an installable PWA.
  • Voice-first entry removes the friction that kills manual trackers.