Description

FinSimple is an AI-powered personal finance web application that transforms bank statements into actionable financial insights through natural conversations. Instead of overwhelming users with spreadsheets, it acts like a best friend who simplifies finances, calls out poor spending habits, celebrates smart decisions, and generates personalized reports, all while keeping financial data private through client-side processing.

How I Made It

• Built with Next.js 16 (App Router) and TypeScript for a modern full-stack architecture.
• Designed a responsive glassmorphism UI using Tailwind CSS v4 with dark and light mode support.
• Implemented client-side parsing for PDF, CSV, XLSX, and DOCX bank statements.
• Developed deterministic financial analysis for transaction categorization, spending insights, subscription detection, and monthly projections.
• Integrated Claude, Gemini, OpenAI, and OpenRouter with automatic AI provider fallback.
• Added real-time streaming AI responses using Server-Sent Events (SSE), interactive Chart.js reports, and Upstash Redis rate limiting.

Challenges Faced

• Reconstructing transaction tables from PDFs with inconsistent layouts and formatting.
• Building a reliable multi-provider AI fallback system while maintaining consistent responses.
• Combining deterministic financial calculations with AI-generated insights without introducing hallucinations.
• Processing large financial documents efficiently while ensuring complete client-side privacy.

Key Wins

• Built a conversational finance assistant that makes budgeting engaging instead of intimidating.
• Ensured uploaded financial documents never leave the user's browser.
• Successfully integrated multiple AI providers with seamless AI fallback.
• Delivered automated transaction categorization, interactive reports, and personalized financial guidance in a polished, production-ready application.