FinanceGPT

Inspiration

I was inspired to build FinanceGPT after seeing how many people struggle to manage their finances effectively. Most individuals lack the tools to track spending, identify wasteful habits, and save consistently. This lack of visibility often leads to poor financial decisions, unnecessary expenses, and long-term financial stress.

FinanceGPT was created to solve this problem by offering a smart, user-friendly platform that provides insight, structure, and automation. By combining AI-driven recommendations with intuitive design, it empowers users to take control of their finances and build better money habits.

What It Does

FinanceGPT is an AI-powered personal finance management tool that helps users track income, expenses, and budgets while offering personalized insights and financial advice. The platform includes a sleek, responsive dashboard where users can monitor their financial activity in real time and set saving goals.

Key features include:

• Bank statement import from PDF, Excel, and CSV files
  
• AI-based transaction categorization with 90%+ accuracy
  
• Duplicate detection and recurring payment recognition
  
• Custom rules, notes, and transaction filters
  
• Text and voice chat with an AI financial advisor
  
• Spending insights, trend detection, and financial health scoring

With FinanceGPT, users get clarity, automation, and proactive suggestions—helping them save more, spend wisely, and achieve financial freedom.

How I Built It

I built FinanceGPT as a full-stack application using:

• Frontend: React + TailwindCSS
  
• Backend: Supabase (authentication, real-time database, file storage)
  
• AI Integration: OpenAI for financial insights, ElevenLabs for voice support
  
• File Parsing:
  
  • Papaparse for CSV
    
  • PDF.js for PDF
    
  • SheetJS for Excel
    
• Environment: Developed and deployed using Bolt.new

Challenges I Ran Into

• Database Setup: I initially faced issues when manually configuring the database. Switching to Supabase via Bolt.new resolved the issue.
  
• Speech-to-Text: Real-time and accurate voice transcription was tricky to implement, especially on mobile.
  
• Mobile Responsiveness: Some UI components didn’t render properly on smaller screens, requiring a mobile-first redesign.

Accomplishments That I'm Proud Of

• Built a fully functional, production-ready AI finance app
  
• Created a robust multi-format bank statement importer
  
• Integrated a voice-based financial assistant
  
• Designed a responsive, modern UI with animations and data visualization
  
• Delivered a solution that solves real-world personal finance challenges

What I Learned

• Leveraging Supabase for secure, scalable backend architecture
  
• Integrating AI into financial workflows, including transaction categorization
  
• Implementing voice AI and optimizing speech-to-text across devices
  
• Adopting a mobile-first UI/UX strategy for better accessibility

What’s Next for FinanceGPT

• Integration of Stripe for billing and subscriptions
  
• Integration of Resend for email notifications and verifications
  
• Live bank account integration using APIs (e.g., Plaid, Mono)
  
• Multi-currency support and localized financial rules
  
• Predictive analytics and goal-based automation

Built With

• bolt.new
• supabase
• tailwindcss-and-react