SpendWise

AI-powered receipt management and tax claim assistant for smarter personal finance.

Comment

Inspiration

I wanted to solve a problem that many people experience every day: managing receipts and organizing personal finances. People often collect paper receipts, forget to record their expenses, and struggle when preparing documents for tax claims. I wanted to build a solution that makes this process simple by combining AI with an intuitive user experience. Rather than creating another budgeting application, I focused on reducing manual work and letting AI handle repetitive tasks while keeping users in control of the final decisions.

What it does

SpendWise is an AI-powered personal finance platform that helps users: Scan receipts using a mobile phone or upload images and PDFs. Extract receipt information automatically using OpenAI. Organize expenses into structured financial records. Suggest tax claim document categories for Malaysian users. Search and manage receipts efficiently. Export reports for record keeping. Users can always review and edit AI-generated information before saving, ensuring transparency and accuracy.

How we built it

The application is built using: React + TypeScript for the frontend. Node.js + Express for the backend. PostgreSQL + Prisma ORM for data storage. Docker for deployment. Cloudflare + Nginx for secure hosting. OpenAI GPT-5.6 for receipt understanding and structured data extraction. During development, I relied heavily on Codex. Instead of only generating code, Codex became my development partner. I used it to implement new features, refactor existing modules, improve UI consistency, debug issues, write tests, and safely deploy updates without breaking production. GPT-5.6 powers the intelligent receipt understanding, while Codex accelerated the engineering workflow throughout the project.

Challenges we ran into

Building a receipt management system turned out to be much harder than expected. Receipts come in many different formats, layouts, and image qualities. Some are blurry, incomplete, or contain inconsistent information. Another challenge was ensuring AI suggestions remained trustworthy. Rather than allowing AI to make financial decisions automatically, I designed the system so users can always verify and edit extracted information before it becomes part of their records. Balancing automation with user control became one of the most important design decisions.

Accomplishments that we're proud of

I'm proud that SpendWise became more than just a prototype. The application includes: AI-powered receipt extraction. Intelligent expense organization. Tax claim document assistance. Responsive web interface. Secure production deployment. Modern dashboard and reporting features. Most importantly, the application solves a real-world problem that people face regularly.

What we learned

This project taught me that AI is most valuable when it assists users instead of replacing their decisions. I also learned how much faster development becomes when combining GPT-5.6 with Codex. Instead of spending hours implementing repetitive tasks, I could focus on improving product quality, refining the user experience, and validating each feature before deployment. AI significantly increased my productivity, but careful testing and human review remained essential throughout development.

What's next for SpendWise

This is only the beginning. The next version of SpendWise will include: Native Android application. Native iOS application. Family expense sharing. AI budgeting recommendations. Smarter financial insights. Subscription management. Bank transaction integration. Multi-language support. My long-term goal is to build SpendWise into an intelligent personal finance assistant that helps users spend smarter, stay organized, and save time.

Built With

Share this project:

Updates