🚨 An autonomous personal finance agent that understands, analyses, and explains UPI transactions automatically, transforming raw bank statement data into actionable financial insights using agentic AI.
- Problem Statement Title: Agentic AI
- Category: Software
UPISensei provides an intelligent, agentic workflow that extracts UPI data from multiple sources and converts it into structured insights. Users can upload monthly bank statements or scan UPI QR codes during payment time. The agent automatically identifies transaction details, assigns categories, and generates summaries and visual explanations.
- Upload bank statements or scan UPI QR codes to extract embedded details automatically.
- AI-driven categorization with user-editable tags.
- Clean, visual summaries of spending patterns and history.
- Offline-first architecture ensuring full privacy and zero external data sharing.
- Removes the need for manual tracking or spreadsheets.
- Converts raw UPI data into meaningful structured insights.
- Reduces categorization errors through continuous learning from user input.
- Helps users maintain financial clarity without complex budgeting apps.
- QR-based logging allows instant capture of transaction info, even before statements arrive.
- Agentic AI explains patterns and generates summaries automatically.
- 100% offline data processing builds strong trust and adoption.
- Adaptive tagging system evolves with each user’s habits.
- Parsing bank statements and UPI QR formats is technically mature and achievable.
- High nationwide UPI usage ensures a broad and relevant user base.
- Offline-first approach keeps costs low and simplifies DevOps.
- No server infrastructure improves long-term sustainability.
- Users prefer automation over manual entry.
- Visual, gamified summaries reduce cognitive load.
- Private, device-side processing increases confidence.
- Accuracy improves the more the user interacts with the system.
- Bank statements differ widely across institutions.
- QR data may sometimes be incomplete depending on the source.
- Initial categorization may require correction to improve model accuracy.
- Expectations may grow beyond the intended core features.
- Build flexible parsers that support multiple bank formats.
- Provide manual fallback when QR metadata is missing.
- Apply incremental learning from user adjustments.
- Communicate scope clearly to prevent unnecessary feature requests.
- NPCI documentation on UPI QR payload structure.
- RBI reports on digital payment adoption in India.
- Consumer behavior research on habit-forming finance tools.
- Studies showing visual spending insights increase user retention.
- India handles billions of UPI transactions monthly.
- Most users do not maintain structured financial tracking.
- Offline personal finance tools show higher long-term engagement rates.
- Users prefer clear visual summaries over spreadsheets.
- Automated QR/statement logging reduces friction significantly.
- Privacy-first tools experience lower churn.
Strong user demand + high technical feasibility + private offline processing = UPISensei solves a real, validated need. Gamified, visual insights match exactly what users expect from a modern UPI-based financial assistant.
- Auto-categorization accuracy rate.
- Number of statement or QR imports per user.
- Monthly active usage and retention.
- Time saved versus manual logging.
- Students gain daily spending clarity.
- Professionals track monthly budgets easily.
- Small businesses log UPI payments instantly.
- Anyone using UPI gains a private, automated finance assistant.
| Platform | Supported? |
|---|---|
| Web (any browser with JS functionality) + Fully Responsive | ✅ |
| Android (non-natively through WebView) | ✅ |
- Frontend: React, Next.js, Typescript, TailwindCSS, shadcn/ui
- Backend: Node.js/Python with AI microservices
- Database: MongoDB/PostgreSQL
- AI/ML: TensorFlow, PyTorch for predictive analytics
- APIs Integration: ...
- Hosting: Vercel/Firebase/AWS/GCP cloud infrastructure
-
Clone & Download the Repo
-
Install NodeJS on your system.
-
Open the project in your preferred IDE.
-
Run in Terminal to Install all dependancies:
npm i
-
Get all api keys in env.template as set them in your env:
-
Run in Terminal to Start Development Server:
npm run dev
- Soon.
| # | Team Member | Role | GitHub Profile |
|---|---|---|---|
| 1 | Fareed Ahmed Owais | 🎯 Team Lead | 🔗 FareedAhmedOwais |
| 2 | Abdur Rahman Qasim | 🔎 Research Engineer | 🔗 Abdur-rahman-01 |
| 3 | Mohammed Saad Uddin | 🚀 Full-stack + AI/ML Developer | 🔗 saad2134 |
| 4 | MD Shoaib Ahmed | 🔗 Backend Support | 🔗 XSHOAIB |
| 5 | Mir Ayan Ali | 🧩 Backend Engineer | 🔗 mirayanali5 |
| 6 | Mohammed Abdul Mugees | 💼 Solutions Engineer | 🔗 mug3es |
Developed with 💖 for the Hacknovate 2025, with heartfelt thanks to Methodist College of Engineering & Technology for the opportunity to build and innovate.
#WebApp #Hacknovate-2025 #fintech #upi #digital-payments #qr-payments #expense-tracker #personal-finance #finance-app #banking #transaction-analysis #data-extraction #ocr #receipt-scanner #categorization #automation #shoaib