Finpulse

Finpulse turns monthly salary into a live daily spending allowance, tracking UPI burns and quick-commerce leaks to end month-end anxiety for young India.

Comment

Finpulse ⚡

Real-Time Daily Budgeting, Burn Pace & Live Banking Intelligence

PayPal Hackathon Next.js 16 Capacitor Android

💡 Inspiration

Every month, millions of young professionals, remote contractors, and digital earners experience the exact same illusion:

"My account has ₹1,00,000 / $3,000 today — I'm wealthy!"

Then reality hits:

  • By Day 18, the balance has mysteriously vanished into a haze of micro-transactions: 10-minute grocery apps (Zepto, Blinkit, Instamart), late-night food deliveries (Swiggy, Zomato, Uber Eats), invisible OTT subscription charges, and dozens of frictionless digital wallet & UPI taps.
  • Monthly budgets on Excel or traditional personal finance apps fail because humans do not experience life in 30-day blocks; we experience life one day at a time.
  • Traditional budgeting apps require tedious manual receipt logging days after the purchase has already done its damage.

We built Finpulse to invert this model: converting arbitrary monthly salary into a hard, live daily spending allowance that ticks down in real time with every transaction, paired with an intelligent AI Financial Copilot and seamless PayPal sandbox integration to enforce automated micro-pacing, subscription radar, and smart wealth escrow.

⚡ What Finpulse Does

Finpulse is an offline-ready, privacy-first personal finance platform and Android app designed to turn impulsive spenders into disciplined wealth builders:

1. 🪙 Real-Time Daily Allowance Engine

  • Converts monthly salary into a realistic, tangible daily allowance (e.g., ₹1,00,000 / 30 = ₹3,333/day).
  • Discretionary Mode: Automatically pre-deducts fixed commitments (Rent, EMIs, SIPs, Cook/Maid, WiFi) so you only see what is genuinely safe to spend.
  • Hourly Burn Pace: Calculates your safe spending velocity per remaining hour of the day.

2. 🤖 Built-In AI Financial Copilot

  • Live Financial Health Audit: Scores your financial posture from 0 to 100 based on liquid reserves, fixed obligations, and impulse volatility.
  • Micro-Leak Radar: Unmasks "The 10-Minute Delivery Tax" — showing exactly how platform fees, surge pricing, and small cart surcharges on quick-commerce drain up to ₹2,400+ monthly.
  • Conversational Financial Intelligence: Ask questions like "Where is my money leaking?", "How does inflation affect my idle cash?", or "Can I afford a weekend trip?" with contextual answers grounded in real macro benchmarks (CPI inflation, repo rates, Nifty/S&P compounding).

3. 💳 PayPal & Digital Payment Integration

  • Smart Escrow & Pacing: Demonstrates automated sandbox settlement and pacing transfers via PayPal, locking away discretionary surpluses into automated savings vaults.
  • Subscription Leak Radar: Auto-detects recurring debits (Netflix, Spotify, Prime, YouTube, Cloud storage), calculates their daily drag on allowance, and provides 1-click management and cancellation links.
  • Global & Cross-Border Ready: Handles multi-currency conversions and remote freelancer inflows seamlessly.

4. 🌙 9 PM Nightly WhatsApp Financial Verdict

  • Sends an automated, punchy 9 PM financial debrief popup and push notification: > "Your daily budget was ₹3,333. You spent ₹2,450. You saved ₹883 today!"
  • Includes culturally resonant, humorous roasts and praises that make financial discipline engaging, plus 1-tap sharing via the WhatsApp Share API.

5. 📂 Bank Statement & Passbook Vault

  • Instant client-side CSV statement parser supporting major institutions (HDFC, ICICI, SBI, Axis, PayPal, UPI logs) with 100% on-device processing.
  • Multi-bank liquid passbook tracker monitoring real-time consolidated balances without sharing sensitive banking credentials.

📐 The Mathematical Formulation

Finpulse is grounded in rigorous financial pacing formulas:

1. Net Discretionary Daily Allowance ((D))

$$D = \frac{S_{\text{monthly}} - \sum_{i=1}^{n} C_{\text{fixed}, i}}{30}$$ Where (S_{\text{monthly}}) is the monthly in-hand income and (C_{\text{fixed}}) represents fixed non-negotiables (Rent, EMIs, utilities, SIP investments).

2. Real-Time Hourly Burn Velocity ((V_{\text{hourly}}))

$$V_{\text{hourly}}(t) = \frac{D_{\text{remaining}}}{24 - t_{\text{hour}}}$$ This informs the user in real time whether their current rate of consumption will survive until midnight.

3. Emergency Runway Metric ((R_{\text{runway}}))

$$R_{\text{runway}} = \frac{B_{\text{liquid}}}{E_{\text{monthly}}}$$ Where (B_{\text{liquid}}) is the total liquid cash across all verified accounts, and (E_{\text{monthly}}) is the full monthly burn rate.

4. Real Purchasing Power Inflation Drag ((\Delta_{\text{drag}}))

$$\Delta_{\text{drag}} = B_{\text{idle}} \times (i_{\text{CPI}} - r_{\text{savings}})$$ Quantifies the hidden annual tax on idle low-interest cash (e.g. (5.4\% \text{ CPI} - 2.8\% \text{ Savings Rate} = 2.6\%) real loss).

🛠️ How We Built It

  • Frontend & Fullstack Core: Next.js 16 with App Router, TypeScript, and Turbopack for sub-second compile and zero-latency UI re-renders.
  • Styling & Design System: Tailwind CSS v4 featuring custom dark-mode glassmorphism, responsive mobile-first views, and emerald/gold fintech accents.
  • Interactive Visualizations: Chart.js & React-Chartjs-2 powering the 7-day daily limit burn bar chart and category distribution donut charts.
  • AI Intelligence Layer: Self-contained client-side AI heuristics and financial reasoning engine (finpulseAI.ts) that runs 100% locally with zero external API dependencies, protecting user data privacy.
  • Mobile Native Packaging: Capacitor Core & Android CLI (@capacitor/android) compiling the app into a standalone Android APK (finpulse.apk).
  • Interactive Micro-Interactions: Web Audio API synthesizer generating clean, procedural digital tap chimes and celebration chords without heavy MP3 assets, complemented by Canvas-Confetti.
  • Data Persistence: Offline-first reactive state management with localStorage hydration and REST API routes.

🧗 Challenges We Faced

  1. Bank Statement Format Discrepancies: Every banking portal formats transaction memos differently. We developed a resilient heuristic parser that strips UTR numbers, UPI IDs, and noise to accurately isolate real merchants and categorize them into 9 distinct buckets.
  2. True Zero-Latency Financial Calculations: Recalculating remaining allowance, hourly velocity, category breakdowns, and AI audits on every keystroke required memoized pure state transitions to avoid mobile UI stutter.
  3. Data Privacy vs. AI Utility: Financial data is deeply sensitive. Instead of transmitting raw bank accounts to third-party cloud LLMs, we engineered a local, privacy-first financial copilot that performs deterministic audits on-device while maintaining natural conversational intelligence.

🏆 Key Accomplishments We're Proud Of

  • ⚡ 0ms Local Latency: Instantaneous feedback loop whenever an expense is logged.
  • 📱 Native Android Deployment: Packaged and verified into a working production Android APK.
  • 🔒 100% Privacy by Design: No bank credentials stored on external servers; everything stays with the user.
  • 🎯 Behavioral Impact: Replacing abstract monthly figures with concrete daily allowances directly curtails mindless impulse purchases.

📚 What We Learned

  • Micro-transactions compound faster than macro expenses: Surcharges of ₹35-₹65 per quick-commerce order quietly sum up to tens of thousands of rupees annually.
  • Auditory and haptic feedback changes behavior: Adding crisp digital audio cues for logging expenses and celebratory chords for staying under budget creates a dopamine feedback loop for saving.
  • AI agents must be anchored in domain benchmarks: Financial advice is only credible when connected to actual economic realities like inflation rates, repo yields, and compounding curves.

🔮 What's Next for Finpulse

  • PayPal Auto-Sweep Payouts: Automated end-of-day sweeps routing unspent daily allowances into PayPal-linked high-yield savings or investment accounts.
  • Open Banking & Account Aggregator (AA) Protocol: Live automated statement fetching across international and domestic accounts.
  • Voice-Activated Expense Logging: Conversational logging ("Log ₹250 for lunch on PayPal") with speech-to-text integration.
  • Predictive Month-End Crash Detection: Machine learning model forecasting the exact calendar day an individual is likely to run out of money based on historical burn curves.

Built With

  • ai
  • android
  • artificial-intelligence
  • banking
  • capacitor
  • chartjs
  • data-analytics
  • financial-agent
  • fintech
  • nextjs
  • offline-first
  • paypal
  • personal-finance
  • privacy-first
  • react
  • tailwindcss
  • turbopack
  • typescript
  • upi
  • web-audio-api
Share this project:

Updates

Submission history