🌟 Carinae - AI-Powered Financial Intelligence Platform

Mumbai Hackathon 2025 | Fintech Track

"Detect Problems Before They Hurt You"

📋 Table of Contents

Overview
Problem Statement
Our Solution
Project Architecture
Key Features
Tech Stack
Projects in This Repository
Getting Started
AI Agents
Team
Documentation
License

🎯 Overview

Carinae is a comprehensive AI-powered financial intelligence platform developed for the Mumbai Hackathon 2025. Our solution addresses critical challenges faced by Small and Medium Enterprises (SMEs) and individual users in India through two specialized applications:

MoneyFyi-Business: AI CFO for SMEs - Fraud detection, compliance tracking, and cashflow forecasting
MoneyFyi-User: Personal finance management with agentic AI-powered investment recommendations

🚨 Problem Statement

For SMEs (MoneyFyi-Business)

Small and medium enterprises in India, especially retail shops and local counters, struggle with:

Financial Fraud: Unauthorized transactions and digital payment frauds in UPI/online payments
Compliance Penalties: Delayed or incorrect GST/TDS compliance leading to penalties
Cashflow Issues: Late payments, unexpected shortages, and poor financial planning
Manual Processes: Traditional apps lack real-time analysis and proactive guidance

For Individuals (MoneyFyi-User)

Irregular Income: Gig workers and informal sector employees struggle with savings
Generic Advice: Lack of personalized financial recommendations
Language Barriers: Financial literacy content not available in regional languages
Investment Complexity: Difficulty in choosing appropriate moderate-risk investments

💡 Our Solution

Agentic AI Challenge Solution

We developed an AI-agent system that autonomously:

✅ Analyzes uploaded transaction data, statements, and invoices using OCR + NLP
✅ Detects anomalies, mismatches, fraud patterns, and compliance risks
✅ Predicts cash-flow forecasts and investment opportunities
✅ Provides proactive alerts and actionable insights in real-time
✅ Delivers recommendations via user-friendly dashboards and WhatsApp notifications

🏗️ Project Architecture

High-Level System Design

┌─────────────────────────────────────────────────────┐
│              User Interface Layer                    │
│  Web Dashboard (Next.js) + Mobile PWA + WhatsApp    │
└──────────────────┬──────────────────────────────────┘
                   │
┌──────────────────▼──────────────────────────────────┐
│         Document Processing Layer                    │
│  OCR (Google Gemini Vision) + Parsing + Validation  │
└──────────────────┬──────────────────────────────────┘
                   │
┌──────────────────▼──────────────────────────────────┐
│            Multi-Agent AI Engine                     │
│  ┌─────────────────────────────────────────────┐   │
│  │  Data Normalizer Agent                       │   │
│  │  ↓                                           │   │
│  │  FraudGuard Agent → Compliance Mate Agent    │   │
│  │  ↓                   ↓                       │   │
│  │  Cashflow Oracle → SmartPayment Agent        │   │
│  │  ↓                                           │   │
│  │  Insight Agent (Recommendations)             │   │
│  └─────────────────────────────────────────────┘   │
└──────────────────┬──────────────────────────────────┘
                   │
┌──────────────────▼──────────────────────────────────┐
│         Data & Storage Layer                         │
│  Supabase PostgreSQL + AES-256 Encrypted Storage    │
└─────────────────────────────────────────────────────┘

Security Architecture

User Data → AES Encryption (Client) → HTTPS/TLS 1.3 → 
FastAPI Backend → Argon2 KDF → Second-Layer AES-256 → 
PostgreSQL (Encrypted at Rest) → Row Level Security (RLS)

Data Privacy Compliance: DPDP Act 2023, RBI Guidelines, ISO 27001 principles

✨ Key Features

🛡️ MoneyFyi-Business (SME Focus)

Fraud Detection & Prevention

Real-time Fraud Scoring: Analyze transactions using 10+ fraud indicators
Duplicate Payment Detection: UTR/reference number validation
Velocity Checks: Detect rapid-fire transactions (potential attacks)
Round-Amount Flagging: Identify suspicious payment patterns
New Vendor Alerts: Flag first-time vendors for verification
After-Hours Alerts: Detect unusual timing (weekends, late nights)

Compliance Automation

GST Invoice Validation: GSTIN format checks, HSN/SAC code verification
TDS Calculation: Automatic 10% TDS verification for services
Filing Reminders: Quarterly GST deadlines with 3-day advance alerts
Document Preparation: Auto-generate compliance reports
Penalty Avoidance: Mismatch detection before filing

Cashflow Forecasting

7-30 Day Predictions: AI-powered balance forecasting
Seasonal Analysis: Pattern recognition for recurring cycles
Shortage Alerts: Proactive warnings before cash crunch
Vendor Risk Scoring: Payment delay analysis (0-100 scale)
Budget Recommendations: Optimize spending based on forecast

Smart Payments

Vendor Verification: Background checks for new vendors
Split Payment Logic: Partial payments based on cashflow stress
Payment Prioritization: Rank payments by urgency and risk
Fraud Prevention: Block/review high-risk transactions

Alert System

🔴 Critical: Suspected fraud (immediate WhatsApp alert)
🟠 High: Compliance deadline <3 days
🟡 Medium: Cashflow shortage forecast
🟢 Low: Weekly summary reports

👤 MoneyFyi-User (Personal Finance)

Investment Recommendations

Personalized Portfolios: Based on income, risk profile, and goals
Market Sentiment Analysis: FinBERT + IndicBERT for news analysis
Stock & Mutual Fund Suggestions: Moderate-risk, data-driven picks
Explainable AI: Clear reasoning for every recommendation

Multi-Agent Intelligence

FinAgent (FinBERT): English financial news sentiment extraction
IndicAgent (IndicBERT): Regional language news & sentiment analysis
VoxAgent (VoxMind ASR): Voice input in Indian languages
TrendAgent (LSTM/Transformer): Price pattern forecasting
FusionAgent (Reinforcement Learning): Adaptive learning from user behavior

Financial Features

Expense Tracking: Automatic categorization and insights
Savings Goals: Track progress toward financial milestones
Investment Monitoring: Portfolio performance tracking
Financial Literacy: Educational content in English + Hindi
Voice Interaction: Speak queries in regional languages

🛠️ Tech Stack

Frontend (Both Projects)

{
  "framework": "Next.js 16 (App Router)",
  "language": "TypeScript 5.0",
  "styling": "Tailwind CSS v4 + shadcn/ui",
  "charts": "Recharts",
  "state": "React Hooks + Context API",
  "forms": "React Hook Form + Zod",
  "auth": "Supabase Auth (JWT)",
  "pwa": "Service Worker + Manifest"
}

Backend (Both Projects)

{
  "framework": "FastAPI (Python 3.11+)",
  "database": "Supabase (PostgreSQL + RLS)",
  "ai_engine": "Google Gemini 1.5 Flash",
  "nlp_models": ["FinBERT", "IndicBERT", "RoBERTa"],
  "ocr": "Google Gemini Vision API",
  "ml_forecasting": ["Prophet", "LSTM", "ARIMA"],
  "storage": "Supabase Storage (AES-256)",
  "notifications": "n8n Webhooks (WhatsApp via Twilio - planned)"
}

AI/ML Stack

NLP: FinBERT, IndicBERT, RoBERTa, VoxMind ASR
Vision: Google Gemini Vision, PaddleOCR
Forecasting: Meta Prophet, LSTM, Transformer, ARIMA
Anomaly Detection: Isolation Forest, DBSCAN, Z-score
Orchestration: LangChain

Infrastructure

Hosting: Vercel (Frontend), Railway/Render (Backend)
Database: Supabase PostgreSQL (with RLS)
Cache: Redis (Upstash)
Task Queue: Celery (planned)
Monitoring: Sentry (planned)
Analytics: Vercel Analytics, PostHog (planned)

📦 Projects in This Repository

1. MoneyFyi-Business (SME Solution)

Purpose: AI CFO for small businesses - fraud detection, compliance, cashflow management

Structure:

MoneyFyi-Business/
├── Backend/
│   ├── ai_engine/              # AI agents (fraud, cashflow, compliance)
│   │   ├── fraudguard_agent.py
│   │   ├── cashflow_oracle.py
│   │   ├── compliance_mate_agent.py
│   │   ├── smartpayment_agent.py
│   │   └── insight_agent.py
│   ├── app/
│   │   ├── routers/            # API endpoints
│   │   ├── services/           # Business logic
│   │   ├── models/             # Data models
│   │   └── main.py             # FastAPI app
│   └── requirements.txt
├── Frontend/
│   ├── app/                    # Next.js pages
│   ├── components/             # React components
│   └── package.json
├── tests/                      # Automated tests
├── README.md
└── sample_analysis_output.json

Key APIs:

POST /api/v1/documents - Upload & process documents
GET /api/v1/transactions - Query transactions
GET /api/v1/alerts - Retrieve alerts
GET /api/v1/insights/executive-summary - Get AI insights

Documentation: MoneyFyi-Business README

2. MoneyFyi-User (Personal Finance)

Purpose: Personal finance management with AI-powered investment recommendations

Structure:

MoneyFyi-User/
├── Backend/
│   ├── app/
│   │   ├── api/v1/             # API routes
│   │   ├── models/             # User, transaction models
│   │   └── main.py
│   └── requirements.txt
├── Frontend/
│   ├── app/                    # Next.js App Router
│   ├── components/
│   └── package.json
└── ML/
    ├── app/
    │   ├── agents/             # FinAgent, IndicAgent, etc.
    │   └── main.py
    ├── notebooks/              # Jupyter notebooks for training
    └── requirements.txt

Key Features:

Agentic AI with 6 specialized agents
Multilingual support (English + Hindi)
Voice interaction (VoxMind ASR)
Sentiment-driven recommendations
Continuous learning from user behavior

Documentation: See individual README files in MoneyFyi-User/

🚀 Getting Started

Prerequisites

Node.js 18+ and npm/pnpm/yarn
Python 3.11+
Supabase account (free tier works)
Google AI Studio account (for Gemini API key)

Quick Start (MoneyFyi-Business)

1. Clone the Repository

git clone https://github.com/yourusername/carinae.git
cd carinae/MoneyFyi-Business

2. Setup Backend

cd Backend
python -m venv venv

# Windows
venv\Scripts\activate

# macOS/Linux
source venv/bin/activate

pip install -r requirements.txt

# Create .env file
echo "SUPABASE_URL=your_supabase_url" > .env
echo "SUPABASE_SERVICE_ROLE_KEY=your_service_key" >> .env
echo "GEMINI_API_KEY=your_gemini_key" >> .env
echo "N8N_WEBHOOK_URL=your_webhook_url" >> .env

# Run backend
uvicorn app.main:app --reload

Backend runs at: http://localhost:8000
API Docs: http://localhost:8000/docs

3. Setup Frontend

cd ../Frontend
npm install

# Create .env.local
echo "NEXT_PUBLIC_SUPABASE_URL=your_supabase_url" > .env.local
echo "NEXT_PUBLIC_SUPABASE_ANON_KEY=your_anon_key" >> .env.local

# Run frontend
npm run dev

Frontend runs at: http://localhost:3000

4. Setup Database (Supabase)

Go to Supabase Dashboard
Create a new project
Go to SQL Editor
Run the migration scripts in Frontend/scripts/ in order:
- 001_create_profiles.sql
- 002_create_transactions.sql
- 003_create_vendors.sql
- 004_create_alerts.sql
- 005_create_cashflow_forecasts.sql
- 006_create_documents.sql

Quick Start (MoneyFyi-User)

Follow similar steps for MoneyFyi-User project. See specific README files:

🤖 AI Agents

MoneyFyi-Business Agents

Agent	Technology	Purpose	Key Features
Data Normalizer	Pandas, NumPy	Clean & structure data	Transaction categorization, amount normalization
FraudGuard	Isolation Forest, Rule Engine	Fraud detection	10+ fraud indicators, risk scoring (0-100)
ComplianceMate	Rule Engine, GST API	Tax compliance	GST validation, TDS checks, filing reminders
Cashflow Oracle	Prophet, ARIMA	Cashflow forecasting	7-30 day predictions, seasonal analysis
SmartPayment	Decision Tree, RL	Payment recommendations	Vendor risk scoring, split payment logic
Insight Agent	Google Gemini LLM	Natural language insights	Executive summaries, action items

MoneyFyi-User Agents

Agent	Model	Purpose
FinAgent	FinBERT	English financial news sentiment
IndicAgent	IndicBERT	Regional language sentiment
InsightAgent	RoBERTa	User intent understanding
VoxAgent	VoxMind ASR	Voice-to-text (Indian languages)
TrendAgent	LSTM/Transformer	Price forecasting
FusionAgent	Reinforcement Learning	Adaptive recommendations

👥 Team

Team Carinae - Mumbai Hackathon 2025

Rajath U - Full Stack Development, AI Integration
Niharika Trivedi - Backend Development, Database Design
Aditya S Hegde - ML Engineering, AI Agent Development
Jayesh RL - Frontend Development, UI/UX Design

📚 Documentation

MoneyFyi-Business

MoneyFyi-User

Reference Documents

🔒 Security & Compliance

Data Privacy (DPDP Act 2023)

✅ Explicit user consent for data processing
✅ AES-256 encryption for sensitive data
✅ Right to erasure (data deletion API)
✅ Breach notification within 72 hours

RBI Guidelines

✅ TLS 1.3 encrypted transmission
✅ Multi-factor authentication (planned)
✅ Audit logs for all data access
✅ No storage of bank credentials
✅ 15-minute session timeout

Technical Security

Encryption: AES-256 (at rest), TLS 1.3 (in transit)
Authentication: JWT tokens, Argon2 password hashing
Database: Row Level Security (RLS) on all tables
API: Rate limiting, CORS policies, input validation
Monitoring: Error tracking, access logs

💰 Business Model (Post-Beta)

MoneyFyi-Business Pricing

Tier	Price	Features
Free	₹0	10 docs/month, basic alerts
Starter	₹499/month	100 docs/month, WhatsApp alerts
Pro	₹1,499/month	Unlimited docs, API access, multi-user
Enterprise	Custom	White-label, on-premise, dedicated support

MoneyFyi-User Pricing

Tier	Price	Features
Free	₹0	Basic tracking, 5 recommendations/month
Premium	₹199/month	Unlimited recommendations, advanced analytics
Family	₹299/month	5 users, shared goals, priority support

🏆 Competitive Advantages

India-Specific: Built for GST/TDS compliance (not generic)
WhatsApp-First: 500M+ WhatsApp users in India
No Bank Integration: Works with uploaded documents (privacy-first)
Multi-Agent AI: More accurate than single-model systems
Agentic Approach: Autonomous decision-making, not just recommendations
Privacy-Focused: AES-256 encryption, DPDP Act 2023 compliant
Affordable: ₹499/month vs ₹10,000+ for enterprise solutions
Multilingual: English + Hindi + regional languages (voice support)

🧪 Testing

Run Backend Tests

cd MoneyFyi-Business/Backend
pytest tests/

Run Frontend Tests (Coming Soon)

cd MoneyFyi-Business/Frontend
npm test

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

Fork the repository
Create your feature branch (git checkout -b feature/AmazingFeature)
Commit your changes (git commit -m 'Add some AmazingFeature')
Push to the branch (git push origin feature/AmazingFeature)
Open a Pull Request

📞 Contact & Support

Project Website: Coming soon
Email: support@moneyfyi.com
LinkedIn: Team Carinae
GitHub Issues: Report a bug

🙏 Acknowledgments

Technologies

Next.js - React Framework
FastAPI - Modern Python API
Supabase - Backend Platform
Google AI - Gemini API
shadcn/ui - UI Components
Vercel - Hosting Platform
Tailwind CSS - CSS Framework

Research & Standards

RBI Authentication Mechanisms for Digital Payment Transactions (2025)
RBI Master Direction on Regulation of Payment Aggregators (2025)
India's DPDP Act (Digital Personal Data Protection Act, 2023)
Information Technology Act, 2000 (with amendments)
The Payment and Settlement Systems Act, 2007
The Prevention of Money Laundering Act (PMLA), 2002

Academic Papers

"UPI Based Financial Fraud Detection Using Deep Learning"
"FAMOS: Robust Privacy-Preserving Authentication"
"Secure Use of the Agent Payments Protocol (AP2)"

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🌟 Star History

If you find this project useful, please consider giving it a ⭐ on GitHub!

🚀 Built with ❤️ by Team Carinae for Mumbai Hackathon 2025

MoneyFyi - Detect Problems Before They Hurt You

Made in India 🇮🇳 | Fintech Track | Agentic AI Challenge

Website • Demo • Documentation • Support

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
MoneyFyi-Business		MoneyFyi-Business
MoneyFyi-User		MoneyFyi-User
docs		docs
README.md		README.md

Folders and files

Latest commit

History

Repository files navigation