MoneyMan is a plug-and-play AI wealth advisory module designed for IDBI Bank's digital ecosystem. It builds a Financial Behaviour Twin from 6 months of transaction history — no questionnaires, no friction. It scores risk, detects anomalies, and delivers proactive, explainable advice through a voice-enabled conversational avatar.
Transaction History (6 months)
│
▼
┌──────────────────────────────┐
│ Behavioural Profiling │ ← Computes risk score, ratios,
│ Engine (server/engines/) │ spend anomalies silently
└──────────────┬───────────────┘
│
▼
┌──────────────────────────────┐
│ Financial Behaviour Twin │ ← Living data model of the customer
│ (Risk · Ratios · Trends) │
└──────────────┬───────────────┘
│
┌───────┴────────┐
▼ ▼
┌─────────────┐ ┌─────────────────┐
│ Explainable │ │ Proactive Nudge │ ← Every tip has a data trigger
│ Recommenda- │ │ Engine │ Every alert has a root cause
│ tions │ └─────────────────┘
└──────┬──────┘
│
▼
┌──────────────────────────────┐
│ Groq LLM (llama-3.1-8b) │ ← Profile context injected as
│ + WebSpeech API │ system prompt; voice I/O free
└──────────────────────────────┘
Financial Behaviour Twin Silently analyzes 6 months of transaction data to auto-compute risk profile, asset ratios (Investment : Savings : Spending), and spending anomaly alerts. Zero questionnaires.
Explainable Advisory Engine Every recommendation surfaces its trigger — "SIP reduced 53%, rebalancing advised." Fully audit-ready for bank compliance.
Voice-First Conversational Avatar Groq Llama-3.1-8B for near-zero latency responses. Browser-native WebSpeech for TTS/STT — zero marginal cost per conversation.
What-If Goal Simulator Drag a slider, watch projected growth and months-to-goal update instantly. Warns when the user falls behind their target.
Neo-Bento Design System
Soft-brutalist bento grid, 40px border radius, font-weight: 900 for financial legibility, full light/dark mode token system.
| Layer | Technology |
|---|---|
| Frontend | React 18 · Vite 5 · Framer Motion |
| Charts | Recharts |
| Voice | Web Speech API (native, zero cost) |
| Backend | Node.js · Express 5 |
| AI | Groq Cloud · llama-3.1-8b-instant |
| Styles | Vanilla CSS custom properties |
Prerequisites: Node.js v18+
# 1. Clone
git clone https://github.com/Durgaprasad-Developer/MoneyMan.git
cd MoneyMan
# 2. Install
npm install
# 3. Environment
echo "GROQ_API_KEY=your_key_here" > .env
# Get a free key → https://console.groq.com/
# 4. Run (starts both Express :3002 and Vite :5173)
npm run dev├── server/
│ ├── engines/
│ │ ├── profiling.js # Risk score, ratios, trend computation
│ │ ├── recommendations.js # Trigger-based explainable advice
│ │ └── nudges.js # Anomaly detection + goal simulation
│ ├── data/ # Mock transaction & customer data
│ └── index.js # Express API (profile, chat, simulate)
└── src/
├── components/ # Dashboard, Chat, Goals, Profile, Nudges
├── context/ # Theme (light/dark) global state
├── hooks/ # API wrapper + speech synthesizer hooks
├── utils/ # Constants, formatters, color tokens
├── App.jsx # Shell layout + sidebar navigation
└── index.css # Design token system (colors, radii, grids)
Monolithic service — Express serves the built React frontend in production.
| Setting | Value |
|---|---|
| Build Command | npm install && npm run build |
| Start Command | node server/index.js |
GROQ_API_KEY |
your Groq key |
NODE_ENV |
production |
Built for IDBI Innovate Hackathon 2026 · Track 01 · Team MoneyMan




