This version addresses the 6 problems identified:

1. 6-Question Onboarding — Comprehensive behavioral profiling
2. Asset Selector Dropdown — Browse Stocks, ETFs, and Mutual Funds with subcategories
3. Conversational Chatbot — Penny is now portfolio-aware and chats naturally
4. Investment Analysis — Every investment shows verdict (Excellent/Good/Caution etc.) for YOU
5. Smart Rebalancing — Tells you WHY, WHERE, and HOW MUCH with specific actions
6. Always-On Recommendations — Even with no portfolio, you get starter pack recommendations

# 1. Navigate to backend folder
cd backend

# 2. Install dependencies
pip install -r requirements.txt

# 3. Create .env file (copy from .env.example)
# Edit it and add your Claude API key:
# ANTHROPIC_API_KEY=sk-ant-your-key-here
# SECRET_KEY=any-random-string-here

# 4. Start backend (will run on http://127.0.0.1:8000)
python -m uvicorn main:app --reload

# 1. Navigate to frontend folder
cd frontend

# 2. Install dependencies
npm install

# 3. Start frontend (will run on http://localhost:5173)
npm run dev

1. Open http://localhost:5173
2. Click "Create one" to register
3. Complete the 6-question onboarding — Penny will analyze your profile
4. Explore: Dashboard, Portfolio, Recommendations, Smart Actions, Ask Penny

penny-app/
├── backend/
│   ├── main.py                    # FastAPI entry point
│   ├── database.py                # SQLite + enhanced profile schema
│   ├── claude_service.py          # AI services (chat, analysis, profiling)
│   ├── market_data.py             # yfinance + asset universe
│   ├── recommendation_engine.py   # Always-on recommendations
│   ├── scoring.py                 # Portfolio analytics & health scoring
│   ├── seed_data.json             # Stocks/ETFs/Mutual Funds catalog
│   ├── requirements.txt
│   ├── .env.example
│   └── routers/
│       ├── auth.py                # JWT auth
│       ├── profile.py             # 6-Q onboarding
│       ├── portfolio.py           # Holdings management
│       ├── market.py              # Asset browsing/search
│       ├── chat.py                # Conversational chatbot
│       ├── actions.py             # Smart actions (rebalance, panic, etc.)
│       └── recommendations.py     # Always-on recs
└── frontend/
    ├── package.json
    ├── vite.config.js
    ├── tailwind.config.js
    ├── postcss.config.js
    ├── index.html
    └── src/
        ├── main.jsx
        ├── App.jsx                # Routing with onboarding gate
        ├── api.js                 # All API endpoints
        ├── index.css              # Tailwind + animations
        ├── pages/
        │   ├── Auth.jsx           # Login/register
        │   ├── Onboarding.jsx     # 6-question flow
        │   ├── Dashboard.jsx      # Hero + health + always-on recs
        │   ├── Portfolio.jsx      # Holdings + asset selector + analysis
        │   ├── Recommendations.jsx # Always-on with filters
        │   ├── SmartActions.jsx   # Rebalance (Why/Where/How), Overlap, Panic, Life, Hype
        │   └── Chat.jsx           # Conversational AI
        └── components/
            ├── layout/Header.jsx
            ├── investment/
            │   ├── AssetSelector.jsx     # Dropdown with subcategories
            │   └── InvestmentAnalysis.jsx # Good/Bad verdict
            ├── portfolio/
            │   ├── HealthScore.jsx
            │   ├── AllocationPie.jsx
            │   └── PortfolioMap.jsx
            ├── recommendations/RecommendationCard.jsx
            ├── insights/
            │   ├── ConcentrationWarning.jsx
            │   └── DiversificationGap.jsx
            └── shared/
                ├── Card.jsx
                ├── Badge.jsx
                └── Button.jsx

6 questions analyze: Goal, timeline, risk tolerance (via 20% drop scenario), monthly investment, experience, age. Claude generates: behavioral type, custom allocation, recommended core holdings, watch-outs, and action steps.

Three tabs: Stocks (by sector), ETFs (by category), Mutual Funds (by category). Subcategory pills filter further. Search anytime. Click any asset → AI analyzes if it fits YOU.

Every investment gets analyzed against your profile:

• Excellent Fit / Good Fit / Decent Fit / Caution / Not Recommended
• Fit Score (0-100)
• Pros/Cons specific to your situation
• Suggested allocation %
• Behavioral note
• Alternative suggestions

Penny knows your profile and holdings. Talks naturally. Uses your name when relevant. Suggests specific tickers. References past conversation. Maintains 12-message context.

Three sections:

• WHY — Headline + explanation + risk if not done
• WHERE & HOW MUCH — Specific BUY/SELL actions with dollar amounts
• IMPACT — Before/after, tax considerations, fee impact
• Plus timing recommendation and behavioral advice

• Empty portfolio → starter pack (VTI, BND, VEA, target date fund)
• Concentration risk → diversifiers
• Missing bonds → bond ETFs (when appropriate to risk)
• Sector concentration → cross-sector ETFs
• Always at least 3 recommendations

Create backend/.env:

ANTHROPIC_API_KEY=sk-ant-your-key-here
SECRET_KEY=any-long-random-string

"ANTHROPIC_API_KEY not set": Make sure .env is in the backend/ folder (not the project root)

bcrypt error: pip install bcrypt==4.0.1 passlib[bcrypt]

npm dev not found: Make sure you ran npm install first

CORS error: Backend must be on port 8000, frontend on 5173

• Backend: FastAPI, SQLite, Anthropic Claude (Sonnet), yfinance, JWT
• Frontend: React 18, Vite, Tailwind CSS, Recharts, Axios, React Router
• AI: Claude Sonnet 4 for profiling, chat, analysis, and rebalancing

Biult by Pratyaksh Singh