Multi-agent financial analysis powered by MLX DeepSeek with smart dual data strategy
- Fundamental Agent: 50+ financial ratios, company health analysis
- Technical Agent: Price trends, momentum indicators, volume analysis
- Sentiment Agent: Real-time news sentiment and market psychology
- Economic Agent: Macroeconomic climate assessment
- Supervisor Agent: Investment synthesis and grade recommendations
- Primary Source: Yahoo Finance (unlimited, reliable)
- ✅ Financial statements, ratios, market data
- ✅ Technical indicators, price history
- ✅ Real-time quotes, no rate limits
- Enhanced Source: Alpha Vantage (optional, cached)
- ⚡ News sentiment analysis
- ⚡ Economic indicators (CPI, GDP, unemployment)
- ⚡ Advanced technical indicators
# Download DeepSeek model (4-bit quantized for Apple Silicon)
huggingface-cli download mlx-community/DeepSeek-R1-Distill-Qwen-14B-4bit --local-dir ~/mlx-models# Get free API key from: https://www.alphavantage.co/support/#api-key
# Add to enhanced_agents.py line 1234 or set as environment variable
export ALPHA_VANTAGE_API_KEY="your_key_here"streamlit run app.py📈 AAPL - Investment Analysis
═══════════════════════════════════════
💰 FUNDAMENTAL ANALYSIS: 8.5/10
- Revenue: $394.3B | ROE: 26.4% | D/E: 1.73
- Strong profitability, moderate leverage
📈 TECHNICAL ANALYSIS: 7.2/10
- RSI: 45.2 | Trend: Bullish above 20-day SMA
- Volume confirmation, healthy momentum
📰 SENTIMENT ANALYSIS: 6.8/10
- Overall sentiment: Neutral-Positive
- 247 articles analyzed, 78% relevance
🌍 ECONOMIC CLIMATE: 6.5/10
- GDP: 2.3% growth | Inflation: 3.1%
- Mixed signals, Fed policy uncertainty
🧠 INVESTMENT GRADE: B+ (7.8/10)
RECOMMENDATION: BUY with 12-month target
- 25 requests per day
- 5 requests per minute
- Requests reset at midnight EST
- ✅ Company overview, financial statements
- ✅ Technical indicators (RSI, MACD, SMA)
- ✅ News sentiment analysis
- ✅ Economic indicators (monthly updates)
- ❌ Real-time intraday data
- ❌ Options data, crypto data
- Economic data: 4-hour cache (reduces daily API usage)
- Sentiment data: 1-hour cache (for active trading)
- Company data: Session cache (until app restart)
- Fallback analysis when rate limited
FinAgent/
├── app.py # Main Streamlit application
├── agents.py # Multi-agent system core
├── alpha_vantage_client.py # API client with caching
├── requirements.txt # Python dependencies
└── README.md # This file