sigma-terminal 3.3.0

Sigma - Finance Research Agent

σ SIGMA

The AI-Powered Finance Research Agent That Actually Gets It Done

Quick Start • Features • Commands • Architecture • Roadmap

╔═══════════════════════════════════════════════════════════════════════════════╗
║                                                                               ║
║   ███████╗██╗ ██████╗ ███╗   ███╗ █████╗                                     ║
║   ██╔════╝██║██╔════╝ ████╗ ████║██╔══██╗                                    ║
║   ███████╗██║██║  ███╗██╔████╔██║███████║     Finance Research Agent         ║
║   ╚════██║██║██║   ██║██║╚██╔╝██║██╔══██║     v3.3.0 | Native macOS          ║
║   ███████║██║╚██████╔╝██║ ╚═╝ ██║██║  ██║                                    ║
║   ╚══════╝╚═╝ ╚═════╝ ╚═╝     ╚═╝╚═╝  ╚═╝                                    ║
║                                                                               ║
╚═══════════════════════════════════════════════════════════════════════════════╝

---

🎯 What is Sigma?

Sigma isn't just another finance app. It's a conversational AI agent that thinks like a quant, analyzes like a hedge fund, and speaks like your smartest friend who happens to be a CFA.

You: "Should I buy NVDA?"

σ: Let me analyze that...
   ⠋ get_stock_quote executing...
   ✓ get_stock_quote complete
   ⠹ technical_analysis executing...
   ✓ technical_analysis complete
   ⠸ get_analyst_recommendations executing...
   ✓ get_analyst_recommendations complete

┌─────────────────────────────────────────────────────────────────────┐
│ σ Sigma                                                             │
├─────────────────────────────────────────────────────────────────────┤
│                                                                     │
│  📊 NVDA Analysis | Current: $875.32 (+2.4%)                       │
│                                                                     │
│  TECHNICAL SIGNALS:                                                 │
│  • RSI: 67.2 (approaching overbought)                              │
│  • MACD: Bullish crossover 3 days ago                              │
│  • Above 20/50/200 SMA ✓                                           │
│                                                                     │
│  FUNDAMENTALS:                                                      │
│  • P/E: 65.2 (premium to sector)                                   │
│  • Revenue Growth: +122% YoY                                        │
│  • AI/Data Center: 83% of revenue                                   │
│                                                                     │
│  ANALYST CONSENSUS: Strong Buy (45 analysts)                        │
│  Price Target: $950 (median) | Range: $620-$1,200                  │
│                                                                     │
│  ⚡ VERDICT: BUY on pullbacks to $820-850 support                   │
│  Risk: High volatility, valuation stretched                         │
│                                                                     │
└─────────────────────────────────────────────────────────────────────┘

---

⚡ Quick Start

One Command Install

pip install sigma-terminal

First Launch = Automatic Setup

sigma

That's it. Sigma detects it's your first time and walks you through:

1. 🤖 Choose AI Provider — Google Gemini, OpenAI, Anthropic, Groq, xAI, or Ollama
2. 🔑 Enter API Key — Or use local Ollama (completely free, no key needed!)
3. 🔍 Auto-detect Integrations — Finds Ollama, LEAN, and more
4. 🚀 Launch Directly — Straight into the beautiful terminal UI

Your config persists at ~/.sigma/ — setup never asks again.

---

🚀 Features

🤖 Multi-Provider AI Engine

Switch between providers on the fly. Use free tiers or bring your own keys.

Provider	Models	Speed	Cost	Tool Calls
Google Gemini	gemini-2.0-flash, 1.5-pro	⚡⚡⚡	Free tier	✓ Native
OpenAI	gpt-4o, o1-preview	⚡⚡	Paid	✓ Native
Anthropic	claude-sonnet-4, opus	⚡⚡	Paid	✓ Native
Groq	llama-3.3-70b	⚡⚡⚡⚡	Free tier	✓ Native
xAI	grok-2	⚡⚡	Paid	✓ Native
Ollama	llama3.2, mistral, phi3	⚡	100% FREE	✓ Native

🛡️ Built-in Rate Limiting — No more API flooding or timeouts. Sigma intelligently manages request rates per provider.

📊 Real-Time Market Intelligence

Every tool is a function call. The AI decides what to use.

Tool	What It Does
get_stock_quote	Live price, change, volume, market cap
technical_analysis	RSI, MACD, Bollinger, MAs, Support/Resistance
get_financial_statements	Income, balance sheet, cash flow
get_analyst_recommendations	Price targets, ratings, consensus
get_insider_trades	Who's buying, who's selling
get_institutional_holders	Track the smart money
compare_stocks	Multi-stock comparison with metrics
get_market_overview	Major indices at a glance
get_sector_performance	Sector rotation analysis

📈 Backtesting Engine

Test strategies before risking capital:

Strategy	Description	Use Case
sma_crossover	20/50 MA crossover	Trend following
rsi	RSI oversold/overbought	Mean reversion
macd	MACD signal crossovers	Momentum
bollinger	Band breakout/bounce	Volatility
momentum	Price momentum	Trend continuation
breakout	S/R level breaks	Breakout trading

# Via CLI
sigma backtest AAPL --strategy sma_crossover --period 2y

# Via chat
You: "Backtest RSI strategy on SPY for the last year"

🔧 LEAN Integration (Auto-Detected!)

Sigma automatically finds your LEAN/QuantConnect installation:

✓ LEAN/QuantConnect detected!
  CLI: /usr/local/bin/lean
  Directory: /Users/you/Lean

Enable LEAN integration? [y/n] (y): y
σ LEAN integration enabled

Don't have LEAN? Sigma offers to install it for you:

! LEAN/QuantConnect not detected
Would you like to [install/manual/skip] (skip): install

σ Installing LEAN CLI via pip...
✓ LEAN CLI installed successfully!

🎨 Beautiful Terminal UI

• Animated Tool Calls — Watch ⠋ ⠙ ⠹ ⠸ ⠼ ⠴ ⠦ ⠧ ⠇ ⠏ spin as tools execute
• Live Ticker Detection — Type AAPL and see $AAPL highlighted instantly
• Smart Autocomplete — Tab to complete commands and tickers
• Pulsing σ Indicator — Subtle animation shows Sigma is thinking
• Rich Responses — Markdown tables, formatting, structure

---

💻 Commands

In-App Commands

Command	Description
/help	Comprehensive help with examples
/clear	Clear chat history
/keys	Configure API keys
/models	Show available models
/status	Current configuration
/provider <name>	Switch AI provider
/model <name>	Switch model
/backtest	Show backtesting strategies
/tickers	Popular tickers by category

Keyboard Shortcuts

Shortcut	Action
Tab	Autocomplete suggestion
Ctrl+L	Clear chat
Ctrl+M	Show models
Ctrl+H	Toggle quick help
Ctrl+P	Command palette
Esc	Cancel operation

CLI Commands

# Launch interactive mode
sigma

# Quick queries (no UI)
sigma ask "analyze AAPL"
sigma quote AAPL MSFT NVDA
sigma compare AAPL MSFT GOOGL

# Backtesting
sigma backtest AAPL --strategy sma_crossover --period 2y

# Configuration
sigma --setup           # Re-run setup wizard
sigma --status          # Show current config
sigma --setkey google YOUR_API_KEY
sigma --provider ollama
sigma --model llama3.2
sigma --list-models     # Show all available models

---

🧠 Natural Language Examples

Sigma understands you. Just talk to it:

Analysis

"What's happening with AAPL today?"
"Give me a full breakdown of NVDA"
"Is TSLA overvalued right now?"

Comparison

"Compare NVDA, AMD, and INTC"
"Which is better: GOOGL or META?"
"Show me the Magnificent 7 stocks side by side"

Technical

"Technical analysis of QQQ"
"What are the support and resistance levels for SPY?"
"Is AAPL overbought?"

Fundamentals

"What's the P/E ratio of MSFT?"
"Show me AMZN's revenue growth"
"Income statement for GOOGL"

Activity

"Who are the biggest institutional holders of AAPL?"
"Any insider trading at TSLA?"
"What are analysts saying about META?"

Market

"How's the market doing?"
"Which sectors are hot right now?"
"Give me a market overview"

Backtesting

"Backtest SMA crossover on SPY"
"Run RSI strategy on AAPL for 2 years"
"Test momentum strategy on QQQ"

---

🏗️ Architecture

sigma/
├── app.py          # Textual TUI application
│                   # - ToolCallDisplay (animated spinners)
│                   # - SigmaIndicator (pulsing σ)
│                   # - TickerHighlight (real-time detection)
│                   # - ChatLog (rich formatting)
│
├── cli.py          # CLI entry point
│                   # - First-run detection
│                   # - Auto-launch setup → interactive
│                   # - Subcommands (ask, quote, backtest)
│
├── config.py       # Settings & persistence
│                   # - detect_lean_installation()
│                   # - detect_ollama()
│                   # - install_lean_cli_sync()
│                   # - ~/.sigma/config.env
│
├── setup.py        # First-run setup wizard
│                   # - Provider selection
│                   # - API key configuration
│                   # - Integration auto-detection
│                   # - LEAN download option
│
├── llm.py          # Multi-provider AI clients
│                   # - RateLimiter (per-provider)
│                   # - GoogleLLM, OpenAILLM, AnthropicLLM
│                   # - GroqLLM, OllamaLLM (with tool calls!)
│
├── tools.py        # Financial data tools
│                   # - 11 tool functions
│                   # - OpenAI-compatible schema
│                   # - yfinance data source
│
├── backtest.py     # Backtesting engine
│                   # - 6 built-in strategies
│                   # - Performance metrics
│                   # - Risk analysis
│
└── analytics/      # Advanced analytics (v4.0)

Key Design Decisions

Decision	Why
Provider Abstraction	All LLMs implement BaseLLM.generate() — easy to add new providers
Tool System	OpenAI-compatible function calling works across all providers
Rate Limiting	Per-provider RateLimiter prevents API flooding automatically
Config Persistence	Simple env file at ~/.sigma/config.env — human readable
First-Run Detection	Marker file .first_run_complete ensures smooth onboarding
Textual TUI	Modern Python TUI framework with async support

---

🔧 Configuration

Config Location

~/.sigma/
├── config.env           # API keys and settings
└── .first_run_complete  # First-run marker

Config File Format

# ~/.sigma/config.env

# API Keys
GOOGLE_API_KEY=AIzaSy...
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GROQ_API_KEY=gsk_...
XAI_API_KEY=xai-...

# Defaults
DEFAULT_PROVIDER=google
DEFAULT_MODEL=gemini-2.0-flash

# Integrations
OLLAMA_HOST=http://localhost:11434
LEAN_ENABLED=true
LEAN_CLI_PATH=/usr/local/bin/lean
LEAN_DIRECTORY=/Users/you/Lean

# Data
OUTPUT_DIR=/Users/you/Documents/Sigma
CACHE_ENABLED=true

---

🗺️ Roadmap

✅ v3.3.0 (Current)

• Auto-setup on first launch
• LEAN auto-detection with install option
• API rate limiting (no more timeouts!)
• Ollama native tool call support
• Enhanced ticker recognition
• Animated tool call display
• Pulsing σ indicator
• Smart autocomplete
• Comprehensive help system

🔜 v3.4.0 (Next)

• Portfolio tracking & management
• Watchlists with price alerts
• Options flow analysis
• Earnings calendar integration
• Multi-ticker comparison charts
• Export to CSV/PDF

🔮 v4.0.0 (Future)

• Full LEAN backtesting integration
• Paper trading mode
• Custom strategy builder (visual)
• Discord/Slack notifications
• REST API for external integrations
• Multi-user support

---

🛠️ Development

Prerequisites

• Python 3.11+
• macOS (for native app bundle)

Setup

# Clone
git clone https://github.com/yourusername/sigma.git
cd sigma

# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install in development mode
pip install -e ".[dev]"

# Run
python -m sigma

Build macOS App Bundle

./scripts/build.sh

# Outputs:
# - dist/sigma_terminal-3.3.0-py3-none-any.whl
# - dist/Sigma.app (drag to /Applications)

Project Structure

sigma/
├── pyproject.toml    # Package configuration
├── README.md         # You are here
├── LICENSE           # Proprietary license
├── scripts/
│   ├── build.sh      # Build script
│   └── create_app.py # macOS app bundle creator
└── sigma/
    ├── __init__.py   # Package exports
    ├── __main__.py   # Entry point
    └── ...           # Core modules

---

🙏 Acknowledgments

Built with:

• Textual — Beautiful TUIs in Python
• Rich — Rich text formatting
• yfinance — Yahoo Finance data
• Plotly — Interactive charts
• Pydantic — Data validation

AI Providers:

• Google Gemini
• OpenAI
• Anthropic
• Groq
• xAI
• Ollama

---

📜 License

Proprietary. All rights reserved.

---

Built with 🧠 by humans who got tired of slow, clunky finance tools.

σ — Because your portfolio deserves an AI that actually works.

_{Star ⭐ this repo if Sigma saved you time!}