12 KiB

Raw Blame History

🎉 New Features in TradingAgents

This document highlights the major enhancements added to TradingAgents!

Overview

We've added four major features to TradingAgents:

✅ Multi-LLM Provider Support - Use Claude, GPT-4, or Gemini
✅ Paper Trading Integration - Practice with real market data
✅ Web Interface - Beautiful GUI for analysis and trading
✅ Docker Support - One-command deployment

1. 🤖 Multi-LLM Provider Support

What's New

You can now use any LLM provider for TradingAgents analysis:

Anthropic Claude (Recommended for deep reasoning)
OpenAI GPT-4 (Proven performance)
Google Gemini (Cost-effective alternative)

Why It Matters

Choose the best model for your needs
Save costs with appropriate models for different tasks
Avoid vendor lock-in - switch providers anytime
Use your existing subscription - Claude, OpenAI, or Google

How to Use

1. Configure Provider

Edit .env:

# Use Claude (Recommended!)
LLM_PROVIDER=anthropic
ANTHROPIC_API_KEY=your_key_here

# Or use OpenAI
LLM_PROVIDER=openai
OPENAI_API_KEY=your_key_here

# Or use Google
LLM_PROVIDER=google
GOOGLE_API_KEY=your_key_here

2. Run Example

from tradingagents.graph.trading_graph import TradingAgentsGraph
from tradingagents.default_config import DEFAULT_CONFIG

# Configure for Claude
config = DEFAULT_CONFIG.copy()
config["llm_provider"] = "anthropic"
config["deep_think_llm"] = "claude-3-5-sonnet-20241022"
config["quick_think_llm"] = "claude-3-5-sonnet-20241022"

# Run analysis
ta = TradingAgentsGraph(config=config)
final_state, signal = ta.propagate("NVDA", "2024-05-10")

print(f"Signal: {signal}")

3. Try It

python examples/use_claude.py

Recommended Models

Provider	Deep Thinking	Quick Thinking	Budget
Anthropic	claude-3-5-sonnet-20241022	claude-3-5-sonnet-20241022	claude-3-5-haiku-20241022
OpenAI	gpt-4o	gpt-4o-mini	gpt-4o-mini
Google	gemini-1.5-pro	gemini-1.5-flash	gemini-1.5-flash

Files:

tradingagents/llm_factory.py - LLM provider factory
examples/use_claude.py - Claude usage example

2. 📈 Paper Trading Integration

What's New

Connect TradingAgents to real broker platforms for paper trading:

Alpaca - FREE paper trading with real market data
Easy order execution
Portfolio tracking
Real-time positions and P&L

Why It Matters

Practice risk-free - No real money involved
Test strategies - Validate before going live
Real market data - Actual prices and execution
Build confidence - Learn without financial risk

How to Use

1. Sign Up for Alpaca (FREE!)

Visit alpaca.markets and create account.

2. Get API Keys

Dashboard → Paper Trading → API Keys

3. Configure

Edit .env:

ALPACA_API_KEY=your_key_id
ALPACA_SECRET_KEY=your_secret
ALPACA_PAPER_TRADING=true

4. Start Trading

from tradingagents.brokers import AlpacaBroker
from decimal import Decimal

# Connect
broker = AlpacaBroker(paper_trading=True)
broker.connect()

# Check account
account = broker.get_account()
print(f"Buying Power: ${account.buying_power:,.2f}")

# Buy AAPL
order = broker.buy_market("AAPL", Decimal("10"))
print(f"Order ID: {order.order_id}")

# Check positions
positions = broker.get_positions()
for pos in positions:
    print(f"{pos.symbol}: {pos.quantity} shares, P&L: ${pos.unrealized_pnl}")

5. Try Examples

# Basic paper trading
python examples/paper_trading_alpaca.py

# TradingAgents + Alpaca integration
python examples/tradingagents_with_alpaca.py

Supported Features

✅ Market orders
✅ Limit orders
✅ Stop-loss orders
✅ Position tracking
✅ Real-time P&L
✅ Account management

Files:

tradingagents/brokers/base.py - Broker interface
tradingagents/brokers/alpaca_broker.py - Alpaca implementation
examples/paper_trading_alpaca.py - Basic example
examples/tradingagents_with_alpaca.py - Full integration

3. 🌐 Web Interface

What's New

Beautiful web-based GUI for TradingAgents:

Chat-based interface
Stock analysis
Order execution
Portfolio management
Real-time updates

Why It Matters

User-friendly - No coding required
Interactive - Chat with your trading assistant
Visual - See analysis and results
Accessible - Use from any browser

How to Use

1. Install Dependencies

Already included in requirements.txt:

chainlit

2. Configure

Edit .env (same as before - no extra setup needed!)

3. Start Web Interface

chainlit run web_app.py -w

4. Open Browser

Visit http://localhost:8000

5. Try Commands

# Analyze a stock
analyze NVDA

# Connect to paper trading
connect

# Check account
account

# View portfolio
portfolio

# Buy shares
buy NVDA 5

# Sell shares
sell NVDA 5

# Get help
help

Available Commands

Command	Description	Example
`analyze TICKER`	AI analysis	`analyze AAPL`
`connect`	Connect broker	`connect`
`account`	Account status	`account`
`portfolio`	View positions	`portfolio`
`buy TICKER QTY`	Buy shares	`buy NVDA 10`
`sell TICKER QTY`	Sell shares	`sell NVDA 5`
`provider NAME`	Change LLM	`provider anthropic`
`settings`	View config	`settings`
`help`	Show help	`help`

Screenshots

Welcome Screen:

🤖 Welcome to TradingAgents!

I'm your AI-powered trading assistant...

Analysis:

📊 Analysis Complete: NVDA

🎯 Trading Signal: BUY

Market Analysis: [Detailed analysis...]
Fundamentals: [Financial metrics...]
News Sentiment: [Recent news...]

Recommendation: BUY

Portfolio:

📈 Current Positions

NVDA
- Quantity: 10 shares
- Avg Cost: $895.50
- Current: $920.00
- P&L: $245.00 (2.73%)

Total Position Value: $9,200.00

Files:

web_app.py - Main web application
.chainlit - Chainlit configuration

4. 🐳 Docker Support

What's New

One-command deployment with Docker:

Pre-configured environment
All dependencies included
Persistent data storage
Easy scaling

Why It Matters

Quick setup - No dependency hell
Reproducible - Same environment everywhere
Isolated - Won't conflict with other projects
Production-ready - Deploy anywhere

How to Use

1. Prerequisites

Install Docker and Docker Compose

2. Configure

# Copy environment file
cp .env.example .env

# Edit with your keys
nano .env

3. Build and Run

# Build container
docker-compose build

# Start services
docker-compose up

# Access at http://localhost:8000

That's it! 🎉

Docker Commands

# Start in background
docker-compose up -d

# View logs
docker-compose logs -f

# Stop
docker-compose down

# Restart
docker-compose restart

# Run Python scripts
docker-compose run tradingagents python examples/portfolio_example.py

# Open shell
docker-compose run tradingagents bash

# Run tests
docker-compose run tradingagents pytest tests/ -v

Optional: Jupyter Notebook

# Start with Jupyter
docker-compose --profile jupyter up

# Access at http://localhost:8888

Data Persistence

Data is automatically persisted in:

./data/              # Market data cache
./eval_results/      # Analysis results
./portfolio_data/    # Portfolio state

Files:

Dockerfile - Container image definition
docker-compose.yml - Multi-service setup
.dockerignore - Build optimization
DOCKER.md - Complete Docker guide

🚀 Getting Started with New Features

Quick Start Guide

1. Clone Repository

git clone https://github.com/TauricResearch/TradingAgents.git
cd TradingAgents

2. Configure Environment

cp .env.example .env
nano .env

Add your API keys:

# LLM Provider (choose one)
ANTHROPIC_API_KEY=your_claude_key
# or
OPENAI_API_KEY=your_openai_key

# Data Provider
ALPHA_VANTAGE_API_KEY=your_av_key

# Paper Trading (optional)
ALPACA_API_KEY=your_alpaca_key
ALPACA_SECRET_KEY=your_alpaca_secret

3. Choose Your Method

Option A: Docker (Easiest)

docker-compose up
# Open http://localhost:8000

Option B: Local Installation

pip install -r requirements.txt
pip install -e .
chainlit run web_app.py -w
# Open http://localhost:8000

4. Try It Out

In the web interface:

analyze NVDA
connect
buy NVDA 5
portfolio

Or run Python scripts:

python examples/use_claude.py
python examples/paper_trading_alpaca.py
python examples/tradingagents_with_alpaca.py

📊 Complete Example Workflow

Here's a complete workflow using all new features:

1. Analyze Stock with Claude

from tradingagents.graph.trading_graph import TradingAgentsGraph
from tradingagents.default_config import DEFAULT_CONFIG

config = DEFAULT_CONFIG.copy()
config["llm_provider"] = "anthropic"
config["deep_think_llm"] = "claude-3-5-sonnet-20241022"

ta = TradingAgentsGraph(config=config)
final_state, signal = ta.propagate("NVDA", "2024-05-10")

2. Execute Trade on Alpaca

from tradingagents.brokers import AlpacaBroker
from decimal import Decimal

broker = AlpacaBroker(paper_trading=True)
broker.connect()

if signal == "BUY":
    order = broker.buy_market("NVDA", Decimal("5"))
    print(f"Bought NVDA: {order.order_id}")

3. Track Performance

positions = broker.get_positions()
for pos in positions:
    print(f"{pos.symbol}: ${pos.unrealized_pnl:,.2f} P&L")

4. Use Web Interface

chainlit run web_app.py -w

Then in browser:

analyze NVDA
buy NVDA 5
portfolio

5. Run in Docker

docker-compose up
# Access http://localhost:8000

🎓 Learning Resources

Documentation

Multi-LLM: tradingagents/llm_factory.py docstrings
Paper Trading: tradingagents/brokers/README.md
Web Interface: web_app.py comments
Docker: DOCKER.md

Examples

examples/use_claude.py - Claude integration
examples/paper_trading_alpaca.py - Basic paper trading
examples/tradingagents_with_alpaca.py - Full integration
examples/portfolio_example.py - Portfolio management

Tests

tests/portfolio/ - Portfolio tests (96% coverage)
tests/backtest/ - Backtesting tests
Run with: pytest tests/ -v

🔮 What's Next?

Future enhancements we're considering:

More Brokers: Interactive Brokers, TD Ameritrade
Advanced Charts: TradingView integration
Alerts: Email/SMS notifications
Strategies: Pre-built trading strategies
Backtesting UI: Visual backtesting in web interface
Mobile App: iOS/Android support

💬 Support

Need help?

Documentation: Check README files in each module
Examples: Run scripts in examples/ directory
Issues: Report bugs on GitHub
Questions: Use GitHub Discussions

🎉 Summary

You now have access to:

✅ Claude Integration - Best-in-class LLM for trading analysis ✅ Paper Trading - Risk-free practice with real market data ✅ Web Interface - User-friendly GUI for everything ✅ Docker Support - One-command deployment anywhere

Get started today:

# Clone and configure
git clone https://github.com/TauricResearch/TradingAgents.git
cd TradingAgents
cp .env.example .env
# Edit .env with your keys

# Start with Docker
docker-compose up

# Or run locally
chainlit run web_app.py -w

# Open http://localhost:8000 and start trading! 🚀

Happy Trading! 📈

12 KiB Raw Blame History

🎉 New Features in TradingAgents

Overview

1. 🤖 Multi-LLM Provider Support

What's New

Why It Matters

How to Use

Recommended Models

2. 📈 Paper Trading Integration

What's New

Why It Matters

How to Use

Supported Features

3. 🌐 Web Interface

What's New

Why It Matters

How to Use

Available Commands

Screenshots

4. 🐳 Docker Support

What's New

Why It Matters

How to Use

Docker Commands

Optional: Jupyter Notebook

Data Persistence

🚀 Getting Started with New Features

Quick Start Guide

📊 Complete Example Workflow

🎓 Learning Resources

Documentation

Examples

Tests

🔮 What's Next?

💬 Support

🎉 Summary

12 KiB

Raw Blame History