Skip to main content

cli-usage

TRADING DECISION REPORT ═══════════════════════════════════════════════

Ticker: NVDA Date: 2024-05-10 Decision: BUY Confidence: 8.5/10

Key Factors: • Strong earnings growth trajectory • Positive AI market sentiment • Technical breakout confirmed • Manageable risk profile

Risk Warnings: • High volatility expected • Market cap concentration risk • Sector rotation possibility

Agent Consensus: 4/4 analysts positive Debate Outcome: Bullish case stronger Risk Level: Moderate

═══════════════════════════════════════════════


## CLI Commands and Shortcuts

### During Analysis
- **Ctrl+C**: Cancel current analysis
- **Enter**: Skip to next phase (where applicable)
- **Space**: Toggle detailed logging

### Result Navigation
- **↑/↓**: Scroll through detailed results
- **Tab**: Switch between result sections
- **Q**: Quit and return to main menu

## Configuration Files

The CLI can save and load configurations:

### Save Configuration
```bash
# Save current settings as default
python -m cli.main --save-config default.json

# Save with custom name
python -m cli.main --save-config my_analysis.json

Load Configuration​

# Load saved configuration
python -m cli.main --load-config default.json

# Quick analysis with preset
python -m cli.main --preset quick_analysis

Batch Processing​

For multiple analyses:

# Analyze multiple tickers
python -m cli.main --batch tickers.txt --date 2024-05-10

# Time series analysis
python -m cli.main --ticker NVDA --date-range 2024-05-01:2024-05-10

Output Options​

Export Results​

  • JSON: Machine-readable format
  • CSV: Spreadsheet-compatible
  • PDF: Human-readable report
  • TXT: Simple text summary
# Export to file
python -m cli.main --export results.json --format json

Logging Levels​

Control CLI verbosity:

  • Minimal: Only final decisions
  • Standard: Key milestones and results
  • Detailed: Full agent communications
  • Debug: Complete trace information

Performance Optimization​

For Faster Analysis​

  • Use gpt-4o-mini models
  • Set research depth to "Quick"
  • Disable detailed logging
  • Use cached data mode

For Comprehensive Analysis​

  • Use o1-preview for deep thinking
  • Set research depth to "Deep"
  • Enable online tools
  • Enable detailed logging

Troubleshooting CLI Issues​

Common Problems​

CLI Won't Start:

# Check Python version
python --version

# Verify installation
pip list | grep tradingagents

API Connection Errors:

# Verify environment variables
echo $OPENAI_API_KEY
echo $FINNHUB_API_KEY

Performance Issues:

  • Switch to smaller models
  • Reduce research depth
  • Check internet connection for online tools

Next Steps​

After mastering the CLI:

CLI Usage

TradingAgents provides a user-friendly command-line interface (CLI) that allows you to interact with the framework without writing code. The CLI provides an interactive experience for stock analysis and trading decisions.

Starting the CLI​

To launch the CLI interface:

python -m cli.main

This command will start an interactive session where you can configure your analysis parameters and view results in real-time.

CLI Interface Overview​

The CLI provides a structured workflow:

  1. Initial Configuration Screen: Select tickers, date, models, and analysis depth
  2. Progress Tracking: Monitor agent progress as analysis runs
  3. Results Display: View comprehensive trading decisions and reasoning

Configuration Options​

Stock Selection​

  • Single Ticker: Analyze one specific stock (e.g., NVDA, AAPL)
  • Multiple Tickers: Compare multiple stocks simultaneously
  • Popular Choices: Quick selection from commonly analyzed stocks

Date Selection​

  • Specific Date: Choose exact date for historical analysis
  • Date Range: Analyze trends over multiple days
  • Recent Data: Use most recent available trading day

Model Selection​

Choose LLM models based on your needs:

For Production Analysis:

  • Deep Thinking: o1-preview for complex reasoning
  • Quick Thinking: gpt-4o for fast responses

For Testing (Cost-Effective):

  • Deep Thinking: gpt-4o-mini
  • Quick Thinking: gpt-4o-mini

Research Depth​

Control the thoroughness of analysis:

  • Quick (1 round): Fast analysis with minimal debate
  • Standard (2-3 rounds): Balanced analysis with moderate debate
  • Deep (4-5 rounds): Comprehensive analysis with extensive debate

Interactive Features​

Real-Time Progress​

The CLI shows live updates as agents work:

  • Agent activation status
  • Current analysis phase
  • Data gathering progress
  • Debate round completion

Agent Activity Display​

Monitor individual agent contributions:

  • Fundamentals Analyst: Company financial analysis
  • Sentiment Analyst: Market sentiment evaluation
  • News Analyst: Recent news impact assessment
  • Technical Analyst: Chart pattern and indicator analysis
  • Researchers: Bullish vs bearish debate process
  • Risk Management: Risk assessment and portfolio impact

Decision Output​

View comprehensive results including:

  • Final Recommendation: BUY/SELL/HOLD decision
  • Confidence Score: Decision certainty level
  • Key Reasoning: Summary of analysis factors
  • Risk Assessment: Identified risks and mitigation
  • Agent Consensus: Level of agreement between agents

CLI Workflow Example​

Here's a typical CLI session flow:

$ python -m cli.main

┌─────────────────────────────────────────────┐
│          TradingAgents CLI v1.0             │
│     Multi-Agent Trading Analysis           │
└─────────────────────────────────────────────┘

Select Stock Ticker: NVDA
Analysis Date: 2024-05-10
Deep Think Model: gpt-4o-mini
Quick Think Model: gpt-4o-mini
Research Depth: Standard (3 rounds)
Online Tools: Yes

Starting Analysis...

[1/6] Initializing Agents... ✓
[2/6] Gathering Market Data... ✓
[3/6] Running Individual Analysis...
  └─ Fundamentals Analyst: Analyzing financials... ✓
  └─ Sentiment Analyst: Processing sentiment... ✓
  └─ News Analyst: Reviewing recent news... ✓
  └─ Technical Analyst: Computing indicators... ✓

[4/6] Researcher Debate (Round 1/3)...
  └─ Bullish Researcher: Presenting bull case... ✓
  └─ Bearish Researcher: Presenting bear case... ✓
  └─ Debate Resolution: Finding consensus... ✓

[5/6] Risk Assessment...
  └─ Risk Manager: Evaluating portfolio impact... ✓
  └─ Portfolio Manager: Final review... ✓

[6/6] Generating Trading Decision... ✓

═══════════════════════════════════════════════