YouTube Analyzer Skill

Extracts and analyzes YouTube videos (podcasts, trading analysis, market commentary) for actionable trading insights using AI-powered transcript analysis.

Purpose

Analyze financial content from YouTube videos to extract:

• Trading signals and stock picks
• Market sentiment and analysis
• Risk factors and warnings
• Actionable trading recommendations
• Key timestamps for important insights

Usage

When the user provides YouTube URL(s):

1. Extract video metadata using yt-dlp
2. Fetch transcript using youtube-transcript-api
3. Analyze content for trading insights using AI
4. Generate actionable recommendations with tickers and confidence scores
5. Save analysis to docs/youtube_analysis/ for reference

Supported Content Types

• Trading strategy videos
• Market analysis and commentary
• Stock picking discussions
• Economic news analysis
• Company earnings reviews
• Investment podcasts
• Financial education content

Output Format

The skill generates comprehensive markdown reports with:

• Executive Summary: Key takeaways in 3-5 bullet points
• Stock Picks: Tickers with bullish/bearish sentiment
• Trading Strategies: Specific approaches mentioned
• Risk Factors: Warnings and cautions
• Actionable Recommendations: What to do next
• Key Timestamps: Jump to important sections
• Full Transcript: Complete text for reference

Integration with Trading System

IMPORTANT: YouTube is a BIAS FILTER, NOT a Trade Trigger

Dec 11, 2025 Clarification: YouTube analysis is ONE OF THREE sentiment sources weighted at 25% in the unified sentiment aggregation. It NEVER directly triggers trades.

Sentiment Source Weights (from src/utils/unified_sentiment.py):

• News: 40%
• Reddit: 35%
• YouTube: 25% (this skill)

Why This Matters: By the time a YouTube video is published, processed, and transcribed, the market "alpha" is likely gone. YouTube insights are valuable for setting daily BIAS (long/short orientation), NOT for direct execution.

What YouTube CAN Inform:

• Daily market bias (bullish/bearish orientation)
• Watchlist additions for further technical screening
• Sentiment context for multi-LLM consensus

What YouTube CANNOT Do:

• Directly execute trades
• Override technical signals (MACD, RSI, Volume)
• Bypass risk management gates
• Act as a standalone trading signal

Integration Flow

Analysis results are used as follows:

• Fed into Multi-LLM Analyzer for sentiment scoring (20% weight)
• Used to inform Tier 2 Growth Strategy watchlist additions
• Tracked over time for accuracy validation
• Compared with actual market performance

RAG Integration

All YouTube insights are stored in RAG for continuous learning:

Storage Locations

• Transcripts: rag_knowledge/youtube/transcripts/
• Insights: rag_knowledge/youtube/insights/
• Cache: data/youtube_cache/processed_videos.json

store_insights_to_rag Tool

Store extracted insights in RAG storage for retrieval.

Parameters:

• video_id: YouTube video ID
• insights: List of extracted insights
• transcript: Full transcript text
• embedding_model: Model for embeddings (default: "text-embedding-3-small")

Returns:

• stored_count: Number of insights stored
• rag_path: Path to RAG storage

Requirements

This skill requires two Python packages installed in the virtual environment:

pip install yt-dlp youtube-transcript-api

Command Examples

# Analyze single video
python3 scripts/analyze_youtube.py --url "https://youtube.com/watch?v=VIDEO_ID"

# Analyze with custom output directory
python3 scripts/analyze_youtube.py --url "URL" --output docs/youtube_analysis/

# Analyze video by ID only
python3 scripts/analyze_youtube.py --video-id "VIDEO_ID"

# Include AI analysis (requires OpenRouter API)
python3 scripts/analyze_youtube.py --url "URL" --analyze

Skill Files

• SKILL.md: This file - skill documentation
• README.md: Detailed usage guide and examples
• scripts/analyze_youtube.py: Python script for video analysis
• scripts/example.sh: Example shell script demonstrating usage

Safety & Ethics

• Only analyze public YouTube videos
• Respect content creator rights
• Use insights for educational purposes
• Verify information before trading
• Never blindly follow trading advice from videos

Future Enhancements

• Automatic daily analysis of top financial channels
• Sentiment trend tracking across multiple videos
• Integration with Alpaca trading alerts
• AI-powered timestamp generation for key insights
• Multi-language transcript support