YouTube Playlist Summary

A Python tool to generate summaries of YouTube playlists using AI. Supports both AWS Bedrock (Claude) and Ollama models for processing.

An Example of entries generated by this project is available here: https://cmanaha.github.io/TechConferences/cloud/AWS/aws_reinvent_2024_compute.html

Features

• Automatic transcript extraction from YouTube videos
• AI-powered video categorization and summarization
• Concurrent video processing for faster results
• Cost tracking for AWS Bedrock usage
• Configurable via environment variables or command line
• Support for multiple LLM providers (AWS Bedrock Claude and Ollama)
• Markdown output with video thumbnails and summaries

Prerequisites

• Python 3.10+
• AWS credentials configured (if using Claude)
• Ollama installed (if using local models)
• YouTube Data API access

Installation

# Clone the repository
git clone https://github.com/cmanaha/youtube-playlist-summary.git
cd youtube-playlist-summary

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

Configuration

Create a .env file in the project root:

PLAYLIST_URL=https://www.youtube.com/playlist?list=YOUR_PLAYLIST_ID
MODEL=claude        # Use 'claude' for AWS Bedrock or 'llama3.2' for Ollama
BATCH_SIZE=10      # Number of concurrent video processes
VIDEO_COUNT=10     # Limit number of videos to process
VERBOSE=true       # Enable detailed logging

Usage

Basic usage:

python src/main.py

Advanced usage with command line arguments:

python src/main.py \
  --playlist-url "https://www.youtube.com/playlist?list=YOUR_PLAYLIST_ID" \
  --model claude \
  --batch-size 10 \
  --videos 5 \
  --categories "ai & ml,security" \
  --verbose

Output

The tool generates a markdown file in the output/ directory with:

• Playlist title and overview
• Table of contents by category
• Video summaries with thumbnails
• Links to original videos

Example output:

# AWS re:Invent 2024 Playlist Summary

## Table of Contents
- [AI & ML](#ai--ml) (5 videos)
- [Security](#security) (3 videos)
- [Serverless](#serverless) (2 videos)

## AI & ML
[Video thumbnail and summary...]

Cost Tracking

When using AWS Bedrock (Claude), the tool tracks and displays:

• Cost per API call
• Token usage (input/output)
• Accumulated session cost

Development

Run tests:

pytest

Format code:

black src/ tests/

License

MIT License - see LICENSE file for details.

Author

Carlos Manzanedo Rueda (@cmanaha)