Summary

This project transforms YouTube video transcripts into EPUB ebooks, targeting users who prefer offline, readable content. It automates fetching videos from specified channels, extracts transcripts, and uses Claude AI to generate polished, magazine-style articles delivered via email or accessible through a web dashboard. The core benefit is converting passive video consumption into a portable, text-based reading experience.

How It Works

The system fetches latest videos from user-defined YouTube channels, filtering out Shorts. It extracts transcripts and employs Claude AI to convert them into structured, magazine-style articles. The output is an EPUB ebook, with optional email delivery. This AI-driven approach enhances the readability and organization of transcribed content beyond raw text.

Quick Start & Requirements

• Install: Clone repo, pip install -r requirements.txt.
• API Keys: Configure YouTube Data API and Anthropic API keys in .env (from .env.example).
• Channels: List YouTube channel handles in channels.txt.
• Run: Execute python main.py.
• Prerequisites: YouTube Data API v3, Anthropic API keys.
• Optional: Streamlit dashboard (pip install streamlit, python -m streamlit run dashboard.py).
• Automation: launchctl scripts provided for macOS.
• Links: GitHub Repo

Highlighted Details

• Automatic YouTube Shorts filtering.
• AI-powered transformation of transcripts into magazine-style articles.
• Generates EPUB ebooks for cross-device readability.
• Optional email delivery with attachments.
• Includes a web dashboard for management.

Maintenance & Community

The README does not detail specific contributors, sponsorships, or community channels (e.g., Discord, Slack).

Licensing & Compatibility

• License: MIT.
• Compatibility: Permissive for free use and modification; suitable for commercial use and closed-source linking.

Limitations & Caveats

The project addresses YouTube API quirks, including Shorts filtering, search chronology, and transcript API changes. Local execution is recommended over cloud platforms like GitHub Actions due to potential blocking. Transcript accuracy may vary; video descriptions can be included in Claude context to improve results. Article truncation mid-sentence can occur and may require adjusting max_tokens. macOS automation requires verifying Python paths for multi-installation environments.