Readtube

Transform YouTube videos into well-written magazine-style articles, delivered as EPUB, PDF, or HTML.

Setup

Option 1: Add as a Claude Code skill (easiest)

code

/add-skill https://raw.githubusercontent.com/unbalancedparentheses/readtube/main/SKILL.md

Then ask: "Create an ebook from https://youtube.com/watch?v=VIDEO_ID"

Option 2: Clone and work from the repo

bash

git clone https://github.com/unbalancedparentheses/readtube.git
cd readtube
pip install -r requirements.txt

Then ask Claude Code: "Create an ebook from https://youtube.com/watch?v=VIDEO_ID"

How to Use This Skill

When the user wants to create an ebook from YouTube videos, follow this workflow:

Step 1: Fetch Video and Transcript

bash

# Single video
python fetch_transcript.py "https://www.youtube.com/watch?v=VIDEO_ID"

# Multiple videos
python fetch_transcript.py "URL1" "URL2" "URL3"

# Playlist (all videos)
python fetch_transcript.py "https://www.youtube.com/playlist?list=PLxxx"

# Playlist with limit
python fetch_transcript.py "https://www.youtube.com/playlist?list=PLxxx" --max 5

# From configured channels
python fetch_transcript.py --channels

# With specific language
python fetch_transcript.py "URL" --lang es

# List available languages
python fetch_transcript.py "URL" --list-languages

# Summary mode (short 2-3 paragraph summary)
python fetch_transcript.py "URL" --summary

# Custom output directory
python fetch_transcript.py "URL" --output-dir ./ebooks

Step 2: Write the Article

Using the transcript output, write a magazine-style article following these guidelines:

Article Writing Guidelines:

•Start with an engaging headline (different from the video title)
•The audience is a curious individual who is smart but not a specialist
•Make it highly engaging and readable - think New Yorker or The Atlantic
•Explain any jargon or obscure references
•Capture key insights, contrarian viewpoints, memorable anecdotes
•Preserve key quotes (clean up filler words or transcription errors)
•Use the video title/description to correct transcription errors
•Do NOT include phrases like "In this video" - write standalone
•Format in clean markdown
•Length depends on content and insight density

Optional: Local LLM Drafting (no Claude Code)

If Claude Code isn't available, you can draft locally:

bash

python fetch_transcript.py "URL" --output-json video.json
python write_article.py video.json --backend claude --output-dir ./drafts
# or use llama.cpp
python write_article.py video.json --backend llama-cpp --model /path/to/model.gguf --output-dir ./drafts

Step 3: Create the Ebook

python

from create_epub import create_ebook

articles = [{
    "title": "Original Video Title",
    "channel": "Channel Name",
    "url": "https://youtube.com/watch?v=...",
    "thumbnail": "https://...",  # Optional: used as cover
    "article": """# Your Article Headline

Your article content in markdown...
"""
}]

# Create EPUB (default)
create_ebook(articles, format="epub")

# Create PDF (requires weasyprint)
create_ebook(articles, format="pdf")

# Create HTML
create_ebook(articles, format="html")

Features

•Playlist support: Process entire playlists with one URL
•Multiple output formats: EPUB, PDF, HTML
•Thumbnail covers: Automatically uses video thumbnail as cover
•Chapter support: Extracts video chapters for article structure hints
•Language selection: Choose preferred transcript language
•Summary mode: Request short summaries instead of full articles
•Transcript caching: Speeds up repeated requests (7-day cache)
•Professional typography: Inspired by Practical Typography
•Speaker labels: Preserves speaker identification when available
•Configurable output: Custom output directories

Requirements

bash

pip install -r requirements.txt

•Python 3.8+
•yt-dlp
•youtube-transcript-api
•ebooklib
•markdown

Optional for PDF: pip install weasyprint

No API keys required!

Testing

bash

make test        # Run all tests
make test-cov    # Run with coverage report
make test-e2e    # Run end-to-end tests

Key Files

code

readtube/
├── fetch_transcript.py  # Fetch video info + transcript
├── create_epub.py       # Create EPUB/PDF/HTML from articles
├── get_videos.py        # Video fetching (yt-dlp)
├── get_transcripts.py   # Transcript extraction
├── tests/               # Test suite
├── Makefile             # Common commands
└── SKILL.md             # This file

Configuring Channels

Edit CHANNELS in get_videos.py:

python

CHANNELS = [
    "@LatentSpacePod",
    "@ycombinator",
    "@DwarkeshPatel",
]

Troubleshooting

No transcript available

Some videos don't have captions. Try --list-languages to see options.

yt-dlp errors

Keep updated: pip install --upgrade yt-dlp

PDF generation fails

Install weasyprint system dependencies. See: https://doc.courtbouillon.org/weasyprint/stable/first_steps.html