ytbrf 0.0.1

A CLI tool to transcribe, summarize and translate YouTube videos

ytbrf (YouTube Brief)

A CLI tool that takes a YouTube video link and generates a summary of the video content. The tool can also translate the summary into different languages.

Features

• Automatic transcript extraction from YouTube (if available)
• Local audio transcription using Whisper when YouTube transcript is not available
• AI-powered text summarization
• Optional translation of summaries
• Progress bars and rich console output
• Support for multiple languages
• Flexible configuration system

Prerequisites

• Python 3.8 or higher
• FFmpeg (for audio processing)
• YouTube API key (for accessing YouTube transcripts)

Installation

1. Clone the repository:

git clone https://github.com/yourusername/ytbrf.git
cd ytbrf

2. Install the package:

pip install -e .

3. Set up your YouTube API key:

export YOUTUBE_API_KEY='your_api_key_here'

Configuration

ytbrf uses a YAML configuration file to customize its behavior. The configuration file can be placed in:

• Current directory (config.yaml)
• User's config directory (~/.config/ytbrf/config.yaml)

Configuration Options

# Summary settings
summary:
  # Relative length of the summary (0.0 to 1.0)
  ratio: 0.2
  # Target language for translation (ISO 639-1 code, e.g., 'en', 'es', 'fr')
  target_language: ""

# Transcription settings
transcription:
  # Path to whisper.cpp executable
  whisper_path: "whisper"
  # Whisper model to use (tiny, base, small, medium, large)
  model: "base"
  # Language to force for transcription (empty for auto-detect)
  force_language: ""

# Output settings
output:
  # Default output directory
  directory: "."
  # File naming pattern (available variables: {title}, {id}, {lang})
  filename_pattern: "{title}-{lang}.txt"

# YouTube settings
youtube:
  # YouTube API key (required for transcript download)
  api_key: ""
  # Preferred audio quality (best, worst, or specific quality like 192k)
  audio_quality: "best"
  # Audio format (mp3, m4a, etc.)
  audio_format: "mp3"

# Translation settings
translation:
  # Translation model to use
  model: "Helsinki-NLP/opus-mt-{src}-{tgt}"
  # Whether to translate the full transcript (true) or just the summary (false)
  translate_full: false

Default Values

If no configuration file is found, ytbrf will use these default values:

• Summary ratio: 20% of original length
• Whisper model: "base"
• Output directory: current directory
• Audio format: mp3
• Translation: disabled by default

Usage

Basic usage:

ytbrf "https://www.youtube.com/watch?v=VIDEO_ID"

With custom summary ratio (e.g., 30% of original length):

ytbrf "https://www.youtube.com/watch?v=VIDEO_ID" --ratio 0.3

With translation to English:

ytbrf "https://www.youtube.com/watch?v=VIDEO_ID" --translate en

With custom output directory:

ytbrf "https://www.youtube.com/watch?v=VIDEO_ID" --output-dir ./summaries

Output Files

For a video titled "Example Video", the tool will generate:

• Example Video.txt - Full transcript
• Example Video-summary-{original_language}.txt - Summary in original language
• Example Video-summary-{target_language}.txt - Translated summary (if translation requested)

How It Works

1. The tool first attempts to get the transcript directly from YouTube using the YouTube API
2. If no transcript is available, it downloads the audio and uses Whisper for local transcription
3. The transcript is then summarized using a transformer model
4. If translation is requested, the summary is translated to the target language
5. All files are saved in the specified output directory

Supported Languages

The tool supports translation to any language available in the Helsinki-NLP models. Common language codes include:

• en - English
• es - Spanish
• fr - French
• de - German
• it - Italian
• pt - Portuguese
• ru - Russian
• zh - Chinese
• ja - Japanese
• ko - Korean

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.