EPUB to Audiobook Converter

A command-line tool and library that converts EPUB ebooks to OGG audiobooks with chapter markers, using local text-to-speech processing with Kokoro.

Features

• Converts EPUB 3.0 files to OGG audiobooks
• Local text-to-speech processing using Kokoro
• Support for 9 languages with multiple voices:
  • 🇺🇸 American English (11F, 9M voices)
  • 🇬🇧 British English (4F, 4M voices)
  • 🇯🇵 Japanese (4F, 1M voices)
  • 🇨🇳 Mandarin Chinese (4F, 4M voices)
  • 🇪🇸 Spanish (1F, 2M voices)
  • 🇫🇷 French (1F voice)
  • 🇮🇳 Hindi (2F, 2M voices)
  • 🇮🇹 Italian (1F, 1M voices)
  • 🇧🇷 Brazilian Portuguese (1F, 2M voices)
• Chapter markers in output files
• Configurable voice selection and speech rate
• Progress reporting with optional quiet mode
• Metadata preservation from EPUB to audio file

Requirements

• Python 3.10 or higher
• Dependencies listed in pyproject.toml

Installation

1. Pip install

pip install epub2audio

2. Clone this repository:

git clone https://github.com/clayrosenthal/epub2audio.git
cd epub2audio

3. Install dev setup using mise (recommended):

mise install

Or manually with a virtual environment:

python -m venv .venv
source .venv/bin/activate 
pip install -e .

Usage

Basic Usage

Convert an EPUB file to an audiobook with default settings:

epub2audio input.epub

The audiobook will be saved as Book_Title.ogg, but can be set with --output.

Voice Selection

The tool supports multiple voices across different languages. Here are some notable voices:

American English

• af_bella (Female, Grade A-) - High quality with extended training
• af_heart (Female, Grade A) - Best overall quality
• af_nicole (Female, Grade B-) - Good quality with extended training
• am_fenrir (Male, Grade C+) - Best male voice option

British English

• bf_emma (Female, Grade B-) - Best British female voice
• bm_fable (Male, Grade C) - Best British male voice

Other Languages

• ff_siwis (French Female, Grade B-)
• jf_alpha (Japanese Female, Grade C+)
• if_sara (Italian Female, Grade C)
• hf_alpha (Hindi Female, Grade C)

To use a specific voice:

epub2audio input.epub --voice af_bella

Advanced Options

epub2audio input.epub \
  --output output.ogg \
  --voice af_bella \
  --speech-rate 1.0 \
  --quiet

Command Line Options

• input.epub: Path to input EPUB file
• --output, -o: Path of output audiobook file, defaults to title of the ebook.
• --voice, -v: Name of the voice to use (default: af_heart)
• --speech-rate, -r: Speech rate multiplier (default: 1.0)
• --quiet, -q: Suppress progress reporting
• --verbose, -v: Output more verbose logs
• --cache, -c: Cache generated audio files for reuse
• --max-chapters, -m: Max number of chapters to generate, or -1 for unlimited
• --format, -f: Output container format

Voice Quality Grades

Voices are graded based on quality and training data:

• A: Exceptional quality, extensive training
• B: Good quality, suitable for most uses
• C: Average quality, may have minor issues
• D: Basic quality, may have noticeable issues
• F: Limited quality, recommended only if necessary

Modifiers (+/-) indicate slight variations within each grade.

Development

Project Structure

src/
├── __init__.py
├── epub_processor.py      # EPUB parsing and text extraction
├── audio_converter.py     # TTS conversion using Kokoro
├── audio_handler.py       # OGG creation, chapter markers, metadata
├── epub2audio.py          # Main class, command line interface
├── voices.py              # Voice definitions and management
├── helpers.py             # Utility functions
└── config.py              # Configuration settings

Running Tests

Using mise:

# Run all tests
mise run test

# Run integration tests
mise run test-integration

# Run with coverage
mise run test-coverage

Or manually:

# Run all tests
pytest

# Run integration tests
pytest --run-integration

# Run with coverage
pytest --cov=src tests/

Code Quality

The project uses:

• Ruff for formatting and linting
• MyPy for type checking
• Pytest for testing

To format and lint code:

mise run format  # Format code
mise run lint    # Check code
mise run fix     # Auto-fix issues

Error Handling

Critical Errors (Exit with error code)

1. Invalid/corrupted EPUB file
2. Invalid voice model selection
3. File system errors (read/write permissions)
4. Insufficient disk space

Non-Critical Errors (Warning and continue)

1. Non-text elements in EPUB
2. Unsupported metadata fields
3. Minor formatting issues

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes
4. Run tests (mise run test)
5. Format code (mise run format)
6. Commit your changes (git commit -m 'Add amazing feature')
7. Push to the branch (git push origin feature/amazing-feature)
8. Open a Pull Request

TO-DO:

• Add support for epub by link
• Add support for taking in multiple epubs on the command kine
• Add a webserver so you could host this
• Add support for more audio output formats
• Add support for images and vector graphics
  • Either their alt text, or generate it with AI
• Better integration tests
• Add support for ONNX runtime
• Add support for other AI models

Development Guidelines

1. Follow Google Python style guide
2. Add tests for new features
3. Update documentation as needed
4. Keep commits focused and atomic

License

AGPL-3.0-or-later - See LICENSE file for details

Acknowledgments

• Kokoro for text-to-speech processing
  • Huggingface for model weights
• ebooklib for EPUB handling
• mutagen for audio metadata
• Voice training data contributors (see individual voice attributions)