Skip to content

Latest commit

 

History

History
98 lines (69 loc) · 2.52 KB

README.md

File metadata and controls

98 lines (69 loc) · 2.52 KB

epub2audiobook

Convert EPUB books to audiobooks using Microsoft Edge TTS. Free, no API key required.

Designed for Russian-language books but works with any language supported by Edge TTS.

Install

macOS

brew install python ffmpeg
git clone https://github.com/firedev/epub2audiobook.git
cd epub2audiobook
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Linux (Debian/Ubuntu)

sudo apt install python3 python3-venv ffmpeg
git clone https://github.com/firedev/epub2audiobook.git
cd epub2audiobook
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Usage

# Russian male voice (default)
python epub2audio.py book.epub

# Russian female voice
python epub2audio.py book.epub --voice ru-RU-SvetlanaNeural

# Custom output directory
python epub2audio.py book.epub --output ./my_audiobook

# Faster speech
python epub2audio.py book.epub --rate "+15%"

# Chapter files only, no combined file
python epub2audio.py book.epub --no-merge

Options

Flag Default Description
epub_file required Path to .epub file
--voice ru-RU-DmitryNeural Edge TTS voice ID
--output ./output Output directory
--no-merge false Keep individual chapter files only, don't concatenate into one MP3
--rate +0% Speech rate (+10%, -5%, etc.)

Chapter splitting

Chapters are split using the EPUB's NCX table of contents, which provides anchor-level granularity. A single HTML file containing multiple chapters (common in FB2-converted EPUBs) is split at each TOC anchor point. Section titles from the TOC (story names, part headings) are preserved as separate short chapters. Falls back to spine-based splitting when no TOC is present.

Output

output/
  book-name/
    01_Chapter_One.mp3
    02_Chapter_Two.mp3
    ...
    book-name_complete.mp3    # all chapters merged

Re-running is safe — existing chapters and chunks are skipped (0-byte files from failed runs are re-generated). If a network error interrupts conversion, re-run the same command and it picks up from the last successful chunk.

Voices

Russian voices:

  • ru-RU-DmitryNeural — male, clear, neutral
  • ru-RU-SvetlanaNeural — female, clear, neutral

List all available voices:

edge-tts --list-voices | grep ru-RU

Works with any Edge TTS voice. For English: en-US-GuyNeural, en-US-JennyNeural, etc.

Requirements

  • Python 3.10+
  • ffmpeg (for concatenating chapter MP3s)

License

MIT