audiobook-generator

Audiobook Generator

Convert any text file into a narrated audiobook. Splits long text into chapters, generates audio for each, and combines them into a single file.

Features

• Read text from a file or stdin
• Automatically split into chapters by paragraph breaks or custom markers
• Generate audio with a consistent narrator voice
• Combine all chapters into a single WAV file
• Individual chapter files saved for easy navigation

Requirements

Base dependencies are installed via the root requirements.txt. See the main README for setup. Add SMALLEST_API_KEY to your .env.

Usage

From a text file:

uv run generate.py story.txt

With custom options:

uv run generate.py story.txt --voice magnus --speed 0.9 --chapter-marker "---"

From inline text (for quick testing):

uv run generate.py --text "Once upon a time, in a land far away, there lived a curious inventor."

Configuration

Parameter	Description	Default
--voice	Narrator voice	sophia
--speed	Narration speed (0.5–2.0)	1.0
--model	TTS model	lightning-v3.1
--chapter-marker	Text that separates chapters	---
--max-chunk	Max characters per API call	500
--text	Inline text instead of file	—

Output

audiobook_output/
├── audiobook.wav        # Full audiobook
├── chapter_01.wav       # Individual chapters
├── chapter_02.wav
└── ...

Tips

• For best results, use clear paragraph breaks in your text
• The --chapter-marker flag lets you control exactly where chapters split
• Long paragraphs are automatically chunked (max 500 chars) for reliable synthesis
• Use --speed 0.9 for a more relaxed narration pace

API Reference

• Lightning v3.1 API

Next Steps

• Podcast Generator — Multi-voice conversation from a topic
• Voices — Pick the perfect narrator voice