ebook2audiobook Turns Any Book Into a Free Voiced Audiobook Across 1158 Languages
ebook2audiobook turns any DRM-free ebook into a chaptered audiobook with voice cloning, 1158 languages, and zero per-book cost
Takeaways
- Free, local audiobook generation: converts DRM-free ebooks to chaptered
.m4baudiobooks with no subscription or per-character cost. - Voice cloning from a 6-second clip: uses XTTSv2 (Coqui TTS) for zero-shot voice cloning across 17 high-quality languages.
- 1158 language support: Meta's Fairseq/MMS engine covers 1,100+ additional languages, including low-resource ones like Yoruba and Swahili.
- 8 TTS engines: XTTSv2, Bark, Fairseq, VITS, Tacotron2, Tortoise, GlowTTS, YourTTS — choose based on hardware and quality needs.
- Runs on 2GB RAM / 1GB VRAM minimum: supports CPU, CUDA, ROCm, Apple Silicon, Intel XPU, and Jetson; Docker images available for all.
- 19,000+ GitHub stars: available at github.com/DrewThomasson/ebook2audiobook under Apache-2.0 license.
ebook2audiobook has quietly become one of the most starred open-source audio tools on GitHub, sitting at nearly 20,000 stars with over 1,600 forks. The premise is simple: drop in an ebook, optionally hand it a short audio clip of any voice, and get back a properly chaptered .m4b audiobook file , entirely for free, running on your own hardware. No subscription, no per-character billing, no cloud dependency.
The project is built around a hard problem that commercial tools sidestep: converting a full-length book is not the same as converting a paragraph. A typical 80,000-word book needs six to nine hours of finalized narration. Services like ElevenLabs charge per character, and the Pro plan at $99/month covers roughly one full novel if you are careful , and the subscription continues whether you create audiobooks that month or not. ebook2audiobook eliminates that math entirely.
The pipeline under the hood
The tool automates the full pipeline: it reads the ebook file, splits it into appropriate segments (chapters, paragraphs), uses text-to-speech models to synthesize audio, optionally applies voice cloning, and outputs a final audiobook. The heavy lifting on format parsing is handled by Calibre, the well-known ebook management tool, which means the input format support is broad.
Supported input formats include .epub, .pdf, .mobi, .txt, .html, .rtf, .fb2, .odt, .cbr, .cbz, and more, with best results from .epub or .mobi for automatic chapter detection.
There is also OCR support for scanned PDFs and image-based pages , useful for digitized physical books.
The output is a proper audiobook file, not just a flat audio dump.
The .m4b format supports chapter markers, cover art, and metadata that sync with iTunes, Audible, and most audiobook players.
Eight TTS engines, one tool
The most technically interesting part of ebook2audiobook is its multi-engine architecture. The tool supports eight distinct TTS engines: XTTSv2, Bark, Fairseq, VITS, Tacotron2, Tortoise, GlowTTS, and YourTTS. This is a strategic architecture that lets you balance quality, speed, and resource usage. XTTSv2 delivers stunning voice cloning with emotional nuance, while YourTTS and Tacotron2 provide lightweight options for CPU-only machines. Bark excels at expressive narration, Fairseq supports the most languages, and Tortoise prioritizes quality over speed.