Convert eBooks to audiobooks with chapters and metadata using Calibre and Coqui XTTS. Supports optional voice cloning and multiple languages!
Click to see images of Web GUI
- 📖 Converts eBooks to text format with Calibre.
- 📚 Splits eBook into chapters for organized audio.
- 🎙️ High-quality text-to-speech with Coqui XTTS.
- 🗣️ Optional voice cloning with your own voice file.
- 🌍 Supports multiple languages (English by default).
- 🖥️ Designed to run on 4GB RAM.
- Huggingface space is running on free cpu tier so expect very slow or timeout lol, just don't give it giant files is all
- Best to duplicate space or run locally.
- Python 3.10
coqui-ttsPython package- Calibre (for eBook conversion)
- FFmpeg (for audiobook creation)
- Optional: Custom voice file for voice cloning
-
Install Python 3.x from Python.org.
-
Install Calibre:
- Ubuntu:
sudo apt-get install -y calibre - macOS:
brew install calibre - Windows (Admin Powershell):
choco install calibre
- Ubuntu:
-
Install FFmpeg:
- Ubuntu:
sudo apt-get install -y ffmpeg - macOS:
brew install ffmpeg - Windows (Admin Powershell):
choco install ffmpeg
- Ubuntu:
-
Optional: Install Mecab (for non-Latin languages):
- Ubuntu:
sudo apt-get install -y mecab libmecab-dev mecab-ipadic-utf8 - macOS:
brew install mecab,brew install mecab-ipadic - Windows: mecab-website-to-install-manually (Note: Japanese support is limited)
- Ubuntu:
-
Install Python packages:
pip install coqui-tts==0.24.2 pydub nltk beautifulsoup4 ebooklib tqdm gradio==4.44.0 python -m nltk.downloader punkt python -m nltk.downloader punkt_tab
For non-Latin languages:
pip install mecab mecab-python3 unidic python -m unidic download
- English (en)
- Spanish (es)
- French (fr)
- German (de)
- Italian (it)
- Portuguese (pt)
- Polish (pl)
- Turkish (tr)
- Russian (ru)
- Dutch (nl)
- Czech (cs)
- Arabic (ar)
- Chinese (zh-cn)
- Japanese (ja)
- Hungarian (hu)
- Korean (ko)
Specify the language code when running the script in headless mode.
-
Run the Script:
python app.py
-
Open the Web App: Click the URL provided in the terminal to access the web app and convert eBooks.
-
For Public Link: Add
--share Trueto the end of it like this:python app.py --share True
- [For More Parameters]: use the
-hparameter like thispython app.py -h
python app.py --headless True --ebook <path_to_ebook_file> --voice [path_to_voice_file] --language [language_code]- <path_to_ebook_file>: Path to your eBook file.
- [path_to_voice_file]: Optional for voice cloning.
- [language_code]: Optional to specify language.
- [For More Parameters]: use the
-hparameter like thispython app.py -h
python app.py --headless True --use_custom_model True --ebook <ebook_file_path> --voice <target_voice_file_path> --language <language> --custom_model <custom_model_path> --custom_config <custom_config_path> --custom_vocab <custom_vocab_path>- <ebook_file_path>: Path to your eBook file.
- <target_voice_file_path>: Optional for voice cloning.
- : Optional to specify language.
- <custom_model_path>: Path to
model.pth. - <custom_config_path>: Path to
config.json. - <custom_vocab_path>: Path to
vocab.json. - [For More Parameters]: use the
-hparameter like thispython app.py -h
python app.py --headless True --use_custom_model True --ebook <ebook_file_path> --voice <target_voice_file_path> --language <language> --custom_model_url <custom_model_URL_ZIP_path>- <ebook_file_path>: Path to your eBook file.
- <target_voice_file_path>: Optional for voice cloning.
- : Optional to specify language.
- <custom_model_URL_ZIP_path>: URL Path to zip of Model folder. For Example this for the xtts_David_Attenborough_fine_tune
https://huggingface.co/drewThomasson/xtts_David_Attenborough_fine_tune/resolve/main/Finished_model_files.zip?download=true - For a custom model a ref audio clip of the voice will also be needed: ref audio clip of David Attenborough
- [For More Parameters]: use the
-hparameter like thispython app.py -h
python app.py -h- This will output the following:
usage: app.py [-h] [--share SHARE] [--headless HEADLESS] [--ebook EBOOK] [--voice VOICE]
[--language LANGUAGE] [--use_custom_model USE_CUSTOM_MODEL]
[--custom_model CUSTOM_MODEL] [--custom_config CUSTOM_CONFIG]
[--custom_vocab CUSTOM_VOCAB] [--custom_model_url CUSTOM_MODEL_URL]
[--temperature TEMPERATURE] [--length_penalty LENGTH_PENALTY]
[--repetition_penalty REPETITION_PENALTY] [--top_k TOP_K] [--top_p TOP_P]
[--speed SPEED] [--enable_text_splitting ENABLE_TEXT_SPLITTING]
Convert eBooks to Audiobooks using a Text-to-Speech model. You can either launch the
Gradio interface or run the script in headless mode for direct conversion.
options:
-h, --help show this help message and exit
--share SHARE Set to True to enable a public shareable Gradio link. Defaults
to False.
--headless HEADLESS Set to True to run in headless mode without the Gradio
interface. Defaults to False.
--ebook EBOOK Path to the ebook file for conversion. Required in headless
mode.
--voice VOICE Path to the target voice file for TTS. Optional, uses a default
voice if not provided.
--language LANGUAGE Language for the audiobook conversion. Options: en, es, fr, de,
it, pt, pl, tr, ru, nl, cs, ar, zh-cn, ja, hu, ko. Defaults to
English (en).
--use_custom_model USE_CUSTOM_MODEL
Set to True to use a custom TTS model. Defaults to False. Must
be True to use custom models, otherwise you'll get an error.
--custom_model CUSTOM_MODEL
Path to the custom model file (.pth). Required if using a custom
model.
--custom_config CUSTOM_CONFIG
Path to the custom config file (config.json). Required if using
a custom model.
--custom_vocab CUSTOM_VOCAB
Path to the custom vocab file (vocab.json). Required if using a
custom model.
--custom_model_url CUSTOM_MODEL_URL
URL to download the custom model as a zip file. Optional, but
will be used if provided. Examples include David Attenborough's
model: 'https://huggingface.co/drewThomasson/xtts_David_Attenbor
ough_fine_tune/resolve/main/Finished_model_files.zip?download=tr
ue'. More XTTS fine-tunes can be found on my Hugging Face at
'https://huggingface.co/drewThomasson'.
--temperature TEMPERATURE
Temperature for the model. Defaults to 0.65. Higher Tempatures
will lead to more creative outputs IE: more Hallucinations.
Lower Tempatures will be more monotone outputs IE: less
Hallucinations.
--length_penalty LENGTH_PENALTY
A length penalty applied to the autoregressive decoder. Defaults
to 1.0. Not applied to custom models.
--repetition_penalty REPETITION_PENALTY
A penalty that prevents the autoregressive decoder from
repeating itself. Defaults to 2.0.
--top_k TOP_K Top-k sampling. Lower values mean more likely outputs and
increased audio generation speed. Defaults to 50.
--top_p TOP_P Top-p sampling. Lower values mean more likely outputs and
increased audio generation speed. Defaults to 0.8.
--speed SPEED Speed factor for the speech generation. IE: How fast the
Narrerator will speak. Defaults to 1.0.
--enable_text_splitting ENABLE_TEXT_SPLITTING
Enable splitting text into sentences. Defaults to True.
Example: python script.py --headless --ebook path_to_ebook --voice path_to_voice
--language en --use_custom_model True --custom_model model.pth --custom_config
config.json --custom_vocab vocab.json⚠️ Legacy-Depricated Old Use Instructions
-
Run the Script:
python custom_model_ebook2audiobookXTTS_gradio.py
-
Open the Web App: Click the URL provided in the terminal to access the web app and convert eBooks.
python ebook2audiobook.py <path_to_ebook_file> [path_to_voice_file] [language_code]- <path_to_ebook_file>: Path to your eBook file.
- [path_to_voice_file]: Optional for voice cloning.
- [language_code]: Optional to specify language.
python custom_model_ebook2audiobookXTTS.py <ebook_file_path> <target_voice_file_path> <language> <custom_model_path> <custom_config_path> <custom_vocab_path>- <ebook_file_path>: Path to your eBook file.
- <target_voice_file_path>: Optional for voice cloning.
- : Optional to specify language.
- <custom_model_path>: Path to
model.pth. - <custom_config_path>: Path to
config.json. - <custom_vocab_path>: Path to
vocab.json.
You can also use Docker to run the eBook to Audiobook converter. This method ensures consistency across different environments and simplifies setup.
To run the Docker container and start the Gradio interface, use the following command:
-Run with CPU only
docker run -it --rm -p 7860:7860 --platform=linux/amd64 athomasson2/ebook2audiobookxtts:huggingface python app.py-Run with GPU Speedup (Nvida graphics cards only)
docker run -it --rm --gpus all -p 7860:7860 --platform=linux/amd64 athomasson2/ebook2audiobookxtts:huggingface python app.pyThis command will start the Gradio interface on port 7860.(localhost:7860)
- For more options like running the docker in headless mode or making the gradio link public add the
-hparameter after theapp.pyin the docker launch command
Example of using docker in headless mode or modifying anything with the extra parameters + Full guide
first for a docker pull of the latest with
docker pull athomasson2/ebook2audiobookxtts:huggingface- Before you do run this you need to create a dir named "input-folder" in your current dir which will be linked, This is where you can put your input files for the docker image to see
mkdir input-folder && mkdir Audiobooks- In the command below swap out YOUR_INPUT_FILE.TXT with the name of your input file
docker run -it --rm \
-v $(pwd)/input-folder:/home/user/app/input_folder \
-v $(pwd)/Audiobooks:/home/user/app/Audiobooks \
--platform linux/amd64 \
athomasson2/ebook2audiobookxtts:huggingface \
python app.py --headless True --ebook /home/user/app/input_folder/YOUR_INPUT_FILE.TXT-
And that should be it!
-
The output Audiobooks will be found in the Audiobook folder which will also be located in your local dir you ran this docker command in
docker run -it --rm \
--platform linux/amd64 \
athomasson2/ebook2audiobookxtts:huggingface \
python app.py -h
and that will output this
user/app/ebook2audiobookXTTS/input-folder -v $(pwd)/Audiobooks:/home/user/app/ebook2audiobookXTTS/Audiobooks --memory="4g" --network none --platform linux/amd64 athomasson2/ebook2audiobookxtts:huggingface python app.py -h
starting...
usage: app.py [-h] [--share SHARE] [--headless HEADLESS] [--ebook EBOOK] [--voice VOICE]
[--language LANGUAGE] [--use_custom_model USE_CUSTOM_MODEL]
[--custom_model CUSTOM_MODEL] [--custom_config CUSTOM_CONFIG]
[--custom_vocab CUSTOM_VOCAB] [--custom_model_url CUSTOM_MODEL_URL]
[--temperature TEMPERATURE] [--length_penalty LENGTH_PENALTY]
[--repetition_penalty REPETITION_PENALTY] [--top_k TOP_K] [--top_p TOP_P]
[--speed SPEED] [--enable_text_splitting ENABLE_TEXT_SPLITTING]
Convert eBooks to Audiobooks using a Text-to-Speech model. You can either launch the
Gradio interface or run the script in headless mode for direct conversion.
options:
-h, --help show this help message and exit
--share SHARE Set to True to enable a public shareable Gradio link. Defaults
to False.
--headless HEADLESS Set to True to run in headless mode without the Gradio
interface. Defaults to False.
--ebook EBOOK Path to the ebook file for conversion. Required in headless
mode.
--voice VOICE Path to the target voice file for TTS. Optional, uses a default
voice if not provided.
--language LANGUAGE Language for the audiobook conversion. Options: en, es, fr, de,
it, pt, pl, tr, ru, nl, cs, ar, zh-cn, ja, hu, ko. Defaults to
English (en).
--use_custom_model USE_CUSTOM_MODEL
Set to True to use a custom TTS model. Defaults to False. Must
be True to use custom models, otherwise you'll get an error.
--custom_model CUSTOM_MODEL
Path to the custom model file (.pth). Required if using a custom
model.
--custom_config CUSTOM_CONFIG
Path to the custom config file (config.json). Required if using
a custom model.
--custom_vocab CUSTOM_VOCAB
Path to the custom vocab file (vocab.json). Required if using a
custom model.
--custom_model_url CUSTOM_MODEL_URL
URL to download the custom model as a zip file. Optional, but
will be used if provided. Examples include David Attenborough's
model: 'https://huggingface.co/drewThomasson/xtts_David_Attenbor
ough_fine_tune/resolve/main/Finished_model_files.zip?download=tr
ue'. More XTTS fine-tunes can be found on my Hugging Face at
'https://huggingface.co/drewThomasson'.
--temperature TEMPERATURE
Temperature for the model. Defaults to 0.65. Higher Tempatures
will lead to more creative outputs IE: more Hallucinations.
Lower Tempatures will be more monotone outputs IE: less
Hallucinations.
--length_penalty LENGTH_PENALTY
A length penalty applied to the autoregressive decoder. Defaults
to 1.0. Not applied to custom models.
--repetition_penalty REPETITION_PENALTY
A penalty that prevents the autoregressive decoder from
repeating itself. Defaults to 2.0.
--top_k TOP_K Top-k sampling. Lower values mean more likely outputs and
increased audio generation speed. Defaults to 50.
--top_p TOP_P Top-p sampling. Lower values mean more likely outputs and
increased audio generation speed. Defaults to 0.8.
--speed SPEED Speed factor for the speech generation. IE: How fast the
Narrerator will speak. Defaults to 1.0.
--enable_text_splitting ENABLE_TEXT_SPLITTING
Enable splitting text into sentences. Defaults to True.
Example: python script.py --headless --ebook path_to_ebook --voice path_to_voice
--language en --use_custom_model True --custom_model model.pth --custom_config
config.json --custom_vocab vocab.json
Click to see images of Web GUI
Models built to be better at a specific voice. Check out my Hugging Face page here.
To use a custom model, paste the link of the Finished_model_files.zip file like this:
David Attenborough fine tuned Finished_model_files.zip
For a custom model a ref audio clip of the voice will also be needed: ref audio clip of David Attenborough
More details can be found at the Dockerfile Hub Page.
To find already fine-tuned XTTS models, visit this Hugging Face link 🌐. Search for models that include "xtts fine tune" in their names.
Rainy day voice
Magic.Tree.House.45.Osborne.Mary.Pope.-.A.Crazy.Day.With.Cobras.mp4
David Attenborough voice
david_atten_alice_text.mp4
- Huggingface space is running on free cpu tier so expect very slow or timeout lol, just don't give it giant files is all
- Best to duplicate space or run locally.
.epub,.pdf,.mobi,.txt,.html,.rtf,.chm,.lit,.pdb,.fb2,.odt,.cbr,.cbz,.prc,.lrf,.pml,.snb,.cbc,.rb,.tcr- Best results:
.epubor.mobifor automatic chapter detection
- Creates an
.m4bfile with metadata and chapters. - Example Output:
- "It's slow!" - On CPU only this is very slow, and you can only get speedups though a NVIDIA GPU. Discussion about this For faster multilingual generation I would suggest my other project that uses piper-tts instead(It doesn't have zero-shot voice cloning though, and is siri quality voices, but it is much faster on cpu.)
- "I'm having dependency issues" - Just use the docker, its fully self contained and has a headless mode, add
-hparameter after theapp.pyin the docker run command for more information. - "Im getting a truncated audio issue!" - PLEASE MAKE AN ISSUE OF THIS, I don't speak every language and I need advise from each person to fine tune my sentense splitting function on any other languages.😊
- "The loading bar is stuck at 30% in the web gui!" - The web gui loading bar is extreamly basic as its just split between the three loading steps, refer to the terminal and what sentense it's on for a more accurate gauge on where is it progress wise.
- Any help from people speaking any of the supported langues to help with proper sentence splitting methods
- Potentially creating readme Guides for Multiple languages(Becuase the only language I know is English 😔)
-
Coqui TTS: Coqui TTS GitHub
-
Calibre: Calibre Website
