My custom Text To Speech API server.
After install and launch, you can use my TTS server in multiple ways :
1 - 🗔 In Linux command line, with the included script : bjornulf_tts.sh
2 - 💬 In my AI bot chat : Bjornulf Lobe Chat Fork
3 - 🎨 Inside Comfyui with my custom node TTS number 31 : Bjornulf Comfyui TTS node
❤️❤️❤️ https://ko-fi.com/bjornulf ❤️❤️❤️
1 - 📝 Text/Chat AI generation : Bjornulf Lobe Chat Fork
2 - 🔊 Speech AI generation : Bjornulf Text To Speech (you are here)
3 - 🎨 Image AI generation : Bjornulf Comfyui custom nodes
Recommended : create virtual environment and activate it.
python3 -m venv ~/venv/xtts2
source ~/venv/xtts2/bin/activate
Choose location of course before cloning.
git clone https://github.com/justUmen/Bjornulf_XTTS/
cd Bjornulf_XTTS
pip install -r requirements.txt
cd xtts_api_server
python bjornulf_xtts_server.py
I use a .desktop file to launch the server, with a custom icon.
Here is mine, change it as needed :
[Desktop Entry]
Name=xtts_server
Comment=xtts_server
Exec=kitty --class "kitty_xtts" --title "kitty - xtts" zsh -i -c 'source /home/umen/venv/xtts/bin/activate && cd /home/umen/SyNc/Forks/Bjornulf_XTTS/xtts_api_server/ && systemd-inhibit --what=sleep --who="bjornulf app" --why="Preventing sleep to complete the task" --mode=block python bjornulf_xtts_server.py || read'
Icon=/home/umen/SyNc/Conf/desktop_applications/icons/speaker.svg
Terminal=false
Type=Application
Categories=Utility;
StartupWMClass=kitty_xtts
Note that i use kitty terminal and a systemd-inhibit to prevent sleep while the server is running. (You can simplify all that if you want)
Of course put your own paths -> where is located your bjornulf_xtts_server, your icon, your virtual environment, etc...
I also use a custom class "kitty_xtts" to manage it in my window manager. (In my case just to keep it in a specific workspace.)
To open the server i just click on this icon.
To close the server, i just go on the kitty window terminal and CTRL + C inside.
Just put a voice sample .wav file in the xtts_api_server/speakers folder. (I recommend creating a subfolder for each language. like en, fr, etc...)
Here you can download a sample with Attenborough voice (English so put in a folder en) : https://drive.google.com/file/d/1JOSpavgN0GS2OswXbQCpqL5kYV0nSr6n/view?usp=sharing
Note : You can actually use an english sample to speak another language. (Not ideal but you can try it out.)
Usage: /home/umen/SyNc/Forks/Bjornulf_XTTS/bjornulf_tts.sh [-l language] [-s speaker] [-f] [-L] [-S] [-m] <text>
Options:
-l language Set the language (default: en)
-s speaker Set the speaker (default: default)
-f Force overwrite of existing audio file (Use if audio is not good enough, it will try another generation)
-L List available languages
-S List available speakers
-m Mute audio playback (download only, don't play the audio)
<text> The text to convert to speech
Example:
./bjornulf_tts.sh -l en -s fr/une_voix "Bonjour tout le monde."
This script will generate a .mp3 file in the path provided in the script, edit if you want to change it : BASE_DIR="$HOME/Audio/Bjornulf/$LANGUAGE/$SPEAKER")
It will not try to regenerate the audio file if it already exists! (So the TTS server doesn't need to be running to use this script - if you know you have the audio file already)
For example, you can generate an audio file system is operational and you can safely use my script to simply play the audio when your computer starts.
I recommend to use with an alias in .bashrc/.zshrc, for example something like :
alias bjornulf_tts='/path/of/your/Bjornulf_XTTS/bjornulf_tts.sh
alias tts='/path/of/your/Bjornulf_XTTS/bjornulf_tts.sh
alias talk='/path/of/your/Bjornulf_XTTS/bjornulf_tts.sh
After that you can use it like this :
tts hello world
talk hello world
You can also create language and speaker aliases in the script if you want to use them often.
Example :
alias parler='/path/of/your/Bjornulf_XTTS/bjornulf_tts.sh -l fr -s fr/une_voix'
After that you can use it like this :
parler "Bonjour tout le monde."