Quick Piper GUI

This is a simple GUI application which allows you to quickly generate and play text-to-speech using Piper. The output of the program can be connected to your favourite VOIP application (on Linux using qpwgraph or Helvum), allowing you to speak even when circumstances don't permit. Pygame is used for the audio output which provides a persistent audio device. Additionally, you can use it to quickly generate text-to-speech audio; the last output is saved to the ./output folder

Requirements

• Python 3.13.2 (tested, should work with Python 3.x versions)
• dimits
• pygame
• Piper TTS model: download an existing model or train your own

Install the necessary packages using pip:

pip install -r requirements.txt

Usage

Add your Piper models together with the coresponding JSON files to: ~/piper (that is where dimits searches, I am not sure if it can be changed). Attention: the JSON files should be named exactly like the model files, so if our model file is named en_US-example-medium.onnx, our JSON file will be named en_US-example-medium.onnx.json.

Start the program with the name of your model as the first parameter without the extension:

python gui.py en_US-example-medium

demo.mp4

Shortcuts

• CTRL+Enter: read text-to-speech
• CTRL+A: clear the input box

Connect output to an application

Disconnect your microphone (optional) and connect one channel of the python3.x [Audio Stream] output to an input. Connecting both channels to an input will double the volume and cause distortion.

qpwgraph_demo.mp4

TODO

These features will probably not be completed in the near future. They are here just for guidance.

• Voice selection in the menu
• Playback control buttons for pause, resume, and stop
• Quick phrases or memory
• Global shortcuts
• Better error handling
• Modern GUI