A simple GUI application for quickly generating A.I. voiceovers using the XTTSv2 model from Coqui.
Note: Immediately after starting up, the application will download nearly 4GB of required TTS model files.
If you've already installed the excellent uv tool, the following command will
- create a virtual environment at
./.venv - activate the virtual environment
- install any missing application dependencies
- launch this app
- deactivate the virtual environment when execution is complete
uv run streamlit run main.py
If you aren't using uv, something like the following should get you up and running:
python -m virtualenv .venv
source .venv/bin/activate
pip install .
streamlit run main.py
deactivate
For my purposes the "Viktor Menelaos" and "Szofi Granger" voices have provided the most consistently realistic intonation for English voiceovers, but your mileage may vary, so try them all out.