from superclonica import Superclonica
tts = Superclonica() # GPU if available, otherwise CPU
audio, sample_rate = tts.synthesize(
"Hello! Your new voice is ready.", voice="voice.json", lang="en", seed=7
)
tts.save("hello.wav", audio)
Long texts, dialogues, custom voices and a comparison with pyttsx3 and gTTS: Text to speech in Python.
Speech from the command line
superclonica --text "Hello! Your new voice is ready." --voice voice.json --lang en --output hello.wav
With Docker
A ready image with an HTTP server and the command line: superclonica/superclonica. The model (about 400 MB) downloads on the first start into the superclonica-models volume.
CPU
docker run -d -p 8000:8000 -v superclonica-models:/models superclonica/superclonica:cpu
NVIDIA GPU
Requires the NVIDIA driver and the NVIDIA Container Toolkit.
docker run -d --gpus all -p 8000:8000 -v superclonica-models:/models superclonica/superclonica:cuda
Speech through the server
The voice goes into the request as the JSON file itself. It is about 0.3 MB, so build the request in a file:
python -c "import json; json.dump({'text': 'Hello! Your new voice is ready.', 'lang': 'en', 'voice': json.load(open('voice.json'))}, open('request.json', 'w'))"
curl -X POST localhost:8000/synthesize -H "Content-Type: application/json" -d @request.json -o hello.wav
API key, settings, custom voices and running in production: Text-to-speech API in Docker.