Serving XTTS with BentoML

XTTS is a voice generation model that lets you clone voices into different languages by using just a quick 6-second audio clip.

This is a BentoML example project, demonstrating how to build a text-to-speech inference API server using the XTTS model. See here for a full list of BentoML example projects.

Install dependencies

git clone https://github.com/bentoml/BentoXTTS.git
cd BentoXTTS

# Recommend Python 3.11
pip install -r requirements.txt

Run the BentoML Service

We have defined a BentoML Service in service.py. Run bentoml serve in your project directory to start the Service. You may also set the environment variable COQUI_TTS_AGREED=1 to agree to the terms of Coqui TTS.

$ COQUI_TOS_AGREED=1 bentoml serve .

2024-01-18T11:13:54+0800 [INFO] [cli] Starting production HTTP BentoServer from "service:XTTS" listening on http://localhost:3000 (Press CTRL+C to quit)
/workspace/codes/examples/xtts/venv/lib/python3.10/site-packages/TTS/api.py:70: UserWarning: `gpu` will be deprecated. Please use `tts.to(device)` instead.
  warnings.warn("`gpu` will be deprecated. Please use `tts.to(device)` instead.")
 > tts_models/multilingual/multi-dataset/xtts_v2 is already downloaded.
 > Using model: xtts

The server is now active at http://localhost:3000. You can interact with it using the Swagger UI or in other different ways.

CURL

curl -X 'POST' \
  'http://localhost:3000/synthesize' \
  -H 'accept: */*' \
  -H 'Content-Type: application/json' \
  -d '{
  "text": "It took me quite a long time to develop a voice and now that I have it I am not going to be silent.",
  "lang": "en"
}' -o output.wav

Python client

This client returns the audio file as a Path object. You can use it to access or process the file. See Clients for details.

import bentoml

with bentoml.SyncHTTPClient("http://localhost:3000") as client:
        result = client.synthesize(
            text="It took me quite a long time to develop a voice and now that I have it I am not going to be silent.",
            lang="en"
        )

For detailed explanations of the Service code, see XTTS: Text to speech.

Deploy to BentoCloud

After the Service is ready, you can deploy the application to BentoCloud for better management and scalability. Sign up if you haven't got a BentoCloud account.

Make sure you have logged in to BentoCloud, then run the following command to deploy it.

bentoml deploy .

Once the application is up and running on BentoCloud, you can access it via the exposed URL.

Note: For custom deployment in your own infrastructure, use BentoML to generate an OCI-compliant image.