Text-to-Speech API

The speech synthesis endpoint accepts text and uses a speech synthesis model running on the device to generate an audio file. This example uses a MeloTTS model package for the AX630C platform.

Preparation

Before calling the endpoint, install the speech synthesis model and tools required by the example. Refer to the model library for platform support.

Before running this example program, please ensure that the following preparations have been completed on the LLM device:

1. Use the apt package manager to install the llm-model-melotts-en-us model package.

apt install llm-model-melotts-en-us

2. Install the ffmpeg tool.

apt install ffmpeg

3. After installation, restart the OpenAI service to make the new model take effect.

systemctl restart llm-openai-api

Example Program

On the PC side, use the OpenAI API to pass text information to implement the text-to-speech function. Before running the example program, modify the IP part of base_url below to the actual IP address of the device.

from pathlib import Path
from openai import OpenAI

client = OpenAI(
    api_key="sk-",
    base_url="http://192.168.20.186:8000/v1"
)

speech_file_path = Path(__file__).parent / "speech.mp3"
with client.audio.speech.with_streaming_response.create(
  model="melotts-en-us",
  voice="alloy",
  input="The quick brown fox jumped over the lazy dog."
) as response:
  response.stream_to_file(speech_file_path)

Request Parameters

Response Example

• The generated audio file data will be stored in the speech_file_path path defined in the example program.