Script to create a custom voice in the Coqui API

create_voice.py

import sys
import requests
import argparse
import json
import os
from requests_toolbelt.multipart.encoder import MultipartEncoder


try:
    API_TOKEN = os.environ["COQUI_API_TOKEN"]
except KeyError:
    print(
        "Set the COQUI_API_TOKEN environment variable with your API token to use this script.",
        file=sys.stderr,
    )
    sys.exit(1)


def create_voice(voice_name: str, file_path: str) -> dict:
    # Open the file in binary mode
    with open(file_path, "rb") as file:
        # Create a multi-part encoded request
        multipart_data = MultipartEncoder(
            fields={
                "name": voice_name,
                "file": ("audio", file, "application/octet-stream"),
            }
        )

        # Set the headers for the request
        headers = {
            "Content-Type": multipart_data.content_type,
            "Authorization": f"Bearer {API_TOKEN}",
        }

        # Make the POST request
        response = requests.post(
            "https://app.coqui.ai/api/v2/voices", data=multipart_data, headers=headers
        )

    # Return the response JSON
    return response.json()


def main():
    parser = argparse.ArgumentParser(description="Create a custom voice clone.")
    parser.add_argument("name", help="The name of the custom voice.")
    parser.add_argument("file_path", help="The path to the file to be sent.")

    args = parser.parse_args()

    # Make the POST request
    response_json = create_voice(args.name, args.file_path)

    # Print the response JSON
    print(json.dumps(response_json, indent=2))


if __name__ == "__main__":
    main()