This service uses hugging face's inference API to query text-to-audio AI models. Any model from the hub available on the inference API that outputs audio and takes a json input with the following structure can be used:
{
"inputs" : "your input text"
}
Check the related documentation for more information.