I’m new to TTS models but is this something I can plug into my own engine like with LLMs, or does it require the Python stack it ships with?