Potential Drawbacks of Using Others' Gradio Apps via API

Hello everyone,

I’m considering using a text-to-speech model via the API of a Gradio app, but I wanted to get your insights on the potential drawbacks. I believe the model likely requires a GPU, which could incur costs for the app owner. How does the billing process work in this case, or is the service provided for free by the owner?

P.S.: Please mention any other drawbacks besides this one.