Meeseeks-Mr/Ethera ? reverse-engineered prompt

Build me a clean, modern web app for generating music with HeartMuLa.

I want a simple interface where someone can type lyrics or a short prompt, pick a style, choose the voice mode, Male, Female, or Instrumental, and set how long the track should be, from a few seconds up to about 4 minutes. The app should feel polished and futuristic, with a glassmorphism look, and it should clearly show a composing state while the music is being generated.

Please make it run locally on my own machine, since privacy matters, and connect it to the backend so it can talk to the music model and return audio when it’s done. If anything about the model setup or API details is unclear, look up the current docs online and wire it up in the simplest working way. I’d also like the page to feel friendly and easy to use, not like a technical dashboard.