BazedFrog/SongGeneration-Studio ? reverse-engineered prompt

Build me a clean, polished web app for making AI songs from text prompts and reference audio.

I want people to be able to write lyrics in sections like verse, chorus, bridge, pick a genre, mood, voice, instruments, and tempo, then generate a full song with vocals and instrumental music. It should also let users upload a reference track so the new song can match the style and vibe, and it should support making just the full mix, vocals only, or instrumental only.

Please include a simple library where saved generations can be organized, renamed, given cover art, and exported as audio or video. I also want batch generation so multiple songs can be queued at once, plus a nice interface that feels easy to use for non experts. If anything needs current setup details, look up the latest docs online and make the best working version.